Yes, AI will be able to generate infinite sequences of appalling takes and devastating take-downs, dos-i-dos-ing off into infinity like pairs of brooms in a sorcerer's apprentice of bad online discourse.
Well said but it only partially addresses a point that Matt missed. Andreesen’s point is taken narrowly if we think of isolation solely as social isolation and, therefore Matt erroneously treats the Internet and social media as one and the same. The Internet is a tremendous advance in communication in terms of accessing and identifying information and connecting people. Many bad consequences, of course, but that is blind to the positive effects.
I think the techno-optimist position is not that all technological progress is always good, but rather it's a bad idea to trust activists and government bureaucrats who generally aren't smart enough to contribute to technological progress to determine which progress is and is not allowed.
Noah Smith has a piece out suggesting America's failure to build things (including new uses of technology) is a result of too *few* bureaucrats with too *little* power:
There is a difference between building things, even things that use new technology and the creation of new technology. With that said I am very familiar with the problem Smith addresses and mostly agree. Governments would likely be better off having more full time staff and less consultants. The idea that the consultants are better is a fallacy in itself.
I think it's more that our culture of outsourcing led to us forgetting what consultants can actually DO.
Consultants can tell you what's wrong with your organization with a kind of insight that is often hard to have when you're in the thick of everything.
Consultants can't do the actual job; if they could, they'd be your competitors, not consultants.
Ed: And I think that's the REAL problem with outsourcing everything to The Groups. They may do a lot of good work advocating for various good things, but they AREN'T design firms, they're activists and non-profits. Affordable housing advocates don't know how to design houses because, surprise, they aren't construction workers, architects, or engineers! That's how we end up with stupid ideas like "La Sombrita". Design by a primarily politically-oriented committee is just not good design.
The most common consultants in local government is technical expertise. Civil engineers for example design roads, surveyors, planners, architects, etc.
It has become the norm to hire consultants to do this work. It could be done in house. It could be that some of it is done in house and some by consultants.
If the government entity has consistent work then having in house professionals doing it (or a part of it) could make more sense.
These consultants are very expensive now. It’s interesting with the rise of HR that these decisions are not more optimized. I have never heard an HR person at my org be curious about how this gets done and if we’re making the best calls. Noah Smiths paper could be a good conversation starter.
Something that folks like Noah don't always recognize, though, is that as expensive as they can be, technical consultants can often be the only reasonable option in rural areas where the need can't justify full-time positions.
The real problem is that in denser jurisdictions, the regulatory burden means that the only people who CAN do the work are expensive consultants.
80 years ago, a disgruntled mid-manager or foreman on a construction crew could grab some buddies and strike out on their own as a contractor. These days, you need an in-house lawyer to navigate zoning laws and safety regs.
Totally right on the rural areas don’t have enough work to justify in house.
In my head, I distinguish between consultants and contractors. That is probably just a social norm. Many midsize governments and beyond could do more in-house professional work (i.e engineering) and they could also build more. Building big projects would be tough. If you’re LA County then you could probably run an enormous construction company in house.
On the design side, single projects can run $500k plus. I think the better place for in house is in doing a collection of the small projects. Then using the consultants on the larger ones. The bigger point is it’s become so commonplace to use consultants that it’s not really being questioned anymore. Pendulums shift and maybe it’s time for this one to shift back?
I am sure a lot of places would push back using a risk based argument but my guess is they don’t have much real data to make that case. They’re just speculating.
My only criticism here is that I think it'd be better to encourage an ecosystem of small-to-larger-mid-size contractors that are competent to consult on mid-to-large projects, rather than having an ecosystem of out-and-out consulting firms.
There is also the problem of how you hire good people to go into a bureaucracy that a) is not so good and b) has gotten used to controlling whether changes can happen around it.
You seem to be sanewashing Andreessen’s piece. He directly says that the enemies aren’t just bureaucrats, but *anyone* who theorizes about risk, worst-case scenarios, etc.
Exactly, see my comment above about his totally pathetic output since he issued the clarion call that it was "Time to Build!" in 2020. He then took all his money, networks, and know-down and built over the next three years... Bored Ape NFTs, crypto-video games, the next WeWork, and Elon's version of Twitter. Yay...?
This during perhaps the most urgent time in recent history for building actually-important things.
Drug delivery, robotic boats, military tech, healthcare tech, restaurant management software, A16Z is really huge, multiple multi million dollar investments per week, and sprawls into all sorts of things. Occasionally there's something funny and/or dumb that the more mainstream press will talk about, like NFTs, but it's a mistake to think that this sort of press is a good summary of A16Z's investment activity.
I'm not denying that a16z has a variety of different investments in its portfolio, but do any of them actually "build" anything important, according to his own definition?
His first argument is that we need to "build our way out of" various serious problems facing humanity, citing past solutions to big problems and basic material needs like the Green Revolution, indoor heating, electric lighting, vaccines, etc. All great things! I agree! Build away!
Some of these things have even been re-invented in the last few years in ways that have really changed the game: mRNA Vaccines, for example. A Nobel Prize-worthy accomplishment! Did a16z invest in those? Not really in his wheelhouse, I guess. (They did write a blog about it, leaving out any mention of the decades of patient, unheralded, government-funded NIH research that made this breakthrough possible). Instead of wasting money on something as high-risk as that, a16z invested in various healthtech apps that haven't revolutionized healthcare in any way, and certainly not commensurate with these aforementioned solutions. One of their top three banner investments in healthcare, PatientPing, isn't really "building" anything, except a kind of ERP/CRM for hospitals. Yay. When I squint, I could see another of their investments--Nautilus Biotechnology and their race to quantify the human proteome--going somewhere, with applications in diagnosis for cancer. But, again, this is no "Moonshot." And it's been quiet for three years since a16z slid their chips over the table.
How about some of the other real challenges of our times? Climate Change? Energy? The Sixth Great Extinction? Food security? Housing? Inequality? Migration? "The Great Stagnation?" Or even to address the more outlandish Existential Risks that Silicon Valley billionaires are really on about these days? Maybe I'm missing something, but I don't see much there in their portfolio. Instead, I see a lot of stuff that offers profitable new B2C and B2B businesses leveraging very incremental innovations that would look familiar to anyone who has followed Silicon Valley for the last 30 years.
Even on areas he has very strong opinions on, like the absolute imperative to (and I quote one of his more famous Tweets from last year) "build 1,000 state of the art nuclear plants in the US and Europe, right now." Has he invested in that? "We won't, but we should." Indeed! Why would he waste his money? It's a terrible business, which is why 1,000 state of the art nuclear plants won't get built in the US and Europe, even if "we should." Instead, a16z invests in conventional things that have a clear line to profitability. That's fine. But it's not the clarion call to the Builders he's trumpeting so grandiosely. And it betrays the fact that what he's really saying here is that he doesn't want the government to regulate him because that might hamper his profits a little bit.
His second argument is that everything good comes from productivity growth, which comes from technological progress. Which is undoubtedly true. It's Econ 101. This is an area that's easier for the likes of a16z to address, surely? Except that the productivity benefit of IT has been surprisingly disappointing since the advent of consumer computing and the Internet. Again, we get stuff like AirBnB, Facebook, Instagram, Pinterest, and Skype, which are certainly successful and useful consumer tech applications. But that haven't, any of them, appreciably increased Total Factor Productivity, which has fallen below the mid-century trend-line since 1973. Non-productivity-enhancing consumer tech brands helping us to mostly "amuse ourselves to death" are BY FAR the largest bucket of investment by a16z over the years, and I don't even need to pick on the likes of Clubhouse or the fallen stars like Buzzfeed or Groupon to make the point that all these billions invested haven't actually increased the productivity of the American economy.
OK, well what about the more serious B2B stuff? Well, though I certainly very much enjoy Slack, is there any evidence that it increased enterprise productivity? Box, Asana, etc... I've used them all. They're not actually "revolutionizing the enterprize," marketing content notwithstanding. You certainly aren't seeing their impact in labor productivity, which has been even more stagnant than TFP, and especially disappointing since the Great Recession. Today, wages are finally up, but in the absence of productivity growth, are just causing "stagflation" fears.
OK, what about the really cool AI/ML stuff like databricks? Miiiiight go somewhere. But that revolution has been nigh for a decade now. I've sold the dream myself for as long. Meanwhile, many of the applications have been gobbled up by the likes of Facebook (Wit.AI) to do what, exactly...? Better ad-targeting!
Again, none of these things are actually game-changing. And maybe that's an unfair criterion. Maybe the Secular Stagnation Hypothesis is right and the truth is that the low-hanging fruit of the caliber that Andreesen alluded to are already invented! Either way, nothing that a16z is betting its money on even has the potential to be that. And I'm sure he knows it.
So the rhetoric sounds really cool and future-y to people who aren't tracking the dollars or the actual outcomes. Much like with Elon Musk's SpaceX, which is really just another DoD contractor for launching military satellites (including proprietary Starlink/Starshield ones) into space, under the veneer of making us a multi-planet species. I don't hate that Elon Musk is building really good rockets and that Mark Andreesen is funding really good SaaS companies. But I wish they would just stop being so vainglorious about it. And that they would both acknowledge how much they actually rely upon the US government for a big chunk of their revenue, the R&D, and of the soft- and hard-infrastructure that make their profits possible.
So which among these 670 funded companies (or "Builders") would you say fits Andreesen's own criteria for "building?"
Are any of them providing "the things we urgently need but don’t have"? In housing, education, manufacturing, transportation, or the other areas specifically cited by him in this essay and its precursor three years ago? Do any of them strike you as "aggressive investment in new products, in new industries, in new factories, in new science, in big leaps forward"?
Because I'm familiar with many of these brands, especially the featured ones, and I'm just not seeing it there. "We admire the great technologists and industrialists who came before us, and we aspire to make them proud of us today." Would the inventors of the Green Revolution, electric lighting, indoor heating, vaccines, or any of the other actually-phase-shifting technologies that he cites in his Manifesto be proud of anything in this portfolio of hundreds of startups?
Yep. Of course it would be ideal to regulate only the amount that is necessary - but the risk too often, is that you get the NRC and totally kill a technology. The counter to that is convincing people to be more pro-technology generally.
I'm referring specifically to Andreessen's manifesto, which I would say is a strawman. I definitely agree that technology is one of the main drivers of a better standard of living. As for regulation, it gets complicated. I'm particularly annoyed about US nuclear and environmental regulation, but mainly in terms of poor implementation rather than on principle. A technology like crypto seems perhaps under-regulated given the amounts of money people have been scammed out of or have foolishly traded away.
The only people who've lost money in crypto scams are scammers themselves. If you buy a worthless digital "asset" in the hopes that you're able to turn around and sell it before it inevitably goes to zero, then you're a scammer.
This is such a bizarre take. Sure, YOU know that crypto is a scam, but a lot of people who were scammed don't understand the technology at all. I have an Uncle who was persuaded by scammers that crypto was safe, that it's returns were guaranteed, and that bitcoin offered him an opportunity to get out of the financial troubles he was in. It wasn't his most brilliant decision, obviously, but it was absolutely a scam that preyed on him. Turning around and blaming him for falling for it because of your far superior understanding of it seems callous and completely at odds with the reality of how these things work.
Crypto has been a huge positive productivity shock to ransom takers. Hospitals paying 1m to not get patient records deleted have lost money due to (and ultimately “in”) crypto without being scammers
It’d be simple to think that when you’re smart enough to understand the difference between “investment” and “speculation,” and have heard of “the bigger fool” concept.
Most haven’t, including most people who “invested” in crypto.
I invested all my liquid wealth into NFTs that I researched by scanning a QR code on a sticker affixed to a light pole. The sticker also featured a cute logo of a pixelated teddy bear or some such. Sure the market is down *now* but I’m in it for the long haul.
There’s a difference between people who came out on top of something they didn’t really understand, and people whose day job is basically to scam those who aren’t lucky enough to come out on top.
Doesn't seem like there is much of a principle (other than "digital") that would separate this definition of scammer from everyone who buys stocks or homes. Unless you really think every crypto customer knew it would "inevitably go[] to zero," which seems very unlikely.
I'm against the government regulating crypto. Doing so would give the government's blessing to something that fundamentally is itself a scam, with no redeeming social value. If people want to blow their money in a speculative frenzy, that's on them.
The NRC actually permits interesting thought experiences. Would it be better to have 1) the current NRC or 2) no NRC? Was there, ex ante, a viable way to establish the NRC that would not have strangled nuclear power?
I think a big unanswered question sitting between Yglesias and Andreeson is at what point should the government start regulating?
For example, I agree that at some point that AI will probably need some regulation. I don’t think we are at that point yet.
To effectively regulate you need to know the actual downsides and tradeoffs for a technology, not imagined ones. Regulate too early and you crush innovation. Regulate too late, and the new tech can potentially cause real and avoidable harm.
So who is supposed to decide how to manage fentanyl and nuclear power and perhaps next generation AI?
I do think Andreessen has gone for the full caricature of techno-optimism that everything should be unregulated, essentially. There are also open source maximalists out there who don't exactly have nuanced positions.
For AI, it'll be managed by the same entities that regulate spreadsheets, operating systems, and all other software: the market in general and the FTC when antitrust concerns arise.
The market is not a regulator. It rewards maximizing externalization of costs/risks, the exact opposite function of regulators. AI will eventually be very, very good at performing these functions.
The challenge I see is developing the expertise necessary to identify AI models that either by design or by flaw pose outsized risk of external harm. For example, imagine an AI that becomes very, very effective at stock manipulation. It recognizes trading patterns that can induce significant movement and constantly seeks to maximize that activity for profit. That is an unqualified negative to the markets, yet the only market response would be to take the same financial positions as the AI to maximize your profits from its trading. You can extrapolate this to various circumstances. AI will be the perfect amoral profit seeker.
In my opinion, at whatever point advanced AI entails a high level of risk, that approach becomes insufficient. Of course, figuring out when that happens is difficult.
* Job loss? Every technology causes that, and it's necessary for higher productivity and a richer world. If we care enough, we can expand our welfare state using our newly created wealth.
* Misinformation? Twitter already produces more misinformation in 60 seconds than a human could consume in a lifetime. We even have computer game videos shared as false combat footage.
* Omnipotent being that kills us all? All this harping is extreme speculation that surpasses even the delusions of crypto bros prophesying the collapse of the global financial system.
If real risks arise, then we can introduce regulation to manage them. Yes, we may wish we had the foresight to proactively introduce regulation and avoid the harms. But the impact is just too unknowable. We humans will muddle through, as we have in every past technological disruption from agriculture to the internet, and be far wealthier for it.
The job loss this time will mostly affect those "intellectual circles [which] are dominated by highly verbal people...". I think this is the underlying reason for the extreme skepticism around AI.
Let’s see… just at the more plausible end of the spectrum:
Highly effective stock market manipulation.
Rapid localization of individuals for any number of illegal purposes.
Fabrication of convincing fraudulent datasets up to falsifying evidence for civil or criminal proceedings.
Generation of false imagery, sound, and video.
But sure, no regulation required for something which would allow foreign state security services to efficiently assassinate a defector hidden by the US government, turn public stock markets into a mockery of efficient capital allocation, completely mis-allocate R&D expenditures, or imprison innocents.
You're arguing that *information networks* are dangerous, not AI. The protocols enable this already, and you aren't putting that genie back in the bottle. The list you have here could have appeared in a breathless Time magazine article about "The World Wide Web" in 2001.
AI risk people usually argue at way too high a level of generality. But giving everyone on earth the skills of a team of computer security researchers and biologists would be incredibly dangerous, for pretty concrete reasons
I think the most-sanewashed version of it is, since tech has outpaced society's ability to deal with failures via politics, we should accelerate innovation in order to make up the difference.
IMO, I don't think that's a good enough reason to stop *trying*.
I think the technologists have some pretty inflated egos. Many of them are very smart, but many bureaucrats are also very smart. This isn't like arguing that no one is smart enough to control a command economy, this guy just thinks he's smarter than everyone else.
Elon Musk has contributed enormously to technological progress, yet he does not strike me as particularly wise. But to be fair, wisdom is not necessarily raw smarts. Yet if one were to evaluate how smart Musk was versus a senior NASA engineer in 2000, I suspect they wouldn't favor him then. Rather, the recent Isaacson biography suggests Musk's success was primarily due to his willingness to put in manic 90 hour weeks and ruthlessly drive his engineers to accomplish something new on a tight deadline. Musk learned a lot about rockets and then drove people to near exhaustion doing things never done before. This pattern was particularly true in enterprises where more vertical integration was feasible, unlike Paypal or Twitter.
Isaacson's book is worth reading because even though it's a bit flowery to gain access, it does a better job demonstrating what was done to produce reusable rockets rather than focusing on the haters. Ultimately, such an obsession with haters misses what produces great technological breakthroughs. Reusable rockets are good. A more helpful question is what motivated Peter Thiel to save it from the brink of bankruptcy in 2008? And what kinds of investments like that today are similarly contingent? The regulatory environment was and is a real issue that demanded constant renegotiation. But it wasn't the main driver of SpaceX. I suspect mRNA vaccines are another great example where the regulatory environment didn't automatically cripple them. But NIH funding politics almost did, if not for the drive of Katalin Kariko to keep bouncing from place to place. It doesn't mean the FDA couldn't use regulatory changes, it absolutely could. It's just not the main story of technology optimism as best I can tell. So it deserves a more proportional analysis in any tech optimism take.
I highly recommend people take a look at my friend Connor's recent essay for more on the mRNA question and hedging our bets in general. I think it approaches this from a better angle; we need to ask where new ideas come from as much and more than than what's conspiring to stop them. https://regions.substack.com/p/how-does-a-society-hedge-its-bets
"Rather, the recent Isaacson biography suggests Musk's success is primarily due to his willingness to put in manic 90 hour weeks and ruthlessly drive his engineers to accomplish something new on a tight deadline."
Well the better question to me then is:
"If the NASA engineers and contractors were so much smarter than Musk, why did they not identify these new things when they probably had decades to do so?"
If the rule is "the people smart enough to build the thing get to decide how to regulate the thing," then we should listen to the people at OpenAI and Anthropic and other leading AI labs, the vast majority of whom are advocating for dramatically increased regulation of AI and are super opposed to what Andreesen is proposing.
These days, Andreesen is a finance guy, not a tech guy. He was a tech guy 30 years ago, but it's not clear to me why that means we should listen to him about AI Governance when the actual cutting-edge tech people are extremely worried and saying we need more AI Governance.
So you only trust the people who have a profit-motive to create new technologies? This line of reasoning would indicate only chemists should be involved in the regulation of pharmaceuticals. Sure, you definitely need to have top chemists involved, but as a rule we ought to include non-technologists for both practical and moral reasons. Andreseen's thesis essentially boils down to "get these g.d. nags out of my way and let the tech bros cook" rather than engaging with the strongest criticisms.
His anti-government screed in this Manifesto is interesting because the only country on the planet with a dominant nuclear power industry is France, which got there through exactly the public-led private sector that he supposedly hates (when a16z isn't invested in Defense Department-funded tech, anyway).
Societies without the Mafia do better than societies with the Mafia. Societies without strong central governments do not do better than societies with strong central governments.
It's the definition of 'strong' that's in question, and how much corruption is present. The Bar association and the American Kennel Club are both non-State organizations, they both are influential in their fields, they both certainly contain some amount of corruption. It's about where the various sliders are set, not about an absolute binary.
One could also ask that about electricity in general (how do we coordinate the paths needed to built transmission wires?) or about the automobile (how do we organize corridors for them to travel freely?) along with a long host of other technologies.
The case that the internet is a beneficial technology shouldn’t be built on loneliness, it should be built on making other science/technology progress faster. The arXiv is legitimately transformative, let alone email, zoom...
Beat me to it. Andreesen's "isolation" framing is absurd, of course. However, is it equally good as the other five things listed alongside it? Perhaps not. But I had a very negative reaction to Matt's "stolen valor" tweet [https://twitter.com/mattyglesias/status/1714047013317210180], when he said that it wasn't even remotely close, when I think the increase in productivity in so many fields can at the very least bring it within remote discussion. I almost pushed back on it in a daily thread last week, but I'm glad I didn't since he wrote this entire article, where it's more appropriate.
It's surprising that neither Andreesen nor Matt brought up the best analogue for the internet in the history of technology, the printing press.
Is anyone willing to make the case against the press? For decades after its invention , there was a great deal of concern about how it was being used to spread "misinformation" (mostly, but not exclusively, heresy) and the mechanical reproduction of texts unsettled existing social relations to an incredible degree, facilitating a staggering amount of human death and suffering. But who among us would give it up, or even say that it has been, on balance, a net negative?
Likewise, the history of the internet is unfinished. With the advent of LLMs, I think it's probably more accurate to say that it's barely begun, and in the end it may dwarf all the other items on Andreesen's list.
The failure of Gutenberg to deploy sufficient Spanish-speaking moderators in 1452 to identify and shut down printing presses outputting flyers urging the ethnic cleansing of the Emirate of Grenada is stark evidence that we need more central government control over mass media!
He invented the modern browser, which is pretty much the modern Internet - it’s certainly what got normies online. He also went against a lot of earlier luminary preferences in how he built it, for example, he resolved a contentious “should we support images” debate by just shipping the img tag and dragging everyone else along.
Yeah, I don’t know how you’d measure it, but the internet seems to have huge productivity value at work, where now you can just fire off an e-mail to a distribution list instead of writing up a memo, getting your secretary to copy it, and then putting it in everyone’s physical inbox
Yep, the internet has been a massive productivity increase for every sort of business and government function. As a social tool though, I think the argument can be made that it's mostly been a failure. Particularly the web 2.0 and beyond form with social media platforms. And yes I realize the irony of stating this on one of said platforms.
Marc Andreesen cannot be stupid, but when you watch his interview with Tyler Cowen about Web3 you begin to wonder if he is. One 'innovation' that is simply bad is Bitcoin, along with other stupid cryptocurrencies. Bitcoin causes massive pollution and waste because of how much electricity is required to 'mine' it. There's also waste in terms of the computers required to mine it, along with waste of engineer talent. Cryptocurrencies are also vector for big scams that cheat people of their money.
I mostly agree with you. It's unfortunate that everyone threw so much money into crypto. However, Ethereum now uses much less energy, I believe, and if I lived in Argentina, I'm sure I would be extremely glad that I can put money in crypto instead of the bank. So even innovations that are mostly a dud still have some benefit.
Using much less energy to do nothing is still not a victory.
And Argentinians mostly put their money in US dollars to avoid inflation, not wildly speculative, unstable cryptocurrencies that are even less reliable than Pesos.
Crypto only is a thing because of quantitative easing. It’s such a joke of a thing. (I mean it facilitates crime, money laundering, and getting cash out of authoritarian regimes.)
The low interest rate environment created opportunities for nonsensical things like crypto, but also stupid border line scams in the stock market. The SPAC craze of 2021 resulted in lots of wasted resources.
Marc Andreesen is certainly not stupid but he is also a VC so it's hard to know when he's just "talking his book" (either consciously or un-consciously). Like does he REALLY think the all the Web3 bullshit was world-changing or does he jus have a huge amount of money invested in Web3 startups? Maybe both!
But, on the plus side, since the publication of that essay, he did help build Bored Ape NFTs, Clubhouse, the newest WeWork grift, and Elon's version of Twitter. It's safe to say that he has used his considerable wealth and influence to construct the most useful things of our times!
If he invested in things that will build, would we know the returns right away in 3.5 years? That seems a little hard to say. Tesla was founded in 2003. The millionth car was sold in March of 2020.
The entire concept of web3 is also stupid. The biggest innovation in the modern Internet has been more and more consolidation by big companies with cloud resources, the opposite of what they want to accomplish with web3.
Can we know which unproven technologies will be good? People once thought cigarettes were pretty harmless. Suppose a Democratically elected body in 1780 or 1820 had debated “should we have an industrial revolution?” Could even the smartest minds have seen the future clearly enough to give a worthwhile answer? Did anyone think that working class wages would increase by over an order of magnitude and life expectancies would more than double? Or were they focused on urban disorder, the risks to the established church, and keeping grain prices high enough for the gentry enjoy the London season.
Upheaval brings countless risks. As the industrial revolution unfolded, armies became and bigger and weapons more lethal. Handloom weavers were reduced to beggary and often starved. The Indian economy collapsed. Urban workers embraced left wing politics and overthrew empires. No responsible moderate would have welcomed these risks.
Status quo bias is strong, and any big change creates enough losers that it will seem big and risky. I have little doubt that the industrial revolution was a good thing, but it’s early years brought misery and progress in almost equal measure. I question whether our species has the guile to know what nascent technologies will be good for us.
was the question “should we have an industrial revolution” ever squarely confronted by a legislature. they asked a series of smaller questions about land enclosure and patents and unions and child labor and the higher level process sort of unfolded so slowly that those who lived through it were only dimly aware.
progress is rarely cost free, and it is easily thwarted when there are too many veto points
Seems to me we need more responsive governance that is less beholden to moneyed interests so that when harms become apparently it doesn’t take a decade or two to act.
If we look at where urban workers embraced left wing politics and overthrew things, it was in non-democracies like China and Russia. In those cases, we're mostly talking about deeply agrarian societies, with a small periphery of radicalized urban labor at the time. By contrast, Marx's prediction of left-wing revolutions in the far more industrial Britain and America never panned out. Electoral politics for reform can resolve most of these problems and most workers voted for an industrial revolution with the their money, in the purchase of mass-manufactured goods.
I'd say it's simply untrue our species lacks the sufficient guile given the British and American experiments in liberal democracy. People deserve more credit. They're not extraordinarily smart or stupid; mass politics can aggregate risk concerns in the form of parties and sort it out. A great deal of industrial development took place as most people wanted it and saw better wages to make. So status quo bias can clearly be overcome. By the 1920s, enormous benefits arrived to middle class households with electrification, indoor plumbing, and automobiles. We were wise enough to see it through then, and judging by McCarthy and Manchin's NEPA reform talks, we could be wise enough now.
Yea, I’m not terribly worried about either regulatory capture or state capacity as issues in the long run. It’s fun to discuss what a perfect world looks like here but we mostly muddle through ok.
We’re always going to struggle to balance the need to constrain law-breaking and externalities with the need to accommodate change and economic/technological advancement, but we do reasonably well.
We’ve done so thus far due to a very flexible and broadly responsive policy apparatus and culture of politics.
That culture, in the fundamentals, remains intact and is mutually reinforcing with our prosperity. Every generation ends up climbing the ladder quickly enough to have a stake and moderate the more absolutist and uncompromising stances of youth.
The effect is to actually make status quo bias work in our favor to an extent because the durable mean to which we revert is a pretty dynamic one.
You should be less pollyannaish about this. The NRC didn't "muddle through". It just killed all technological innovation in nuclear energy for decades and decades.
Nuclear checks all the boxes for the sort of megaprojects that the US has failed at for quite a few decades now. Untangling that knot goes far beyond the NRC, to the very fundaments of how our construction and project delivery ecosystem has come to function.
I’m really quite irritated at Matt for peddling a simple view of the world which is mainly untrue, still. He’s seen enough discussion of the topic here and elsewhere to understand that ALARA is far from the only, or even biggest, problem that nuclear power faces.
I think you made some very convincing arguments about ALARA. However, the broader issue with large construction you point to is significantly impacted by the broader and more dense regulatory thicket that has enveloped most construction projects in the US.
As with everything else, both more and less regulation needed, only a portion of it within the NRC’s remit.
The NRC needs to be able to do shit faster, which basically means breaking it and starting over.
The Vogtle project went off the rails mainly due to coordination failures and gross mismanagement of the critical path operations. Because the reactor pressure vessel design and assembly sequence wasn’t finalized until well into construction, concrete reinforcement was removed, altered, and reinstalled in response to changing internal configurations and plant design.
For nothing, because the reinforcement turned out to be wildly off the critical path due to fabrication and full-depth welding challenges, anyway.
There was a bunch of other shit, but those reconstruction needs and the financing costs incurred from the delays they caused probably account for half the overrun on their own.
We should, at this point, try building 4-6 more of them with the same EPC contractor and see if we start to see any kind of experiential learning, but it’d be “unfair” to other EPC groups and therefore won’t happen.
Whereas the ROK’s program uses the same group of contractors and has a bunch of trades organizations and unions feeding into them alongside university research into fabrication and loaned welding and fabrication experts from their formidable shipbuilding interests.
The more equitably the benefits of new technology and automation are distributed, the more open society will be to it. A strong social safety net to help people cope with the upheaval brought about by technological progress will assuage short term concerns. And regulations should be based on the best data we have about costs and benefits, with ample funding for research to help enlighten us on that front.
Kids, remember your Uncle Barney's number one rule? "New is always better."
Joking aside, while Matt's general thesis is correct, it seems like it should be recognized as a problem that Andreessen felt he needed to make his foolish argument to begin with. Matt has a good point about how most public intellectuals are writers, but one thing he doesn't mention is that most of these said intellectuals also don't like and/or aren't good at math. This probably means they have difficulty quantifying the actual improvements in human life from technological development, and instead depend on anecdotes and vibes.
I think it’s more laziness than innumeracy. While I’m not a public intellectual, I’m a highly verbal person who used to be kind of afraid of math, owing to some unfortunate experiences in middle/high school. But eventually I needed a basic grasp of statistics to understand some things I was reading (and to do my job properly). So I actually applied myself and it wasn’t so difficult! But one can *choose* to rely on anecdotes and vibes, which is much easier and neater than trying to analyze complex issues in detail.
One also has the option of keeping one's damn fool mouth shut when they know they're out of their depth. You'll rarely see me comment here on matters of monetary policy, and that's even under a pseudonym!
Marc Andreessen has that rich successful guy problem where he’s been celebrated for so long and surrounded himself with lickspittles. There’s no one around to tell him when he’s making silly arguments/out of his depth. His appearance on the Sam Harris podcast discussing AI was a tour de force of bluster and empty calories.
This is one reason why I tend to think it's important for authority figures to have to deal with hostile questioning on a regular basis. Most politicians have to do that, but so many rich businesspeople end up running a company built around them and just don't deal with it.
In fact, I think Trump's hostility to hostile questioning and avoidance of any media interaction that wasn't entirely on his own terms was a useful early sign of his personal problems as a politician (as distinct from the problems I have with his political views).
I'm not sure where we're re supposed to be drawing lines about what's a strawman and what isn't here, but I don't think Matt's really succeeding if his goal is actually torching Andreesen's. Matt's take, which is obviously correct to me, basically amounts to, "Even the most successful and productive technologies still produce some negative externalities." Instagram is a negative externality of the Internet. Fentanyl is a negative externality of chemically synthesis of pharmaceuticals. You have to get waaaaayyy lost in the sauce to think that Instagram means the Internet was a bad idea, or that fentanyl OD means synthetic painkillers shouldn't have been invented.
Andreesen is a bit of a schmuck, but on this he's a far better approximation of correct than Paul Ehrlich has ever been about anything.
What do you see as Andreesen’s take? Because he really does have a part that seems to say that any time you’re worrying about risk you’re the enemy, without a clear “but of course there’s fentanyl” qualifier.
Andreesen thinks he is in an existential war against annoying lefty journalists on Twitter. He's not really, but if you have to choose between them, he's obviously on the correct side
Reading over the comments to this post, I think a lot of people are thinking about it this way, that they should defend Andreesen because he has the right enemies or because the alternative is being one of the people on “the other side”--tribalism, basically. I prefer Matt’s approach of judging on the merits.
I also prefer Matt's approach. But it seems so obvious to me I'm not sure it was worth writing? Like I feel like Andreesen has just spent so much time on Twitter it's addled his brain a bit. Obviously slow boring commenters are going to agree with Matt. Not sure there's a great upshot to this takedown post
What I'm saying is that quite a lot of readers seem to be agreeing with Andreesen, or steelmanning him as taking a reasonable perspective when he's not. As long as that's happening, the takedown seems warranted.
If I had to sum it up in one quote from the piece it would be:
> Our enemy is the Precautionary Principle, which would have prevented virtually all progress since man first harnessed fire. The Precautionary Principle was invented to prevent the large-scale deployment of civilian nuclear power, perhaps the most catastrophic mistake in Western society in my lifetime. The Precautionary Principle continues to inflict enormous unnecessary suffering on our world today. It is deeply immoral, and we must jettison it with extreme prejudice.
Right but armed with that absolute he argues against the concept of "risk management" and "trust and safety," too, which I was suggesting to Dave Coffin is a much broader target than just the people that killed nuclear power.
I think he goes way too far for sure, but I can see his point somewhat. If you can't ban something outright you can always nitpick it to death with concern trolling and maximally precautionary regulation. Exhibit A would of course be how NEPA is used to block all manner of development. Is it sometimes true that NEPA blocks genuinely environmentally harmful things from happening? Sure of course, but you might look at NEPA review in the aggregate and conclude that it's a net-negative even if you ONLY care about the environment.
I agree one can! I think Andreesen goes beyond "NEPA is a bad way to assess environmental harms" into "any time you're worried about environmental harms, or harms in general, you're proposing more NEPA." Which basically makes it impossible to reach common cause with people who agree the current system isn't working well but aren't on board with stripping away every guardrail. And those people outnumber the "damn the torpedos" folks for sure, it seems to me. In any event, it seems like we are pretty much in agreement about where he's onto something and where he goes too far.
Below is the part I was referring to, where he really is saying that "risk management" and "tech ethics" are just another way of saying de-growth:
"Our present society has been subjected to a mass demoralization campaign for six decades – against technology and against life – under varying names like “existential risk”, “sustainability”, “ESG”, “Sustainable Development Goals”, “social responsibility”, “stakeholder capitalism”, “Precautionary Principle”, “trust and safety”, “tech ethics”, “risk management”, “de-growth”, “the limits of growth”."
Do you think that paragraph is limited to literal de-growthers and people who oppose all new inventions? Because it seems to me the target is much wider than that and that he is throwing some important babies out with the bathwater.
The core of the fight is about dueling visions of "sustainability". Andreesen's version is about sustaining growth through technological advancement. He understands the other side, with "sustainability" in scare quotes, as being about sustainability through an imposed calcification of some preferred status quo that constrains human society to some "rightful" equilibrium in the natural order. The "enemies" are the people who are promoting that calcified vision of humanity.
Right, I agree that's what he's saying: any time you're worried about what the future and tech progress might bring, you're the enemy. And I agree with Matt that the only way to get there is to just ignore numerous instances where a failure to think about risk caused severe problems.
And it's one thing to say, "We should be responsive to the negative externalities of progress." It's very much another to say, "Maybe we just shouldn't let people do new things, at the very least not without extensive gatekeeping by those vested with power by the status quo."
I think the argument he would make is that a lot of those are used to launder de-growth ideas. I don't really agree and think that you can absolutely worry about sustainability or ESG in a good-faith way, but as way of analogy it's like NIMBY's concern trolling about parking/noise/crime/infrastrcuture/etc. when really they just don't want anything in their neighborhood to change.
Sure, and that's a real thing! He's right that some people are invoking those important concepts in bad faith. But the reason they're doing that is that those really are important concepts, which makes it all the more important to establish when they apply and when they don't, or how to tell the good-faith invocations form the bad-faith ones. He can't just say "let's presume it's all bullshit and go about our business without it."
Andreesen, like Elon, is an edge lord who enjoys the attention of controversy. His "manifesto" was likely written from a list of maximally quotable outrage bait given to an intern (or ChatGPT). Whoever wrote it was lazy enough to largely keep the list structure.
Over/Under 6 months Andreesen becomes the 5th permanent member of the "All In" Podcast.
I also suspect a lot of this is politically calculated. I know people like Josh Hawley and other members of the GOP make a lot of noises about going after Bit Tech, "woke" corporations and social media platforms who supposedly "silence" right wing voices. And I don't think this new shift is entirely cynical (the belief that social media is suppressing right wing voices seems genuinely held by a number of GOP pols). But let's face it, if we're talking about which party at this moment is more likely to pass legislation or put people in charge of Federal agencies that will threaten Big Tech profits, overwhelmingly to this day that's more likely to come from Democrats.
Mostly came here to show appreciation for the Roger Waters and/or Neil Postman shoutout.
However, I think this has been a big problem in the Bay Area/tech world with innocuous descriptors becoming weird cultish subcultures. I'm for altruism being effective, but if I say I'm an Effective Altruist that means something else. I'm optimistic about technology but that doesn't mean I agree with Andreessen. Rationalism has precious little to do with rationality, etc. etc.
This is obviously a common problem for humans organising into groups in general, but the recent SF ones have confused a lot of subjects and resulted in people being opposed to things they should actually be for and vice versa.
A related idea I’ve been toying with for some time is the notion of a sort of “universal context collapse” that naturally happens to every ideology as it grows in popularity.
Ideas that start out good in the abstract eventually leave their original context and get misapplied by their worst fans/followers [ed: who are just too dumb to understand the original idea in its abstract perfection].
Communism would be the canonical example: it starts out as a general call for economic egalitarianism intended to meet the specific crisis facing an industrializing Europe in the 1820s-1840s, but then its followers overindex on a dumb interpretation of historical materialism, and THEN they decide to start taking violent shortcuts, so we end up with the Russian Revolution 70 years later.
Think of it as a derivative of the "death of the author" concept. Basically, once Marx comes up with socialism/communism/historical materialism, the idea leaves his control and takes on a life of its own.
There's a lot of what communists call "ethical socialism" in that "general call for economic egalitarianism" in the 1820s; if you go read the Communist Manifesto, there's a whole section where Marx is denouncing the ethical socialists (and it's definitely Marx; from the rest of their writings, it's his style, not Engels').
Indeed, there would be a (not entirely unreasonable) argument that it is the overindexing that defines communism from other types of socialism.
I feel like this is what happens to any progressive idea or intellectual tool once it gets adopted by Twitter or Tumblr. If a progressive idea gets popular enough on those sites, it gets warped beyond recognition to apply in every context, especially by slacktivists who can only remember five or so intellectual frameworks at a time. Sometimes a hammer should only be a hammer.
I try not to overindex on FdB, but I think he's spot-on about how Tumblr's demise basically took a bunch of self-absorbed introverts who had been safely over-externalizing their neuroses within their own communities for a while, and turned them loose on the rest of the internet (mainly Twitter).
It's like, if BPD was a platform, it would've been Tumblr.
Indeed, I've done that several times now regarding this idea.
Because it basically implies that most good ideas are doomed to fail in some aspect, regardless of their abstract merit. Obviously, the root of it is just that humanity is imperfect - there's no way for us to live in perfect epistemological communion with each other, so we're doomed to keep misunderstanding each other and implementing each other's ideas poorly.
But what's so tantalizing about it is being able to see those aspects of perfection, to understand how they would work, while knowing that they'll never work like that.
Ultimately, though, it all just kind of leads you back to places like SB here. We have literally no choice in the pursuit of progress BUT to resign ourselves to the slow boring of hard boards.
And it's basically the reason why I'm such a discourse-cop around here. People grinding their pet grievances isn't going to solve anything, satisfying as it may be. We just need to keep coming up with ideas, and hoping those ideas filter up from us to gatekeepers like Milan, Kate, and Maya, to throne-whisperers like Matt, and up to the thrones of power; or alternatively, that these ideas gain traction within the broader commentariat and can eventually find some political purchase. It's frustrating and slow, but... it's also just how the universe works.
Postman’s book is *so* good! I reread it last year (I had read it as an undergrad quite some time ago) and was blown away by the salience of his arguments.
"I'm for altruism being effective, but if I say I'm an Effective Altruist that means something else."
I've been banging this drum for a long time. It would be great to have a place to measure which avenues are most helpful for furthering a cause of one's own choice, and not have someone saying that you should be contributing to entirely different causes altogether as opposed from the one you want to contribute to.
Are you just talking about Charity Navigator? Something that looks into which charities actually spend their donations on charitable work?
I don’t mean to restart the EA wars but in general I don’t understand the apparent sense that whether to donate to Cause A or Cause B should not be a valid topic of discussion, debate, and persuasion. Just tune it out if your mind is made up!
"You can’t answer specific questions at such a high level of generality."
This.
If you give me any technological improvement without any other information, then the statistical average is that it will make the world better.
But the worst cases are worse than the best cases are good (the worst cases are existential risks to the species, eg nuclear war).
Which means that "technology, good or bad?" isn't a meaningful or useful question - any given technology could be either, and when we are (collectively) making a decision about a specific technology, we do have information about that technology and we should use it, not pretend we can apply a generic answer and save ourselves looking at the specific case.
>But the worst cases are worse than the best cases are good (the worst cases are existential risks to the species, eg nuclear war).
Nuclear war is not a technology though, it's a feared consequence of nuclear weapons. But nuclear weapons have in general been good for humanity. The long pause in hot great power conflict the world has enjoyed for nearly a century is directly due to them, and the only reason that pause is going to end in the Taiwan Strait is because China and the US consider each other to be extremely unlikely to escalate to nuclear use over such a conflict.
This post sort of hits on something I've touched on before, but I feel like needs to be discussed more; the educational polarization has had a bit of a weird effect on elite level discourse.
Matt noted in this post but it seems like Marc's general turn (including this technology manifesto) last 10 years is driven in part by living in the Bay Area and being surrounded by people who sit on the left side of the political spectrum. So he probably encounters a lot of Boomer left wing "degrowthers' and feels the need to write this manifesto when the number of people out there who hold left wing "degrowth" views happen to be quite small but quite influential in Bay Area (hence left wing San Francisco NIMBY).
This post actually ties together with Matt's interview with Melissa Kearney on Thursday. I actually really enjoyed the interview and found I agreed with most of it (not surprisingly given I subscribe to this Substack). But something bothered me a bit. They discussed a lot the left wing pushback Melissa has gotten and I thought she had reasonably decent response to this. And to lesser extent they discussed some right wing pushback; namely writers who blame the welfare state for being too generous and this somehow destroying marriage. I thought (maybe not surprisingly given my political leanings) they did a good job of how this probably has the opposite effect; being too stingy likely leads to more financial distress which likely leads to less marriage not more (and why expanding EITC is a good idea).
But there was something missing from the conversation; the religious right. Like I don't know how you can have an entire conversation about the "controversy" around the word marriage and not talk extensively about the power of right wing Christianity in this country. Because I know for me when I hear the words "we need to promote marriage", especially from a GOP politician, I'm quite aware that what that often (and I'd say majority of time) means is a) Feminism has gone too far and women need to go back to traditional roles and b) I'm against gay marriage. Like I'm sorry, I truly don't think you can discuss the rise of Trump without discussing the fact he has such an oafish and over the top "masculinity" and is clearly an avatar for a particular view of what a lot of people in this country still view what marriage should be; one where men clearly have the power including legally. Do I need to remind everyone that Dobbs just happened? And while gay marriage continues to garner more and more support in this country, there is still a very large minority who still have lots of influence in the GOP. I know it's been discussed ad nauseum what's really in the these "Don't say gay" bills in FL. And they are deliberately written in a vague way as a sort of "plausible deniability" mechanism. But it seems inarguable to me that they are in part a way to cater to religious right to say "don't worry we're still with you on what marriage is supposed to mean".
This ties together with Andreesen in the sense that I suspect both Kearney and Andreesen are not interacting that much with a lot of traditional conservatives*. Andreesen himself is aligning more with the "PayPal mafia" which is clearly more of an Ayn Rand/libertarian strain of conservatism. And I think it's really having an impact on the discourse more generally.
*In the case of Kearney, I'm basing this on the podcast she had with Matt. I got the sense from the interview that most of the pushback she's gotten over email or publicly is from the Left of the anti-welfare right. Maybe because this was a podcast aimed at Matt's mostly centrist and center left audience/readers. If she has gotten a lot of feedback from the religious right, I'm actually quite curious what they've said to her.
Yes, I think this is correct and I've noticed the same thing. I think it's both about education polarization and (perhaps downstream of that) ideological polarization in the media. If you are an upwardly mobile educated professional (like basically all of my friends...) who lives in a costal city, then you are exposed to a lot of nutty left-wing ideas. But you are not generally not exposed to nutty right-wing ideas because you don't watch Fox News or read Breitbart. So you might know people who will say explicitly they want to abolish the police or decolonize TPS reports but not anyone who wants to ban abortion without exception or thinks that wives should be subservient to their husbands. You may actually know people who believe those things but you exist in a social context in which they know they aren't allowed to say so. You start to think that abolishing the police is a thing that might actually happen and banning abortion will never happen (how could it, you've never met anyone who actually wants to do it!) and start to develop a pretty skewed perspective on what the real risks are.
I don't disagree that the hazard you describe exists, but there are two failure modes for the self-sorted blue American population.
First, the scenario you outlined.
Second, its inverse: where people derive their knowledge of conservative Americans from caricatures created by media outlets that fight for limited attention spans by means of clickbait, hype, and catastrophism.
I don't know how one would ever pin down which one is more skewed, but the second population appears to be much bigger.
I don't really see what you're saying about Melissa Kearney and the religious right. The religious right doesn't send emails to Melissa Kearney because they agree with what she's saying regarding marriage.
You probably agree with this, but the biggest reason to be techno-optimist is that:
1) It's hard to conduct a cost-benefit analysis on things that haven't been invented yet
2) Technological progress is on net good
Often kind hearted-sounding restrictions, like rules around pharma trials, banning human-challenge trials the like, have killed millions of people through preventing discoveries. This means we should be biased towards seeing if harms actually materialize before regulating a whole field away.
MY writes: "The people involved in creating ARPANET seem to disagree about why they did it. Most say it’s a myth that they were trying to create a command-and-control system that would be robust to a Soviet nuclear attack"
Boy does this ever cry out for [citation needed]. Because that is *exactly* why the ARPANET was created. Just read the official DARPA account of its origins: https://www.darpa.mil/attachments/ARPANET_final.pdf
Yeah, I was very surprised to learn that it's a "myth" that ARPANET was intended to be a redundant communications system in the event of nuclear war given that coming up with that kind of thing was DARPA's day-to-day job.
I happened to be in the RAND Corporation classified library one day some years ago and saw some folks from a government security agency processing some old RAND reports. They were going through old classified documents and determining which ones were okay to declassify. I sauntered over and looked at the one on the top of the "declassified" pile they had just stamped as no longer classified. It was a short report by Paul Baran from 1962 in which he proposed the idea of "packet switching" as a way to maintain communications during and after a Soviet nuclear attack. And the rest, as they say, is history.
The internet may not reduce isolation, but it does serve as a matching app for strawmen and people itching to torch them.
The promise of AI: strawmen too cheap to meter.
Meter? I haven't even swiped yet!
Yes, AI will be able to generate infinite sequences of appalling takes and devastating take-downs, dos-i-dos-ing off into infinity like pairs of brooms in a sorcerer's apprentice of bad online discourse.
it already is
When we vote for an MVP of the Slow Boring comment section, this will be the one that puts you over the top. I applaud you, sir.
Aww, I'm blushing. A warm compliment from an enthalpic guy.
The real J. Willard Gibbs would note that the entropic component is more important!
Hold on -- you mean you are *not* the real J. Willard Gibbs?
I hate it when people don't use their real names -- how can you trust anything they say?
I assume that means you're an heir to the Treadmill family fortune?
As I have disclosed before, I come from a long line of Treadmills, or a short line that wraps under in back and then comes up again in the front.
Well said but it only partially addresses a point that Matt missed. Andreesen’s point is taken narrowly if we think of isolation solely as social isolation and, therefore Matt erroneously treats the Internet and social media as one and the same. The Internet is a tremendous advance in communication in terms of accessing and identifying information and connecting people. Many bad consequences, of course, but that is blind to the positive effects.
I think the techno-optimist position is not that all technological progress is always good, but rather it's a bad idea to trust activists and government bureaucrats who generally aren't smart enough to contribute to technological progress to determine which progress is and is not allowed.
Noah Smith has a piece out suggesting America's failure to build things (including new uses of technology) is a result of too *few* bureaucrats with too *little* power:
https://www.noahpinion.blog/p/america-needs-a-bigger-better-bureaucracy
There is a difference between building things, even things that use new technology and the creation of new technology. With that said I am very familiar with the problem Smith addresses and mostly agree. Governments would likely be better off having more full time staff and less consultants. The idea that the consultants are better is a fallacy in itself.
I think it's more that our culture of outsourcing led to us forgetting what consultants can actually DO.
Consultants can tell you what's wrong with your organization with a kind of insight that is often hard to have when you're in the thick of everything.
Consultants can't do the actual job; if they could, they'd be your competitors, not consultants.
Ed: And I think that's the REAL problem with outsourcing everything to The Groups. They may do a lot of good work advocating for various good things, but they AREN'T design firms, they're activists and non-profits. Affordable housing advocates don't know how to design houses because, surprise, they aren't construction workers, architects, or engineers! That's how we end up with stupid ideas like "La Sombrita". Design by a primarily politically-oriented committee is just not good design.
I thought La Sombrita was designed to shame the city, not to actually provide shade.
I mean, it *brought* shame upon the city, but I'm not sure if that was the non-profit's goal.
The most common consultants in local government is technical expertise. Civil engineers for example design roads, surveyors, planners, architects, etc.
It has become the norm to hire consultants to do this work. It could be done in house. It could be that some of it is done in house and some by consultants.
If the government entity has consistent work then having in house professionals doing it (or a part of it) could make more sense.
These consultants are very expensive now. It’s interesting with the rise of HR that these decisions are not more optimized. I have never heard an HR person at my org be curious about how this gets done and if we’re making the best calls. Noah Smiths paper could be a good conversation starter.
Something that folks like Noah don't always recognize, though, is that as expensive as they can be, technical consultants can often be the only reasonable option in rural areas where the need can't justify full-time positions.
The real problem is that in denser jurisdictions, the regulatory burden means that the only people who CAN do the work are expensive consultants.
80 years ago, a disgruntled mid-manager or foreman on a construction crew could grab some buddies and strike out on their own as a contractor. These days, you need an in-house lawyer to navigate zoning laws and safety regs.
Totally right on the rural areas don’t have enough work to justify in house.
In my head, I distinguish between consultants and contractors. That is probably just a social norm. Many midsize governments and beyond could do more in-house professional work (i.e engineering) and they could also build more. Building big projects would be tough. If you’re LA County then you could probably run an enormous construction company in house.
On the design side, single projects can run $500k plus. I think the better place for in house is in doing a collection of the small projects. Then using the consultants on the larger ones. The bigger point is it’s become so commonplace to use consultants that it’s not really being questioned anymore. Pendulums shift and maybe it’s time for this one to shift back?
I am sure a lot of places would push back using a risk based argument but my guess is they don’t have much real data to make that case. They’re just speculating.
>>Then using the consultants on the larger ones.
My only criticism here is that I think it'd be better to encourage an ecosystem of small-to-larger-mid-size contractors that are competent to consult on mid-to-large projects, rather than having an ecosystem of out-and-out consulting firms.
There is also the problem of how you hire good people to go into a bureaucracy that a) is not so good and b) has gotten used to controlling whether changes can happen around it.
(uninformed opinion)
Someone ask Noah to invite a guest bureaucrat to write on bureaucracy topics.
Can we get a TPS report before we green light that?
You seem to be sanewashing Andreessen’s piece. He directly says that the enemies aren’t just bureaucrats, but *anyone* who theorizes about risk, worst-case scenarios, etc.
Andreessen isn’t exactly all that smart himself, though. He makes for a shit standard bearer of his own cause.
Exactly, see my comment above about his totally pathetic output since he issued the clarion call that it was "Time to Build!" in 2020. He then took all his money, networks, and know-down and built over the next three years... Bored Ape NFTs, crypto-video games, the next WeWork, and Elon's version of Twitter. Yay...?
This during perhaps the most urgent time in recent history for building actually-important things.
Amen. The techno-optimists we actually have are not the ones he's wishing we had, and that includes himself.
Take a look at A16Z's actual investments, not just the ones made fun of in the press. For example -
https://app.dealroom.co/investors/andreessen_horowitz/news
Drug delivery, robotic boats, military tech, healthcare tech, restaurant management software, A16Z is really huge, multiple multi million dollar investments per week, and sprawls into all sorts of things. Occasionally there's something funny and/or dumb that the more mainstream press will talk about, like NFTs, but it's a mistake to think that this sort of press is a good summary of A16Z's investment activity.
I'm not denying that a16z has a variety of different investments in its portfolio, but do any of them actually "build" anything important, according to his own definition?
His first argument is that we need to "build our way out of" various serious problems facing humanity, citing past solutions to big problems and basic material needs like the Green Revolution, indoor heating, electric lighting, vaccines, etc. All great things! I agree! Build away!
Some of these things have even been re-invented in the last few years in ways that have really changed the game: mRNA Vaccines, for example. A Nobel Prize-worthy accomplishment! Did a16z invest in those? Not really in his wheelhouse, I guess. (They did write a blog about it, leaving out any mention of the decades of patient, unheralded, government-funded NIH research that made this breakthrough possible). Instead of wasting money on something as high-risk as that, a16z invested in various healthtech apps that haven't revolutionized healthcare in any way, and certainly not commensurate with these aforementioned solutions. One of their top three banner investments in healthcare, PatientPing, isn't really "building" anything, except a kind of ERP/CRM for hospitals. Yay. When I squint, I could see another of their investments--Nautilus Biotechnology and their race to quantify the human proteome--going somewhere, with applications in diagnosis for cancer. But, again, this is no "Moonshot." And it's been quiet for three years since a16z slid their chips over the table.
How about some of the other real challenges of our times? Climate Change? Energy? The Sixth Great Extinction? Food security? Housing? Inequality? Migration? "The Great Stagnation?" Or even to address the more outlandish Existential Risks that Silicon Valley billionaires are really on about these days? Maybe I'm missing something, but I don't see much there in their portfolio. Instead, I see a lot of stuff that offers profitable new B2C and B2B businesses leveraging very incremental innovations that would look familiar to anyone who has followed Silicon Valley for the last 30 years.
Even on areas he has very strong opinions on, like the absolute imperative to (and I quote one of his more famous Tweets from last year) "build 1,000 state of the art nuclear plants in the US and Europe, right now." Has he invested in that? "We won't, but we should." Indeed! Why would he waste his money? It's a terrible business, which is why 1,000 state of the art nuclear plants won't get built in the US and Europe, even if "we should." Instead, a16z invests in conventional things that have a clear line to profitability. That's fine. But it's not the clarion call to the Builders he's trumpeting so grandiosely. And it betrays the fact that what he's really saying here is that he doesn't want the government to regulate him because that might hamper his profits a little bit.
His second argument is that everything good comes from productivity growth, which comes from technological progress. Which is undoubtedly true. It's Econ 101. This is an area that's easier for the likes of a16z to address, surely? Except that the productivity benefit of IT has been surprisingly disappointing since the advent of consumer computing and the Internet. Again, we get stuff like AirBnB, Facebook, Instagram, Pinterest, and Skype, which are certainly successful and useful consumer tech applications. But that haven't, any of them, appreciably increased Total Factor Productivity, which has fallen below the mid-century trend-line since 1973. Non-productivity-enhancing consumer tech brands helping us to mostly "amuse ourselves to death" are BY FAR the largest bucket of investment by a16z over the years, and I don't even need to pick on the likes of Clubhouse or the fallen stars like Buzzfeed or Groupon to make the point that all these billions invested haven't actually increased the productivity of the American economy.
OK, well what about the more serious B2B stuff? Well, though I certainly very much enjoy Slack, is there any evidence that it increased enterprise productivity? Box, Asana, etc... I've used them all. They're not actually "revolutionizing the enterprize," marketing content notwithstanding. You certainly aren't seeing their impact in labor productivity, which has been even more stagnant than TFP, and especially disappointing since the Great Recession. Today, wages are finally up, but in the absence of productivity growth, are just causing "stagflation" fears.
OK, what about the really cool AI/ML stuff like databricks? Miiiiight go somewhere. But that revolution has been nigh for a decade now. I've sold the dream myself for as long. Meanwhile, many of the applications have been gobbled up by the likes of Facebook (Wit.AI) to do what, exactly...? Better ad-targeting!
Again, none of these things are actually game-changing. And maybe that's an unfair criterion. Maybe the Secular Stagnation Hypothesis is right and the truth is that the low-hanging fruit of the caliber that Andreesen alluded to are already invented! Either way, nothing that a16z is betting its money on even has the potential to be that. And I'm sure he knows it.
So the rhetoric sounds really cool and future-y to people who aren't tracking the dollars or the actual outcomes. Much like with Elon Musk's SpaceX, which is really just another DoD contractor for launching military satellites (including proprietary Starlink/Starshield ones) into space, under the veneer of making us a multi-planet species. I don't hate that Elon Musk is building really good rockets and that Mark Andreesen is funding really good SaaS companies. But I wish they would just stop being so vainglorious about it. And that they would both acknowledge how much they actually rely upon the US government for a big chunk of their revenue, the R&D, and of the soft- and hard-infrastructure that make their profits possible.
This is just wrong, even if it makes you feel good to believe it: https://a16z.com/portfolio/
So which among these 670 funded companies (or "Builders") would you say fits Andreesen's own criteria for "building?"
Are any of them providing "the things we urgently need but don’t have"? In housing, education, manufacturing, transportation, or the other areas specifically cited by him in this essay and its precursor three years ago? Do any of them strike you as "aggressive investment in new products, in new industries, in new factories, in new science, in big leaps forward"?
Because I'm familiar with many of these brands, especially the featured ones, and I'm just not seeing it there. "We admire the great technologists and industrialists who came before us, and we aspire to make them proud of us today." Would the inventors of the Green Revolution, electric lighting, indoor heating, vaccines, or any of the other actually-phase-shifting technologies that he cites in his Manifesto be proud of anything in this portfolio of hundreds of startups?
His interview on Sam Harris's podcast was truly terrible.
Yep. Of course it would be ideal to regulate only the amount that is necessary - but the risk too often, is that you get the NRC and totally kill a technology. The counter to that is convincing people to be more pro-technology generally.
I don't think taking an extreme strawman position is the most effective way to convince people to be more pro-technology.
The position that:
1. Technology has been a massive driving force for human improvement.
2. You should be more skeptical of regulation in light of well evidenced historical failures
Is neither extreme nor a strawman.
I'm referring specifically to Andreessen's manifesto, which I would say is a strawman. I definitely agree that technology is one of the main drivers of a better standard of living. As for regulation, it gets complicated. I'm particularly annoyed about US nuclear and environmental regulation, but mainly in terms of poor implementation rather than on principle. A technology like crypto seems perhaps under-regulated given the amounts of money people have been scammed out of or have foolishly traded away.
The only people who've lost money in crypto scams are scammers themselves. If you buy a worthless digital "asset" in the hopes that you're able to turn around and sell it before it inevitably goes to zero, then you're a scammer.
This is such a bizarre take. Sure, YOU know that crypto is a scam, but a lot of people who were scammed don't understand the technology at all. I have an Uncle who was persuaded by scammers that crypto was safe, that it's returns were guaranteed, and that bitcoin offered him an opportunity to get out of the financial troubles he was in. It wasn't his most brilliant decision, obviously, but it was absolutely a scam that preyed on him. Turning around and blaming him for falling for it because of your far superior understanding of it seems callous and completely at odds with the reality of how these things work.
Crypto has been a huge positive productivity shock to ransom takers. Hospitals paying 1m to not get patient records deleted have lost money due to (and ultimately “in”) crypto without being scammers
It’d be simple to think that when you’re smart enough to understand the difference between “investment” and “speculation,” and have heard of “the bigger fool” concept.
Most haven’t, including most people who “invested” in crypto.
I invested all my liquid wealth into NFTs that I researched by scanning a QR code on a sticker affixed to a light pole. The sticker also featured a cute logo of a pixelated teddy bear or some such. Sure the market is down *now* but I’m in it for the long haul.
I think that’s a flattening oversimplification.
There’s a difference between people who came out on top of something they didn’t really understand, and people whose day job is basically to scam those who aren’t lucky enough to come out on top.
I disagree with this analysis for a few reasons.
1. Cryptocurrency has enabled many new scams against people not previously involved in cryptocurrency, such as ransomware.
2. Many people don't really understand cryptocurrency and have gotten roped into this without intent of being scammers.
3. Many people are true believers and have been scammed in incidents like FTX.
Doesn't seem like there is much of a principle (other than "digital") that would separate this definition of scammer from everyone who buys stocks or homes. Unless you really think every crypto customer knew it would "inevitably go[] to zero," which seems very unlikely.
I'm against the government regulating crypto. Doing so would give the government's blessing to something that fundamentally is itself a scam, with no redeeming social value. If people want to blow their money in a speculative frenzy, that's on them.
Great example just from yesterday -
"PM hopes gathering will ensure Britain's leading position in regulating the technology"
Like - does the UK even have the expertise?
https://www.telegraph.co.uk/business/2023/10/22/silicon-valley-bosses-jet-in-ai-summit-boost-sunak/
The NRC actually permits interesting thought experiences. Would it be better to have 1) the current NRC or 2) no NRC? Was there, ex ante, a viable way to establish the NRC that would not have strangled nuclear power?
Error was uniting promotion and safety. You need both in conflict. Even then a cost benefit rule for safety could have been chosen.
It wasn’t JUST that, though. It was the fights over waste disposal, over siting, the general public fearmongering about three headed frogs.
It didn’t help that the clearly best place for waste storage happened to be in the swing state that hosted the senate leader of one of the parties.
They have the ASN in France and it didn't stifle them.
In other words, let the smart people (aka philosophers… sorry tech bros) rule the world? What a terribly original idea!
Thoses dumbasses at the SEC are the dregs of the finance world, nowhere near as smart as bankers and hedge funders--why does it even exist???
I think a big unanswered question sitting between Yglesias and Andreeson is at what point should the government start regulating?
For example, I agree that at some point that AI will probably need some regulation. I don’t think we are at that point yet.
To effectively regulate you need to know the actual downsides and tradeoffs for a technology, not imagined ones. Regulate too early and you crush innovation. Regulate too late, and the new tech can potentially cause real and avoidable harm.
So hire people who /are/ smart enough to be the bureaucrats?
So who is supposed to decide how to manage fentanyl and nuclear power and perhaps next generation AI?
I do think Andreessen has gone for the full caricature of techno-optimism that everything should be unregulated, essentially. There are also open source maximalists out there who don't exactly have nuanced positions.
For AI, it'll be managed by the same entities that regulate spreadsheets, operating systems, and all other software: the market in general and the FTC when antitrust concerns arise.
The market is not a regulator. It rewards maximizing externalization of costs/risks, the exact opposite function of regulators. AI will eventually be very, very good at performing these functions.
The challenge I see is developing the expertise necessary to identify AI models that either by design or by flaw pose outsized risk of external harm. For example, imagine an AI that becomes very, very effective at stock manipulation. It recognizes trading patterns that can induce significant movement and constantly seeks to maximize that activity for profit. That is an unqualified negative to the markets, yet the only market response would be to take the same financial positions as the AI to maximize your profits from its trading. You can extrapolate this to various circumstances. AI will be the perfect amoral profit seeker.
pretty sure they already have that
In my opinion, at whatever point advanced AI entails a high level of risk, that approach becomes insufficient. Of course, figuring out when that happens is difficult.
What is the risk, though?
* Job loss? Every technology causes that, and it's necessary for higher productivity and a richer world. If we care enough, we can expand our welfare state using our newly created wealth.
* Misinformation? Twitter already produces more misinformation in 60 seconds than a human could consume in a lifetime. We even have computer game videos shared as false combat footage.
* Omnipotent being that kills us all? All this harping is extreme speculation that surpasses even the delusions of crypto bros prophesying the collapse of the global financial system.
If real risks arise, then we can introduce regulation to manage them. Yes, we may wish we had the foresight to proactively introduce regulation and avoid the harms. But the impact is just too unknowable. We humans will muddle through, as we have in every past technological disruption from agriculture to the internet, and be far wealthier for it.
The job loss this time will mostly affect those "intellectual circles [which] are dominated by highly verbal people...". I think this is the underlying reason for the extreme skepticism around AI.
It’s contentless, vapid expressions of pique like this that risk turning this comment section into Twitter.
I expected better of you and I’m sad to see how many likes this received just for punching at a disfavored group.
Let’s see… just at the more plausible end of the spectrum:
Highly effective stock market manipulation.
Rapid localization of individuals for any number of illegal purposes.
Fabrication of convincing fraudulent datasets up to falsifying evidence for civil or criminal proceedings.
Generation of false imagery, sound, and video.
But sure, no regulation required for something which would allow foreign state security services to efficiently assassinate a defector hidden by the US government, turn public stock markets into a mockery of efficient capital allocation, completely mis-allocate R&D expenditures, or imprison innocents.
You're arguing that *information networks* are dangerous, not AI. The protocols enable this already, and you aren't putting that genie back in the bottle. The list you have here could have appeared in a breathless Time magazine article about "The World Wide Web" in 2001.
AI risk people usually argue at way too high a level of generality. But giving everyone on earth the skills of a team of computer security researchers and biologists would be incredibly dangerous, for pretty concrete reasons
This seems like a good example of another techno-optimist fallacy, the Safe Uncertainty fallacy: https://www.astralcodexten.com/p/mr-tries-the-safe-uncertainty-fallacy
There is a risk that "job loss" will be though of as a risk!
I think the most-sanewashed version of it is, since tech has outpaced society's ability to deal with failures via politics, we should accelerate innovation in order to make up the difference.
IMO, I don't think that's a good enough reason to stop *trying*.
I think the technologists have some pretty inflated egos. Many of them are very smart, but many bureaucrats are also very smart. This isn't like arguing that no one is smart enough to control a command economy, this guy just thinks he's smarter than everyone else.
Elon Musk has contributed enormously to technological progress, yet he does not strike me as particularly wise. But to be fair, wisdom is not necessarily raw smarts. Yet if one were to evaluate how smart Musk was versus a senior NASA engineer in 2000, I suspect they wouldn't favor him then. Rather, the recent Isaacson biography suggests Musk's success was primarily due to his willingness to put in manic 90 hour weeks and ruthlessly drive his engineers to accomplish something new on a tight deadline. Musk learned a lot about rockets and then drove people to near exhaustion doing things never done before. This pattern was particularly true in enterprises where more vertical integration was feasible, unlike Paypal or Twitter.
Isaacson's book is worth reading because even though it's a bit flowery to gain access, it does a better job demonstrating what was done to produce reusable rockets rather than focusing on the haters. Ultimately, such an obsession with haters misses what produces great technological breakthroughs. Reusable rockets are good. A more helpful question is what motivated Peter Thiel to save it from the brink of bankruptcy in 2008? And what kinds of investments like that today are similarly contingent? The regulatory environment was and is a real issue that demanded constant renegotiation. But it wasn't the main driver of SpaceX. I suspect mRNA vaccines are another great example where the regulatory environment didn't automatically cripple them. But NIH funding politics almost did, if not for the drive of Katalin Kariko to keep bouncing from place to place. It doesn't mean the FDA couldn't use regulatory changes, it absolutely could. It's just not the main story of technology optimism as best I can tell. So it deserves a more proportional analysis in any tech optimism take.
I highly recommend people take a look at my friend Connor's recent essay for more on the mRNA question and hedging our bets in general. I think it approaches this from a better angle; we need to ask where new ideas come from as much and more than than what's conspiring to stop them. https://regions.substack.com/p/how-does-a-society-hedge-its-bets
"Rather, the recent Isaacson biography suggests Musk's success is primarily due to his willingness to put in manic 90 hour weeks and ruthlessly drive his engineers to accomplish something new on a tight deadline."
Well the better question to me then is:
"If the NASA engineers and contractors were so much smarter than Musk, why did they not identify these new things when they probably had decades to do so?"
If the rule is "the people smart enough to build the thing get to decide how to regulate the thing," then we should listen to the people at OpenAI and Anthropic and other leading AI labs, the vast majority of whom are advocating for dramatically increased regulation of AI and are super opposed to what Andreesen is proposing.
These days, Andreesen is a finance guy, not a tech guy. He was a tech guy 30 years ago, but it's not clear to me why that means we should listen to him about AI Governance when the actual cutting-edge tech people are extremely worried and saying we need more AI Governance.
So you only trust the people who have a profit-motive to create new technologies? This line of reasoning would indicate only chemists should be involved in the regulation of pharmaceuticals. Sure, you definitely need to have top chemists involved, but as a rule we ought to include non-technologists for both practical and moral reasons. Andreseen's thesis essentially boils down to "get these g.d. nags out of my way and let the tech bros cook" rather than engaging with the strongest criticisms.
But could you actually build and operate a nuclear power station without government support? How would you insure it?
His anti-government screed in this Manifesto is interesting because the only country on the planet with a dominant nuclear power industry is France, which got there through exactly the public-led private sector that he supposedly hates (when a16z isn't invested in Defense Department-funded tech, anyway).
Sometimes the Mafia actually did provide protection to local businesses. That doesn't mean the Mob is a beneficient entity.
Societies without the Mafia do better than societies with the Mafia. Societies without strong central governments do not do better than societies with strong central governments.
It's the definition of 'strong' that's in question, and how much corruption is present. The Bar association and the American Kennel Club are both non-State organizations, they both are influential in their fields, they both certainly contain some amount of corruption. It's about where the various sliders are set, not about an absolute binary.
Substitute "effective" for "strong" and Tom's argument becomes more persuasive.
One could also ask that about electricity in general (how do we coordinate the paths needed to built transmission wires?) or about the automobile (how do we organize corridors for them to travel freely?) along with a long host of other technologies.
The case that the internet is a beneficial technology shouldn’t be built on loneliness, it should be built on making other science/technology progress faster. The arXiv is legitimately transformative, let alone email, zoom...
Beat me to it. Andreesen's "isolation" framing is absurd, of course. However, is it equally good as the other five things listed alongside it? Perhaps not. But I had a very negative reaction to Matt's "stolen valor" tweet [https://twitter.com/mattyglesias/status/1714047013317210180], when he said that it wasn't even remotely close, when I think the increase in productivity in so many fields can at the very least bring it within remote discussion. I almost pushed back on it in a daily thread last week, but I'm glad I didn't since he wrote this entire article, where it's more appropriate.
It's surprising that neither Andreesen nor Matt brought up the best analogue for the internet in the history of technology, the printing press.
Is anyone willing to make the case against the press? For decades after its invention , there was a great deal of concern about how it was being used to spread "misinformation" (mostly, but not exclusively, heresy) and the mechanical reproduction of texts unsettled existing social relations to an incredible degree, facilitating a staggering amount of human death and suffering. But who among us would give it up, or even say that it has been, on balance, a net negative?
Likewise, the history of the internet is unfinished. With the advent of LLMs, I think it's probably more accurate to say that it's barely begun, and in the end it may dwarf all the other items on Andreesen's list.
The failure of Gutenberg to deploy sufficient Spanish-speaking moderators in 1452 to identify and shut down printing presses outputting flyers urging the ethnic cleansing of the Emirate of Grenada is stark evidence that we need more central government control over mass media!
But Andreessen didn’t invent or make possible the Internet, right? He made possible Facebook.
He invented the modern browser, which is pretty much the modern Internet - it’s certainly what got normies online. He also went against a lot of earlier luminary preferences in how he built it, for example, he resolved a contentious “should we support images” debate by just shipping the img tag and dragging everyone else along.
Yeah, that's fair!
Yeah, I don’t know how you’d measure it, but the internet seems to have huge productivity value at work, where now you can just fire off an e-mail to a distribution list instead of writing up a memo, getting your secretary to copy it, and then putting it in everyone’s physical inbox
And how many marriages have been saved because wives no longer yell at husbands for refusing to ask for directions?
How many marriages were ended by the Ashley Madison data leak?
None that didn’t deserve to be ended!
Ha!
Simply having access to shared data over the Web is huge. Here, have an Excel file....
Yep, the internet has been a massive productivity increase for every sort of business and government function. As a social tool though, I think the argument can be made that it's mostly been a failure. Particularly the web 2.0 and beyond form with social media platforms. And yes I realize the irony of stating this on one of said platforms.
Marc Andreesen cannot be stupid, but when you watch his interview with Tyler Cowen about Web3 you begin to wonder if he is. One 'innovation' that is simply bad is Bitcoin, along with other stupid cryptocurrencies. Bitcoin causes massive pollution and waste because of how much electricity is required to 'mine' it. There's also waste in terms of the computers required to mine it, along with waste of engineer talent. Cryptocurrencies are also vector for big scams that cheat people of their money.
I mostly agree with you. It's unfortunate that everyone threw so much money into crypto. However, Ethereum now uses much less energy, I believe, and if I lived in Argentina, I'm sure I would be extremely glad that I can put money in crypto instead of the bank. So even innovations that are mostly a dud still have some benefit.
Using much less energy to do nothing is still not a victory.
And Argentinians mostly put their money in US dollars to avoid inflation, not wildly speculative, unstable cryptocurrencies that are even less reliable than Pesos.
Crypto only is a thing because of quantitative easing. It’s such a joke of a thing. (I mean it facilitates crime, money laundering, and getting cash out of authoritarian regimes.)
The low interest rate environment created opportunities for nonsensical things like crypto, but also stupid border line scams in the stock market. The SPAC craze of 2021 resulted in lots of wasted resources.
The 2010s were really silly when you think back on it.
Marc Andreesen is certainly not stupid but he is also a VC so it's hard to know when he's just "talking his book" (either consciously or un-consciously). Like does he REALLY think the all the Web3 bullshit was world-changing or does he jus have a huge amount of money invested in Web3 startups? Maybe both!
Three and a half years ago he wrote an essay called “It’s Time to Build.” Has he built anything?
Well, he and his wife certainly proved themselves very much *against* building some stuff (like, say, 131 multi-family units to help address California's extreme housing deficit) in his own affluent community: https://www.theatlantic.com/ideas/archive/2022/08/marc-andreessens-opposition-housing-project-nimby/671061/
But, on the plus side, since the publication of that essay, he did help build Bored Ape NFTs, Clubhouse, the newest WeWork grift, and Elon's version of Twitter. It's safe to say that he has used his considerable wealth and influence to construct the most useful things of our times!
If he invested in things that will build, would we know the returns right away in 3.5 years? That seems a little hard to say. Tesla was founded in 2003. The millionth car was sold in March of 2020.
The entire concept of web3 is also stupid. The biggest innovation in the modern Internet has been more and more consolidation by big companies with cloud resources, the opposite of what they want to accomplish with web3.
Can we know which unproven technologies will be good? People once thought cigarettes were pretty harmless. Suppose a Democratically elected body in 1780 or 1820 had debated “should we have an industrial revolution?” Could even the smartest minds have seen the future clearly enough to give a worthwhile answer? Did anyone think that working class wages would increase by over an order of magnitude and life expectancies would more than double? Or were they focused on urban disorder, the risks to the established church, and keeping grain prices high enough for the gentry enjoy the London season.
Upheaval brings countless risks. As the industrial revolution unfolded, armies became and bigger and weapons more lethal. Handloom weavers were reduced to beggary and often starved. The Indian economy collapsed. Urban workers embraced left wing politics and overthrew empires. No responsible moderate would have welcomed these risks.
Status quo bias is strong, and any big change creates enough losers that it will seem big and risky. I have little doubt that the industrial revolution was a good thing, but it’s early years brought misery and progress in almost equal measure. I question whether our species has the guile to know what nascent technologies will be good for us.
Yes, but there’s no way to actually avoid that question. Saying we can’t answer it just results in someone else answering it.
was the question “should we have an industrial revolution” ever squarely confronted by a legislature. they asked a series of smaller questions about land enclosure and patents and unions and child labor and the higher level process sort of unfolded so slowly that those who lived through it were only dimly aware.
progress is rarely cost free, and it is easily thwarted when there are too many veto points
Seems to me we need more responsive governance that is less beholden to moneyed interests so that when harms become apparently it doesn’t take a decade or two to act.
The more responsive, ubiquitous, and powerful the State is, the more money will be dedicated to controlling it
Thought-provoking. Where does this idea originate?
Empiricism.
As in observation?
If we look at where urban workers embraced left wing politics and overthrew things, it was in non-democracies like China and Russia. In those cases, we're mostly talking about deeply agrarian societies, with a small periphery of radicalized urban labor at the time. By contrast, Marx's prediction of left-wing revolutions in the far more industrial Britain and America never panned out. Electoral politics for reform can resolve most of these problems and most workers voted for an industrial revolution with the their money, in the purchase of mass-manufactured goods.
I'd say it's simply untrue our species lacks the sufficient guile given the British and American experiments in liberal democracy. People deserve more credit. They're not extraordinarily smart or stupid; mass politics can aggregate risk concerns in the form of parties and sort it out. A great deal of industrial development took place as most people wanted it and saw better wages to make. So status quo bias can clearly be overcome. By the 1920s, enormous benefits arrived to middle class households with electrification, indoor plumbing, and automobiles. We were wise enough to see it through then, and judging by McCarthy and Manchin's NEPA reform talks, we could be wise enough now.
Yea, I’m not terribly worried about either regulatory capture or state capacity as issues in the long run. It’s fun to discuss what a perfect world looks like here but we mostly muddle through ok.
We’re always going to struggle to balance the need to constrain law-breaking and externalities with the need to accommodate change and economic/technological advancement, but we do reasonably well.
We’ve done so thus far due to a very flexible and broadly responsive policy apparatus and culture of politics.
That culture, in the fundamentals, remains intact and is mutually reinforcing with our prosperity. Every generation ends up climbing the ladder quickly enough to have a stake and moderate the more absolutist and uncompromising stances of youth.
The effect is to actually make status quo bias work in our favor to an extent because the durable mean to which we revert is a pretty dynamic one.
You should be less pollyannaish about this. The NRC didn't "muddle through". It just killed all technological innovation in nuclear energy for decades and decades.
That’s just not true.
Taps the sign: “ALARA accounts for 15% of costs.”
Nuclear checks all the boxes for the sort of megaprojects that the US has failed at for quite a few decades now. Untangling that knot goes far beyond the NRC, to the very fundaments of how our construction and project delivery ecosystem has come to function.
I’m really quite irritated at Matt for peddling a simple view of the world which is mainly untrue, still. He’s seen enough discussion of the topic here and elsewhere to understand that ALARA is far from the only, or even biggest, problem that nuclear power faces.
I think you made some very convincing arguments about ALARA. However, the broader issue with large construction you point to is significantly impacted by the broader and more dense regulatory thicket that has enveloped most construction projects in the US.
As with everything else, both more and less regulation needed, only a portion of it within the NRC’s remit.
The NRC needs to be able to do shit faster, which basically means breaking it and starting over.
The Vogtle project went off the rails mainly due to coordination failures and gross mismanagement of the critical path operations. Because the reactor pressure vessel design and assembly sequence wasn’t finalized until well into construction, concrete reinforcement was removed, altered, and reinstalled in response to changing internal configurations and plant design.
For nothing, because the reinforcement turned out to be wildly off the critical path due to fabrication and full-depth welding challenges, anyway.
There was a bunch of other shit, but those reconstruction needs and the financing costs incurred from the delays they caused probably account for half the overrun on their own.
We should, at this point, try building 4-6 more of them with the same EPC contractor and see if we start to see any kind of experiential learning, but it’d be “unfair” to other EPC groups and therefore won’t happen.
Whereas the ROK’s program uses the same group of contractors and has a bunch of trades organizations and unions feeding into them alongside university research into fabrication and loaned welding and fabrication experts from their formidable shipbuilding interests.
The more equitably the benefits of new technology and automation are distributed, the more open society will be to it. A strong social safety net to help people cope with the upheaval brought about by technological progress will assuage short term concerns. And regulations should be based on the best data we have about costs and benefits, with ample funding for research to help enlighten us on that front.
one great thing about a support net is it is technology agnostic. it doesn’t care why a worker was displaced, it just helps him
Kids, remember your Uncle Barney's number one rule? "New is always better."
Joking aside, while Matt's general thesis is correct, it seems like it should be recognized as a problem that Andreessen felt he needed to make his foolish argument to begin with. Matt has a good point about how most public intellectuals are writers, but one thing he doesn't mention is that most of these said intellectuals also don't like and/or aren't good at math. This probably means they have difficulty quantifying the actual improvements in human life from technological development, and instead depend on anecdotes and vibes.
I think it’s more laziness than innumeracy. While I’m not a public intellectual, I’m a highly verbal person who used to be kind of afraid of math, owing to some unfortunate experiences in middle/high school. But eventually I needed a basic grasp of statistics to understand some things I was reading (and to do my job properly). So I actually applied myself and it wasn’t so difficult! But one can *choose* to rely on anecdotes and vibes, which is much easier and neater than trying to analyze complex issues in detail.
One also has the option of keeping one's damn fool mouth shut when they know they're out of their depth. You'll rarely see me comment here on matters of monetary policy, and that's even under a pseudonym!
A sadly underused option in The Discourse!
Marc Andreessen has that rich successful guy problem where he’s been celebrated for so long and surrounded himself with lickspittles. There’s no one around to tell him when he’s making silly arguments/out of his depth. His appearance on the Sam Harris podcast discussing AI was a tour de force of bluster and empty calories.
This is one reason why I tend to think it's important for authority figures to have to deal with hostile questioning on a regular basis. Most politicians have to do that, but so many rich businesspeople end up running a company built around them and just don't deal with it.
In fact, I think Trump's hostility to hostile questioning and avoidance of any media interaction that wasn't entirely on his own terms was a useful early sign of his personal problems as a politician (as distinct from the problems I have with his political views).
I'm not sure where we're re supposed to be drawing lines about what's a strawman and what isn't here, but I don't think Matt's really succeeding if his goal is actually torching Andreesen's. Matt's take, which is obviously correct to me, basically amounts to, "Even the most successful and productive technologies still produce some negative externalities." Instagram is a negative externality of the Internet. Fentanyl is a negative externality of chemically synthesis of pharmaceuticals. You have to get waaaaayyy lost in the sauce to think that Instagram means the Internet was a bad idea, or that fentanyl OD means synthetic painkillers shouldn't have been invented.
Andreesen is a bit of a schmuck, but on this he's a far better approximation of correct than Paul Ehrlich has ever been about anything.
What do you see as Andreesen’s take? Because he really does have a part that seems to say that any time you’re worrying about risk you’re the enemy, without a clear “but of course there’s fentanyl” qualifier.
Andreesen thinks he is in an existential war against annoying lefty journalists on Twitter. He's not really, but if you have to choose between them, he's obviously on the correct side
Reading over the comments to this post, I think a lot of people are thinking about it this way, that they should defend Andreesen because he has the right enemies or because the alternative is being one of the people on “the other side”--tribalism, basically. I prefer Matt’s approach of judging on the merits.
I also prefer Matt's approach. But it seems so obvious to me I'm not sure it was worth writing? Like I feel like Andreesen has just spent so much time on Twitter it's addled his brain a bit. Obviously slow boring commenters are going to agree with Matt. Not sure there's a great upshot to this takedown post
What I'm saying is that quite a lot of readers seem to be agreeing with Andreesen, or steelmanning him as taking a reasonable perspective when he's not. As long as that's happening, the takedown seems warranted.
Pointing and laughing is also not an ideology
Say what you will about pointing and laughing, but at least it’s not an ethos?
If I had to sum it up in one quote from the piece it would be:
> Our enemy is the Precautionary Principle, which would have prevented virtually all progress since man first harnessed fire. The Precautionary Principle was invented to prevent the large-scale deployment of civilian nuclear power, perhaps the most catastrophic mistake in Western society in my lifetime. The Precautionary Principle continues to inflict enormous unnecessary suffering on our world today. It is deeply immoral, and we must jettison it with extreme prejudice.
All the rest sort of follows from that.
Right but armed with that absolute he argues against the concept of "risk management" and "trust and safety," too, which I was suggesting to Dave Coffin is a much broader target than just the people that killed nuclear power.
I think he goes way too far for sure, but I can see his point somewhat. If you can't ban something outright you can always nitpick it to death with concern trolling and maximally precautionary regulation. Exhibit A would of course be how NEPA is used to block all manner of development. Is it sometimes true that NEPA blocks genuinely environmentally harmful things from happening? Sure of course, but you might look at NEPA review in the aggregate and conclude that it's a net-negative even if you ONLY care about the environment.
I agree one can! I think Andreesen goes beyond "NEPA is a bad way to assess environmental harms" into "any time you're worried about environmental harms, or harms in general, you're proposing more NEPA." Which basically makes it impossible to reach common cause with people who agree the current system isn't working well but aren't on board with stripping away every guardrail. And those people outnumber the "damn the torpedos" folks for sure, it seems to me. In any event, it seems like we are pretty much in agreement about where he's onto something and where he goes too far.
In a sentence? "Degrowthers are dumb and if you think you can tell people not to invent new stuff you can go fuck yourself."
Below is the part I was referring to, where he really is saying that "risk management" and "tech ethics" are just another way of saying de-growth:
"Our present society has been subjected to a mass demoralization campaign for six decades – against technology and against life – under varying names like “existential risk”, “sustainability”, “ESG”, “Sustainable Development Goals”, “social responsibility”, “stakeholder capitalism”, “Precautionary Principle”, “trust and safety”, “tech ethics”, “risk management”, “de-growth”, “the limits of growth”."
Do you think that paragraph is limited to literal de-growthers and people who oppose all new inventions? Because it seems to me the target is much wider than that and that he is throwing some important babies out with the bathwater.
The core of the fight is about dueling visions of "sustainability". Andreesen's version is about sustaining growth through technological advancement. He understands the other side, with "sustainability" in scare quotes, as being about sustainability through an imposed calcification of some preferred status quo that constrains human society to some "rightful" equilibrium in the natural order. The "enemies" are the people who are promoting that calcified vision of humanity.
Right, I agree that's what he's saying: any time you're worried about what the future and tech progress might bring, you're the enemy. And I agree with Matt that the only way to get there is to just ignore numerous instances where a failure to think about risk caused severe problems.
And it's one thing to say, "We should be responsive to the negative externalities of progress." It's very much another to say, "Maybe we just shouldn't let people do new things, at the very least not without extensive gatekeeping by those vested with power by the status quo."
I think the argument he would make is that a lot of those are used to launder de-growth ideas. I don't really agree and think that you can absolutely worry about sustainability or ESG in a good-faith way, but as way of analogy it's like NIMBY's concern trolling about parking/noise/crime/infrastrcuture/etc. when really they just don't want anything in their neighborhood to change.
Sure, and that's a real thing! He's right that some people are invoking those important concepts in bad faith. But the reason they're doing that is that those really are important concepts, which makes it all the more important to establish when they apply and when they don't, or how to tell the good-faith invocations form the bad-faith ones. He can't just say "let's presume it's all bullshit and go about our business without it."
Out: hippy-punching
In: VC-nerd-punching
and I am HERE FOR IT.
And two legs for kick/stomping!
Andreesen, like Elon, is an edge lord who enjoys the attention of controversy. His "manifesto" was likely written from a list of maximally quotable outrage bait given to an intern (or ChatGPT). Whoever wrote it was lazy enough to largely keep the list structure.
I disagree. From listening to him on different podcasts, I think he's sincerely a techno-optimist and has been for a long time.
Over/Under 6 months Andreesen becomes the 5th permanent member of the "All In" Podcast.
I also suspect a lot of this is politically calculated. I know people like Josh Hawley and other members of the GOP make a lot of noises about going after Bit Tech, "woke" corporations and social media platforms who supposedly "silence" right wing voices. And I don't think this new shift is entirely cynical (the belief that social media is suppressing right wing voices seems genuinely held by a number of GOP pols). But let's face it, if we're talking about which party at this moment is more likely to pass legislation or put people in charge of Federal agencies that will threaten Big Tech profits, overwhelmingly to this day that's more likely to come from Democrats.
Mostly came here to show appreciation for the Roger Waters and/or Neil Postman shoutout.
However, I think this has been a big problem in the Bay Area/tech world with innocuous descriptors becoming weird cultish subcultures. I'm for altruism being effective, but if I say I'm an Effective Altruist that means something else. I'm optimistic about technology but that doesn't mean I agree with Andreessen. Rationalism has precious little to do with rationality, etc. etc.
This is obviously a common problem for humans organising into groups in general, but the recent SF ones have confused a lot of subjects and resulted in people being opposed to things they should actually be for and vice versa.
A related idea I’ve been toying with for some time is the notion of a sort of “universal context collapse” that naturally happens to every ideology as it grows in popularity.
Ideas that start out good in the abstract eventually leave their original context and get misapplied by their worst fans/followers [ed: who are just too dumb to understand the original idea in its abstract perfection].
Communism would be the canonical example: it starts out as a general call for economic egalitarianism intended to meet the specific crisis facing an industrializing Europe in the 1820s-1840s, but then its followers overindex on a dumb interpretation of historical materialism, and THEN they decide to start taking violent shortcuts, so we end up with the Russian Revolution 70 years later.
Think of it as a derivative of the "death of the author" concept. Basically, once Marx comes up with socialism/communism/historical materialism, the idea leaves his control and takes on a life of its own.
There's a lot of what communists call "ethical socialism" in that "general call for economic egalitarianism" in the 1820s; if you go read the Communist Manifesto, there's a whole section where Marx is denouncing the ethical socialists (and it's definitely Marx; from the rest of their writings, it's his style, not Engels').
Indeed, there would be a (not entirely unreasonable) argument that it is the overindexing that defines communism from other types of socialism.
I feel like this is what happens to any progressive idea or intellectual tool once it gets adopted by Twitter or Tumblr. If a progressive idea gets popular enough on those sites, it gets warped beyond recognition to apply in every context, especially by slacktivists who can only remember five or so intellectual frameworks at a time. Sometimes a hammer should only be a hammer.
I try not to overindex on FdB, but I think he's spot-on about how Tumblr's demise basically took a bunch of self-absorbed introverts who had been safely over-externalizing their neuroses within their own communities for a while, and turned them loose on the rest of the internet (mainly Twitter).
It's like, if BPD was a platform, it would've been Tumblr.
I wish this was a Friday comment and not a Monday comment. It makes me want to go take a contemplative walk.
Indeed, I've done that several times now regarding this idea.
Because it basically implies that most good ideas are doomed to fail in some aspect, regardless of their abstract merit. Obviously, the root of it is just that humanity is imperfect - there's no way for us to live in perfect epistemological communion with each other, so we're doomed to keep misunderstanding each other and implementing each other's ideas poorly.
But what's so tantalizing about it is being able to see those aspects of perfection, to understand how they would work, while knowing that they'll never work like that.
Ultimately, though, it all just kind of leads you back to places like SB here. We have literally no choice in the pursuit of progress BUT to resign ourselves to the slow boring of hard boards.
And it's basically the reason why I'm such a discourse-cop around here. People grinding their pet grievances isn't going to solve anything, satisfying as it may be. We just need to keep coming up with ideas, and hoping those ideas filter up from us to gatekeepers like Milan, Kate, and Maya, to throne-whisperers like Matt, and up to the thrones of power; or alternatively, that these ideas gain traction within the broader commentariat and can eventually find some political purchase. It's frustrating and slow, but... it's also just how the universe works.
OK, time to take a break and do some gardening.
It implies that most good ideas are actually “good” ideas and that there is nothing new under the sun.
Not necessarily. It actually implies that even the new ideas have a hard limit on how well they can be implemented.
I don't see how your definition differs from mine.
Postman’s book is *so* good! I reread it last year (I had read it as an undergrad quite some time ago) and was blown away by the salience of his arguments.
It is a little funny to read someone be like "It all started going downhill with that newfangled technology, the telegraph."
According to Plato the Ur sin was the invention of writing.
"I'm for altruism being effective, but if I say I'm an Effective Altruist that means something else."
I've been banging this drum for a long time. It would be great to have a place to measure which avenues are most helpful for furthering a cause of one's own choice, and not have someone saying that you should be contributing to entirely different causes altogether as opposed from the one you want to contribute to.
Are you just talking about Charity Navigator? Something that looks into which charities actually spend their donations on charitable work?
I don’t mean to restart the EA wars but in general I don’t understand the apparent sense that whether to donate to Cause A or Cause B should not be a valid topic of discussion, debate, and persuasion. Just tune it out if your mind is made up!
It's a valid topic of discussion, as long as we all know that we're bringing opinions to the table on what we think is best.
Andreesen’s manifesto puts off real “In This House We Believe” vibes
"You can’t answer specific questions at such a high level of generality."
This.
If you give me any technological improvement without any other information, then the statistical average is that it will make the world better.
But the worst cases are worse than the best cases are good (the worst cases are existential risks to the species, eg nuclear war).
Which means that "technology, good or bad?" isn't a meaningful or useful question - any given technology could be either, and when we are (collectively) making a decision about a specific technology, we do have information about that technology and we should use it, not pretend we can apply a generic answer and save ourselves looking at the specific case.
>But the worst cases are worse than the best cases are good (the worst cases are existential risks to the species, eg nuclear war).
Nuclear war is not a technology though, it's a feared consequence of nuclear weapons. But nuclear weapons have in general been good for humanity. The long pause in hot great power conflict the world has enjoyed for nearly a century is directly due to them, and the only reason that pause is going to end in the Taiwan Strait is because China and the US consider each other to be extremely unlikely to escalate to nuclear use over such a conflict.
This post sort of hits on something I've touched on before, but I feel like needs to be discussed more; the educational polarization has had a bit of a weird effect on elite level discourse.
Matt noted in this post but it seems like Marc's general turn (including this technology manifesto) last 10 years is driven in part by living in the Bay Area and being surrounded by people who sit on the left side of the political spectrum. So he probably encounters a lot of Boomer left wing "degrowthers' and feels the need to write this manifesto when the number of people out there who hold left wing "degrowth" views happen to be quite small but quite influential in Bay Area (hence left wing San Francisco NIMBY).
This post actually ties together with Matt's interview with Melissa Kearney on Thursday. I actually really enjoyed the interview and found I agreed with most of it (not surprisingly given I subscribe to this Substack). But something bothered me a bit. They discussed a lot the left wing pushback Melissa has gotten and I thought she had reasonably decent response to this. And to lesser extent they discussed some right wing pushback; namely writers who blame the welfare state for being too generous and this somehow destroying marriage. I thought (maybe not surprisingly given my political leanings) they did a good job of how this probably has the opposite effect; being too stingy likely leads to more financial distress which likely leads to less marriage not more (and why expanding EITC is a good idea).
But there was something missing from the conversation; the religious right. Like I don't know how you can have an entire conversation about the "controversy" around the word marriage and not talk extensively about the power of right wing Christianity in this country. Because I know for me when I hear the words "we need to promote marriage", especially from a GOP politician, I'm quite aware that what that often (and I'd say majority of time) means is a) Feminism has gone too far and women need to go back to traditional roles and b) I'm against gay marriage. Like I'm sorry, I truly don't think you can discuss the rise of Trump without discussing the fact he has such an oafish and over the top "masculinity" and is clearly an avatar for a particular view of what a lot of people in this country still view what marriage should be; one where men clearly have the power including legally. Do I need to remind everyone that Dobbs just happened? And while gay marriage continues to garner more and more support in this country, there is still a very large minority who still have lots of influence in the GOP. I know it's been discussed ad nauseum what's really in the these "Don't say gay" bills in FL. And they are deliberately written in a vague way as a sort of "plausible deniability" mechanism. But it seems inarguable to me that they are in part a way to cater to religious right to say "don't worry we're still with you on what marriage is supposed to mean".
This ties together with Andreesen in the sense that I suspect both Kearney and Andreesen are not interacting that much with a lot of traditional conservatives*. Andreesen himself is aligning more with the "PayPal mafia" which is clearly more of an Ayn Rand/libertarian strain of conservatism. And I think it's really having an impact on the discourse more generally.
*In the case of Kearney, I'm basing this on the podcast she had with Matt. I got the sense from the interview that most of the pushback she's gotten over email or publicly is from the Left of the anti-welfare right. Maybe because this was a podcast aimed at Matt's mostly centrist and center left audience/readers. If she has gotten a lot of feedback from the religious right, I'm actually quite curious what they've said to her.
Yes, I think this is correct and I've noticed the same thing. I think it's both about education polarization and (perhaps downstream of that) ideological polarization in the media. If you are an upwardly mobile educated professional (like basically all of my friends...) who lives in a costal city, then you are exposed to a lot of nutty left-wing ideas. But you are not generally not exposed to nutty right-wing ideas because you don't watch Fox News or read Breitbart. So you might know people who will say explicitly they want to abolish the police or decolonize TPS reports but not anyone who wants to ban abortion without exception or thinks that wives should be subservient to their husbands. You may actually know people who believe those things but you exist in a social context in which they know they aren't allowed to say so. You start to think that abolishing the police is a thing that might actually happen and banning abortion will never happen (how could it, you've never met anyone who actually wants to do it!) and start to develop a pretty skewed perspective on what the real risks are.
I don't disagree that the hazard you describe exists, but there are two failure modes for the self-sorted blue American population.
First, the scenario you outlined.
Second, its inverse: where people derive their knowledge of conservative Americans from caricatures created by media outlets that fight for limited attention spans by means of clickbait, hype, and catastrophism.
I don't know how one would ever pin down which one is more skewed, but the second population appears to be much bigger.
I don't really see what you're saying about Melissa Kearney and the religious right. The religious right doesn't send emails to Melissa Kearney because they agree with what she's saying regarding marriage.
You probably agree with this, but the biggest reason to be techno-optimist is that:
1) It's hard to conduct a cost-benefit analysis on things that haven't been invented yet
2) Technological progress is on net good
Often kind hearted-sounding restrictions, like rules around pharma trials, banning human-challenge trials the like, have killed millions of people through preventing discoveries. This means we should be biased towards seeing if harms actually materialize before regulating a whole field away.
MY writes: "The people involved in creating ARPANET seem to disagree about why they did it. Most say it’s a myth that they were trying to create a command-and-control system that would be robust to a Soviet nuclear attack"
Boy does this ever cry out for [citation needed]. Because that is *exactly* why the ARPANET was created. Just read the official DARPA account of its origins: https://www.darpa.mil/attachments/ARPANET_final.pdf
Yeah, I was very surprised to learn that it's a "myth" that ARPANET was intended to be a redundant communications system in the event of nuclear war given that coming up with that kind of thing was DARPA's day-to-day job.
An anecdote.
I happened to be in the RAND Corporation classified library one day some years ago and saw some folks from a government security agency processing some old RAND reports. They were going through old classified documents and determining which ones were okay to declassify. I sauntered over and looked at the one on the top of the "declassified" pile they had just stamped as no longer classified. It was a short report by Paul Baran from 1962 in which he proposed the idea of "packet switching" as a way to maintain communications during and after a Soviet nuclear attack. And the rest, as they say, is history.
Yeah, the DoD has pretty specific goals when they fund things ...