There's a structural similarity between this issue and (believe it or not) moving centrist on Democratic branding. In both cases, Matt is calling the bluff of people who claim to be scared of something.
Are you scared of climate change? Really, really scared? Do you think it may be an extinction-level event for human beings? Then you should be willing to try a lot of things to avoid it -- even giving up some of your previous lefty/liberal stances on nuclear.
Are you scared of a second Trump presidency? Really, really scared? Do you think it may be an extinction-level event for American democracy? Then you should be willing to try a lot of other things to avoid it -- even giving up some of your previous lefty/liberal stances on [uncompromising leftism, a variety of social issues, you name it.]
" I'll do anything to stop Trump/climate change, but I won't compromise on historical preservation!" is a pretty good way to show that we should not take your predictions of catastrophe seriously.
Thanks for this, Matt. As a trained and experienced nuclear reactor operator (Navy propulsion) I can tell anyone that reactors are safe, reliable, and can be built. The Navy gets around the NRC by using a different organization: Naval Reactors. NR is very stringent but they also understand that reactors must be built. So we've had new classes of submarines, aircraft carriers, approved in timely fashion and relied on for national security for decades. Not a single nuclear incident. They are refueled every 25 years. Compare to my other ship while serving, which ran on jet fuel and had to refuel every other day. Nuclear power is an amazing achievement of engineering and science, and I wish that more people were doing cost-benefits here instead of knee-jerking and feeling proud of themselves because they "failed conservative".
This passage really alarmed me: "And what gives people in the industry heartburn is that the NRC has, in fact, never approved a nuclear reactor from start to finish since its creation in 1975."
Human knowledge needs to be continually passed on, from person to person, in order to continue to be useful and applied in the world. That includes knowledge of how to regulate the construction and operation of a new power plant. The longer we go without successfully building a new power plant, the more old timers working in these agencies are going to retire or die, and then we won't have anyone around who's actually done the thing that these agencies are supposed to know how to do. We need to get on it!
The fact that so many parts of our government are unable or unwilling to fulfill their stated functions is a major problem for the progressive movement.
I’m not totally sure what can be done. Could a focused executive team reform these departments? I’d like to think so, but Obama didn’t and Biden seems disinterested. Meanwhile, passport processing is up to six months.
I had an internship at Idaho National Lab and although it was a nice time it was clear to me I didn't want to work in the nuclear industry because it seemed to me like it was on the decline and there might not be a real future in it. Not to mention there were people ten years older than me who couldn't point to a single thing that they'd concretely done because even the smallest design changes can take 15 years - I worked on usability so we're not talking about new reactors, we're talking better knobs and switches.
By the way, most of the nuclear plants in the US use the original control room designs from 60 - 70 years ago, so instead of big monitors showing temperatures you still have moving needles printing on rolls of paper like old lie detector tests. But getting instrumentation updates through the NRC was just incredibly difficult.
The main reason people finally got serious about updating is that they literally can't buy the old equipment any more, analog instrumentation like that literally isn't used in anything else.
Articles like this make me think that (unfortunately) I was right to take a different path (the medical field, where I deal with the other regulator that this Substack likes to bash).
There is a Law of Bureaucracies that flows from the old saying "Victory has a thousand fathers; defeat is an orphan".
If the NRC approves a new design that goes on to be a success, that's a great job by some faceless bureaucrats we will never hear about.
But if the new design goes BOOM, so do the careers of the formerly faceless.
The incentive of any sane careerist is to say "no" outright, or to raise more questions. "Yes" can only end badly. (This goes on in the credit review department at banks, as a private sector example.)
Understanding that dynamic is the key to any kind of reform or oversight. As an offhand example, it is an argument *in favor* of the infamous revolving door - a regulator who says 'yes' occasionally can cash in by going over to the other side. (Obviously, that creates other undesirable incentives...)
I'm glad to see this topic covered, but I'd love a bit more detail on the specifics with respect to problems with the NRC. Matt points out bad communications and a lack of well-defined process, but I'd love to know in this specific case whether or not the NRC was being unreasonable.
After the CDC and FDA's abysmal performance during the pandemic, I'm more convinced than ever that many of our government bureaucracies aren't fit for purpose, but it'd be good to know what the NRC's response is to the assertion that they don't communicate well, don't have a well defined process, and don't weigh the overall lives saved from a project. If not an NRC response, even a third party assessment that got into some specifics would be good.
Nuclear power advocates un the US always bring up the NRC as a big part of the reason nuclear can't succeed, but they can't point to anywhere else in the world where nuclear power is succeeding. Even France can't seem to build the darn things anymore:
Matt’s concerns about the footprint of solar power are overblown. It would take 22,000 square miles of solar to generate electricity equal to current American consumption. That’s less than 1% of the continental US. Meanwhile 21% of America is shrubland and an additional 17% is grass and pasture land. Generating solar in the northeast might require knocking down forests, but there are lots of lightly used, uninhabited forests in Maine, New Brunswick and Quebec. I’m not talking about virgin forests. I’m talking ugly, second growth forests that have been harvested every 50-70 years since the early 19th century.
No one’s beloved urban or suburban forest has to die to throw up solar panels. Those who suggest otherwise are either naive or really cynical.
Hard not to draw a through-line between the NRC and the FDA and CDC's slowness and inactivity-bias during the pandemic. And at least somewhat hard for me to believe that they're the only agencies that are like this.
How do you systematically address the tenancy for regulators to become slower and slower, more and more risk averse (to the risks of activity, not inactivity) over time? In private enterprise, there concept is that if you become too hidebound, someone else eats your lunch, but obviously not a problem for regulators.
This is just a microcosm of most (all?) US regulation: the lack cost-benefit analysis. Now it's find for Congress to oversee what parameters an agency uses. What is the value of a ton of CO2 emission avoided? What's the value of a life lost in a nuclear accident? What's the value of a vaccine that saves a life (through externalities and directly) earlier rather than later? And GAO could make sure that agencies are conducting the analyses.
Great points about the NRC that made me hear resonance with various complaints I’ve had about the FDA the last two years: they’re 100% ok with the risks of doing nothing since this ensures they won’t be blamed for anything, even if their inaction costs lives.
Politically, what can Biden do? Since nuclear is still taboo for many on the left it doesn’t seem to be obviously a winning issue, unless the tortoise- and bird-lovers decide they like next-gen nuclear more than solar and wind. Thoughts?
Is there some reason this needs to be built in the US? Like...there are other countries out there. Whatever happened to regulatory arbitrage and shopping jurisdictions and all that? Oklo KNEW that the NRC has never approved an application but they still decided to set up shop in the US. I don't mean that decision was necessarily irrational but it is all a blackbox to me and I'd like to see a journalist unpack it.
Like, why can't Oklo go to Nigeria and convince them to built a testbed reactor? Or Malaysia? Or Saudi Arabia? Or Pakistan? Or Israel? (Or, as Noah Smith suggests, why not Jamaica?)
The agency that this most resembles is the part of the FDA that regulates medical devices, because both nuclear reactors and medical devices are engineered systems. In the case of the Center for Devices and Radiological Health, one of the requirements is that companies study their product in order to collect the evidence that will characterize safety and effectiveness (e.g., the clinical trial). So there is basically a two-gate process: First the FDA can evaluate the company's strategy for collecting safety and effectiveness evidence. If the study is approved, then the company can build devices and run the clinical trial. This is also where the company gets experience scaling up its manufacturing and putting all the processes in place to capture safety data, etc. Then when the study is done, they submit the FDA for full market approval. So the FDA gets another chance to evaluate before full commercial release. This seems like a better fit for these experimental reactor designs. There needs to be a way to set them up on a small scale, study their performance, get the safety protocols in place, work out bugs in manufacturing scale-up, and then submitting for full commercial release based on the evidence collected to that point. Even at the FDA, this is not a perfect system. But for a safety-based regulatory body, having a two-stage process of evaluation is critical, particularly for engineered systems where modifications can be made (in contrast to drugs and biologics, which can't be "fixed" if the safety profile is bad).
I looked up the letter from the NRC when they rejected Oklo’s application, and it reads a lot worse for Oklo than I expected. I’m not entirely sure if the founders are being as forthright as they might seem when they say that they didn’t really understand what the NCR needs.
By letter dated November 17, 2020 (ADAMS Accession No. ML20308A677), the NRC staff informed Oklo that the RAI responses, audit documents, and audit discussions enhanced the NRC staff’s understanding of Oklo’s novel approach to the Aurora design but did not provide sufficient information to define the scope of the full technical review of the custom combined license application. This letter informed Oklo that (1) resolution on several aspects of the MCA was needed, (2) resolution on several aspects of classification of SSCs was needed, and (3) the topic of quality assurance was being tracked as part of the safety classification of SSCs rather than as a separate issue. The letter also informed Oklo that the topic of applicability of regulations was closed6 and that Step 1 of the review was being extended to allow time for Oklo to address the topics of MCA and safety classification of SSCs.
On December 2, 2020, during a routine scheduling call, Oklo requested that the NRC staff temporarily pause its review and stop developing additional RAIs for the Aurora custom combined license application; Oklo confirmed its request in a follow-up email dated
December 3, 2020 (ADAMS Accession No. ML20338A510). By letter dated December 21, 2020 (ADAMS Accession No. ML20357A001), Oklo informed the NRC staff that it was reviewing the specific items outlined by the NRC staff in the Step 1 extension letter for the topics of MCA and classification of SSCs, and that it would propose next steps for the Step 1 review. Discussions with Oklo on the next steps for the review took place with NRC management in early 2021. Ultimately, Oklo decided to submit generic topical reports to address the topics of MCA and safety classification of SSCs for the Step 1 review of these topics, including the specific questions in the RAIs.
Topical reports for the MCA methodology and safety classification of SSCs did not resolve the open Step 1 issues
By letter dated July 2, 2021 (ADAMS Accession No. ML21184A001), Oklo submitted two topical reports that contained insufficient technical information to address the open Step 1 issues for NRC staff review. The first, “Maximum Credible Accident Methodology,” Revision 27 (ADAMS Accession No. ML21184A002), described Oklo’s approach to the MCA analysis. The second, “Performance Based Licensing Methodology,” Revision 0 (ADAMS Accession
No. ML21187A001), attempted to describe, in part, Oklo’s process for safety classification of SSCs. The NRC staff performed completeness reviews of the topical reports using the Office of Nuclear Reactor Regulation’s Office Instruction LIC-500, “Topical Report Process” (ADAMS Accession No. ML19123A252). The staff determined that neither topical report contained sufficient information to initiate detailed technical reviews. Each report contained conceptual information, rather than repeatable methodologies, and each left many issues unresolved and open for future potential applicants referencing the topical reports to address. The NRC staff informed Oklo of the insufficiency of the topical reports by two emails dated August 5, 2021 (ADAMS Accession Nos. ML21201A079 and ML21201A111), that included attachments describing in detail the supplemental information Oklo must provide for the NRC staff to begin the detailed review of each topical report (NRC Forms 898 – ADAMS Accession
Nos. ML21201A094 and ML21201A113). The NRC staff identified five areas where additional information was needed for the MCA methodology and three areas where additional information was needed for the PBLM methodology. The NRC staff held public meetings with Oklo on September 1, 16, and 28, 2021 (meeting summaries available at ADAMS Accession
Nos. ML21259A260, ML21266A428, and ML21293A329, respectively). During these meetings NRC staff responded to Oklo’s requests for clarification on the information needed to address.
There's a structural similarity between this issue and (believe it or not) moving centrist on Democratic branding. In both cases, Matt is calling the bluff of people who claim to be scared of something.
Are you scared of climate change? Really, really scared? Do you think it may be an extinction-level event for human beings? Then you should be willing to try a lot of things to avoid it -- even giving up some of your previous lefty/liberal stances on nuclear.
Are you scared of a second Trump presidency? Really, really scared? Do you think it may be an extinction-level event for American democracy? Then you should be willing to try a lot of other things to avoid it -- even giving up some of your previous lefty/liberal stances on [uncompromising leftism, a variety of social issues, you name it.]
" I'll do anything to stop Trump/climate change, but I won't compromise on historical preservation!" is a pretty good way to show that we should not take your predictions of catastrophe seriously.
Thanks for this, Matt. As a trained and experienced nuclear reactor operator (Navy propulsion) I can tell anyone that reactors are safe, reliable, and can be built. The Navy gets around the NRC by using a different organization: Naval Reactors. NR is very stringent but they also understand that reactors must be built. So we've had new classes of submarines, aircraft carriers, approved in timely fashion and relied on for national security for decades. Not a single nuclear incident. They are refueled every 25 years. Compare to my other ship while serving, which ran on jet fuel and had to refuel every other day. Nuclear power is an amazing achievement of engineering and science, and I wish that more people were doing cost-benefits here instead of knee-jerking and feeling proud of themselves because they "failed conservative".
This passage really alarmed me: "And what gives people in the industry heartburn is that the NRC has, in fact, never approved a nuclear reactor from start to finish since its creation in 1975."
Human knowledge needs to be continually passed on, from person to person, in order to continue to be useful and applied in the world. That includes knowledge of how to regulate the construction and operation of a new power plant. The longer we go without successfully building a new power plant, the more old timers working in these agencies are going to retire or die, and then we won't have anyone around who's actually done the thing that these agencies are supposed to know how to do. We need to get on it!
The fact that so many parts of our government are unable or unwilling to fulfill their stated functions is a major problem for the progressive movement.
I’m not totally sure what can be done. Could a focused executive team reform these departments? I’d like to think so, but Obama didn’t and Biden seems disinterested. Meanwhile, passport processing is up to six months.
I had an internship at Idaho National Lab and although it was a nice time it was clear to me I didn't want to work in the nuclear industry because it seemed to me like it was on the decline and there might not be a real future in it. Not to mention there were people ten years older than me who couldn't point to a single thing that they'd concretely done because even the smallest design changes can take 15 years - I worked on usability so we're not talking about new reactors, we're talking better knobs and switches.
By the way, most of the nuclear plants in the US use the original control room designs from 60 - 70 years ago, so instead of big monitors showing temperatures you still have moving needles printing on rolls of paper like old lie detector tests. But getting instrumentation updates through the NRC was just incredibly difficult.
The main reason people finally got serious about updating is that they literally can't buy the old equipment any more, analog instrumentation like that literally isn't used in anything else.
Articles like this make me think that (unfortunately) I was right to take a different path (the medical field, where I deal with the other regulator that this Substack likes to bash).
There is a Law of Bureaucracies that flows from the old saying "Victory has a thousand fathers; defeat is an orphan".
If the NRC approves a new design that goes on to be a success, that's a great job by some faceless bureaucrats we will never hear about.
But if the new design goes BOOM, so do the careers of the formerly faceless.
The incentive of any sane careerist is to say "no" outright, or to raise more questions. "Yes" can only end badly. (This goes on in the credit review department at banks, as a private sector example.)
Understanding that dynamic is the key to any kind of reform or oversight. As an offhand example, it is an argument *in favor* of the infamous revolving door - a regulator who says 'yes' occasionally can cash in by going over to the other side. (Obviously, that creates other undesirable incentives...)
I'm glad to see this topic covered, but I'd love a bit more detail on the specifics with respect to problems with the NRC. Matt points out bad communications and a lack of well-defined process, but I'd love to know in this specific case whether or not the NRC was being unreasonable.
After the CDC and FDA's abysmal performance during the pandemic, I'm more convinced than ever that many of our government bureaucracies aren't fit for purpose, but it'd be good to know what the NRC's response is to the assertion that they don't communicate well, don't have a well defined process, and don't weigh the overall lives saved from a project. If not an NRC response, even a third party assessment that got into some specifics would be good.
Nuclear power advocates un the US always bring up the NRC as a big part of the reason nuclear can't succeed, but they can't point to anywhere else in the world where nuclear power is succeeding. Even France can't seem to build the darn things anymore:
https://www.reuters.com/business/energy/edf-announces-new-delay-higher-costs-flamanville-3-reactor-2022-01-12
Matt’s concerns about the footprint of solar power are overblown. It would take 22,000 square miles of solar to generate electricity equal to current American consumption. That’s less than 1% of the continental US. Meanwhile 21% of America is shrubland and an additional 17% is grass and pasture land. Generating solar in the northeast might require knocking down forests, but there are lots of lightly used, uninhabited forests in Maine, New Brunswick and Quebec. I’m not talking about virgin forests. I’m talking ugly, second growth forests that have been harvested every 50-70 years since the early 19th century.
No one’s beloved urban or suburban forest has to die to throw up solar panels. Those who suggest otherwise are either naive or really cynical.
Hard not to draw a through-line between the NRC and the FDA and CDC's slowness and inactivity-bias during the pandemic. And at least somewhat hard for me to believe that they're the only agencies that are like this.
How do you systematically address the tenancy for regulators to become slower and slower, more and more risk averse (to the risks of activity, not inactivity) over time? In private enterprise, there concept is that if you become too hidebound, someone else eats your lunch, but obviously not a problem for regulators.
"mustache-twirling fossil fuel barons"
All things considered, the NIMBY/environmentalists are likely to be actually "mustache-twirling" these days.
This is just a microcosm of most (all?) US regulation: the lack cost-benefit analysis. Now it's find for Congress to oversee what parameters an agency uses. What is the value of a ton of CO2 emission avoided? What's the value of a life lost in a nuclear accident? What's the value of a vaccine that saves a life (through externalities and directly) earlier rather than later? And GAO could make sure that agencies are conducting the analyses.
Great points about the NRC that made me hear resonance with various complaints I’ve had about the FDA the last two years: they’re 100% ok with the risks of doing nothing since this ensures they won’t be blamed for anything, even if their inaction costs lives.
Politically, what can Biden do? Since nuclear is still taboo for many on the left it doesn’t seem to be obviously a winning issue, unless the tortoise- and bird-lovers decide they like next-gen nuclear more than solar and wind. Thoughts?
Is there some reason this needs to be built in the US? Like...there are other countries out there. Whatever happened to regulatory arbitrage and shopping jurisdictions and all that? Oklo KNEW that the NRC has never approved an application but they still decided to set up shop in the US. I don't mean that decision was necessarily irrational but it is all a blackbox to me and I'd like to see a journalist unpack it.
Like, why can't Oklo go to Nigeria and convince them to built a testbed reactor? Or Malaysia? Or Saudi Arabia? Or Pakistan? Or Israel? (Or, as Noah Smith suggests, why not Jamaica?)
The agency that this most resembles is the part of the FDA that regulates medical devices, because both nuclear reactors and medical devices are engineered systems. In the case of the Center for Devices and Radiological Health, one of the requirements is that companies study their product in order to collect the evidence that will characterize safety and effectiveness (e.g., the clinical trial). So there is basically a two-gate process: First the FDA can evaluate the company's strategy for collecting safety and effectiveness evidence. If the study is approved, then the company can build devices and run the clinical trial. This is also where the company gets experience scaling up its manufacturing and putting all the processes in place to capture safety data, etc. Then when the study is done, they submit the FDA for full market approval. So the FDA gets another chance to evaluate before full commercial release. This seems like a better fit for these experimental reactor designs. There needs to be a way to set them up on a small scale, study their performance, get the safety protocols in place, work out bugs in manufacturing scale-up, and then submitting for full commercial release based on the evidence collected to that point. Even at the FDA, this is not a perfect system. But for a safety-based regulatory body, having a two-stage process of evaluation is critical, particularly for engineered systems where modifications can be made (in contrast to drugs and biologics, which can't be "fixed" if the safety profile is bad).
World capacity of offshore wind now: 34GW.
Amount of new offshore wind announced by the Scottish Government today: 25GW
https://www.gov.scot/news/offshore-wind-development/
I looked up the letter from the NRC when they rejected Oklo’s application, and it reads a lot worse for Oklo than I expected. I’m not entirely sure if the founders are being as forthright as they might seem when they say that they didn’t really understand what the NCR needs.
https://www.nrc.gov/docs/ML2135/ML21357A034.pdf
(Here are some relevant paragraphs)
—————
By letter dated November 17, 2020 (ADAMS Accession No. ML20308A677), the NRC staff informed Oklo that the RAI responses, audit documents, and audit discussions enhanced the NRC staff’s understanding of Oklo’s novel approach to the Aurora design but did not provide sufficient information to define the scope of the full technical review of the custom combined license application. This letter informed Oklo that (1) resolution on several aspects of the MCA was needed, (2) resolution on several aspects of classification of SSCs was needed, and (3) the topic of quality assurance was being tracked as part of the safety classification of SSCs rather than as a separate issue. The letter also informed Oklo that the topic of applicability of regulations was closed6 and that Step 1 of the review was being extended to allow time for Oklo to address the topics of MCA and safety classification of SSCs.
On December 2, 2020, during a routine scheduling call, Oklo requested that the NRC staff temporarily pause its review and stop developing additional RAIs for the Aurora custom combined license application; Oklo confirmed its request in a follow-up email dated
December 3, 2020 (ADAMS Accession No. ML20338A510). By letter dated December 21, 2020 (ADAMS Accession No. ML20357A001), Oklo informed the NRC staff that it was reviewing the specific items outlined by the NRC staff in the Step 1 extension letter for the topics of MCA and classification of SSCs, and that it would propose next steps for the Step 1 review. Discussions with Oklo on the next steps for the review took place with NRC management in early 2021. Ultimately, Oklo decided to submit generic topical reports to address the topics of MCA and safety classification of SSCs for the Step 1 review of these topics, including the specific questions in the RAIs.
Topical reports for the MCA methodology and safety classification of SSCs did not resolve the open Step 1 issues
By letter dated July 2, 2021 (ADAMS Accession No. ML21184A001), Oklo submitted two topical reports that contained insufficient technical information to address the open Step 1 issues for NRC staff review. The first, “Maximum Credible Accident Methodology,” Revision 27 (ADAMS Accession No. ML21184A002), described Oklo’s approach to the MCA analysis. The second, “Performance Based Licensing Methodology,” Revision 0 (ADAMS Accession
No. ML21187A001), attempted to describe, in part, Oklo’s process for safety classification of SSCs. The NRC staff performed completeness reviews of the topical reports using the Office of Nuclear Reactor Regulation’s Office Instruction LIC-500, “Topical Report Process” (ADAMS Accession No. ML19123A252). The staff determined that neither topical report contained sufficient information to initiate detailed technical reviews. Each report contained conceptual information, rather than repeatable methodologies, and each left many issues unresolved and open for future potential applicants referencing the topical reports to address. The NRC staff informed Oklo of the insufficiency of the topical reports by two emails dated August 5, 2021 (ADAMS Accession Nos. ML21201A079 and ML21201A111), that included attachments describing in detail the supplemental information Oklo must provide for the NRC staff to begin the detailed review of each topical report (NRC Forms 898 – ADAMS Accession
Nos. ML21201A094 and ML21201A113). The NRC staff identified five areas where additional information was needed for the MCA methodology and three areas where additional information was needed for the PBLM methodology. The NRC staff held public meetings with Oklo on September 1, 16, and 28, 2021 (meeting summaries available at ADAMS Accession
Nos. ML21259A260, ML21266A428, and ML21293A329, respectively). During these meetings NRC staff responded to Oklo’s requests for clarification on the information needed to address.