Why A.I. might kill us

Plus some ICE chatter, Austin mass transit, and the trouble with vice presidents

Will A.I. kill us all? (Photo by Jeff Robbins and DALL-E 3)

I really want to put two Democrats running U.S. House races in Texas on the radar of anyone interested in pragmatic politics. Bobby Pulido in TX-15 is a Tejano music star running in South Texas, and Johnny Garcia in TX-35 is a sheriff’s deputy running in San Antonio and some of its suburbs.

Both districts have been drawn to be quite red as part of Texas’s new gerrymander, but their redness hinges in critical ways on the idea that heavily Hispanic areas that traditionally voted for Democrats will retain their Trump-era allegiance to the G.O.P.

Both candidates believe that this is wrong, that there is a backlash to Trump along with a willingness on the part of Hispanic swing voters to consider voting for old-school, common-sense Democrats in the mold of Vicente Gonzalez and Henry Cuellar.

The big obstacle is that both Pulido and Garcia need to win primaries in March, but they are not the kind of candidates who are fashionable right now with either online small donors or the very climate-pilled Dem mega-donors. That said, I think they both have a decent chance of winning if they can raise money, and if they were to win, that would help reset the narrative in the national party in terms of what working-class communities of color are actually looking for in a congressional representative.

Speaking of party building, the Majority Democrats asked me to extend an invitation to their upcoming D.C. happy hour on Thursday, January 22 to Slow Boring readers. I cannot make it that day, but I will be at future Majority Democrats happy hours, and I think this is a great networking opportunity for people who are interested in the issues and ideas we write about in this newsletter and want to meet other broadly like-minded people.

---

Siddhartha Roychowdhury: I think Renee Good’s shooting was completely unnecessary and avoidable. I also think ICE has a legal right to detain and remove illegal immigrants. Do you think it’s a good idea for activists to physically insert themselves in situations to prevent the ICE agents from doing their job?

Alison: First ever mailbag question! You’ve previously suggested that young progressive-minded people concerned about police reform should consider becoming police officers themselves, both to improve policing from within and to develop realistic, credible perspectives on criminal justice reform based on real experience (something I completely agree with). Would you say the same about ICE, or have they crossed a threshold where this kind of strategy could make a difference? I’d be interested in other thoughts on immigration enforcement reform as well—it’s starting to seem to me like the next Democratic president really will need to abolish ICE (something I scoffed at 5 years ago).

Michael Adelman: There has been a strikingly authoritarian propaganda campaign by both the Trump Administration and right-aligned media in the wake of the shooting of Renee Good. Their very clear implication is that if you’re someone who doesn’t appeal to conservative sensibilities, they can shoot you in the street, make a bunch of allegations against you like domestic terrorism, and get away with it. I feel like I have Brian Beutler on one shoulder telling me that liberals need to forcefully contest this kind of frankly eliminationist rhetoric, and Matt Yglesias on the other shoulder reminding me that on any controversy pitting law enforcement against a lib-coded victim, the public will break overwhelmingly on the pro-cop side, especially when the controversy is immigration-adjacent. And they both have good points! What’s the right balance between directly contesting hard-core authoritarianism while also trying to win in November (which, inter alia, is ALSO critical to contesting hard-core authoritarianism)? What would you tell a Blue Dog House member to do here?

I’m going to address the Minnesota/ICE situation in a post next week, so I’m not going to say too much about it in the mailbag context.

Broadly, though, I’ve decided that I want to try to do a better job than I did with the Abrego Garcia case of practicing what I preach in terms of issue salience. All things considered, I think it is better for Trump if the 2026 midterms are a referendum on immigration than if they are a referendum on health care or the economy. That doesn’t mean that most voters think Good’s killing was justified (they don’t, and it wasn’t) or that the general issue of Trump’s overreach on immigration isn’t important on substance (it is) or that the world needs me to scold other people for talking too much about this (that only encourages them to talk even more about it).

But at the margin, I’m not worried that the issue is under-discussed or under-covered. Plenty of people are talking about it, and public approval of Trump’s handling of immigration has plummeted.

The cautionary note I would flag for Democrats is that, in two-party politics, Trump doesn’t need to outrun the bear — he just needs to outrun the Democrats. Even with his numbers way down, this is an area where Republicans are significantly more trusted and Democrats should think about how they can try to rebuild trust. They should also consider that they had a strong political hand on immigration by November of 2020. The reason they ended up with a huge trust deficit is that the Biden administration blundered in ways that voters find hard to forgive. It’s important to think about the substantive public-policy question of how to avoid blundering.

Alexander McCoy: Why do some people think that AI could be an “existential risk” to humanity?

I can totally imagine a badly aligned causing the economy to crash, or making the internet unusable, and other things that would be bad - but the only straight answer I’ve been able to find about why AI would actually be an existential risk seems to be the assumption that militaries would give AI control of nuclear weapons. To me, this seems 1) simple and easy to avoid, just don’t give the AI nukes, and 2) more of an argument that nukes are the existential risk, not AI.

What’s your take?

I believe the most intelligent non-human animals on the planet are chimpanzees, dolphins, and elephants, all of which have been pushed to the brink of near-extinction by human activity. In most cases, this activity is not even particularly malicious or designed to harm the dolphins. It’s just that humans have our own reasons for doing things, and these things that we do end up having negative downstream consequences for other species — species that, though intelligent in the scheme of things, simply do not have the ability to fight back effectively.