Packy...Let's help you with some perspective and context. You write: Once a regulation is in place, it’s incredibly hard to roll it back. Oh, not at all. The Supreme Court is busy flipping over decisions you thought were solid and something you could count on. Think like a woman. Use your imagination to ask, what fresh new hell could some fool possibly come up with that would reverse what I am now believing?
I'm with Packy and Ben Thompson on the problems of regulatory capture and self-dealing among early AI incumbents. My sole point of pushback is that one doesn't have to be a full-blown AI doomsayer to think some degree of caution may be a good idea. I'd guess that's more or less the average person's stance on it.
The argument for optimism-based AI laws to, "...promote research, competition, and experimentation to push U.S.-based companies to the frontier of A.I. and ensure the nation is at the cutting edge..." sounds a lot like the arguments we heard for why gain-of-function research in virology was worth the possible risks. Oopsies.
Then again, I'm as skeptical of this (or any) administration's wisdom and restraint when it comes to the proper role of government as I am about their sophistication regarding the future of AI research. Probably not a great sign.
You asked for feedback, I’m not throwing shade, just giving you honest feedback from my (possibly inaccurate) viewpoint.
I finished listening to E3: Nuclear Economics. I like what you’re doing, but I think you’re getting some significant things wrong. The characterization of the issues with concrete at Vogtle 3 is highly inaccurate, the concrete was fine when it was finally poured, there was an issue with the basemat (backfilled dirt) settling more than predicted that was resolved by changing the concrete mix, then there was the bigger issue with the rebar. The rebar got pulled out and replaced before any concrete was poured. The specifics of the rebar would have been much more enlightening.
The idea of having to spend more money on buying more expensive “nuclear grade” concrete that cost much more than the cost of materials is also silly, the concrete batch plant was built on site and the raw materials were bought, but the mixing was done on site. Yes, building a medium sized concrete plant is a capital expense, but there is a lot of extra testing that needs to be done before each pour and concrete doesn’t store so having a short travel distance is a good thing.
I’m also frustrated every time I hear someone saying the lack of a trained workforce is an issue. This is mostly wrong, and without getting more specific tends to imply the craft people were the problem. I mean sure, there were a few troublemakers, and some were lazy, but most were professionals and knew their craft. Building a nuclear plant is different, but not much of that difference is at the craft level.
Fun story: I was attached to a submarine during defueling. One fine day the Naval Reactors rep (NR, the Navy’s NRC equivalent) decided to climb up to talk to the crane operator.
NR: How do you like your job moving highly radioactive fuel.
Crane operator: Trash cans or fuel cells, it’s all the same from up here.
An hour later all work was stopped for three days while the crane operators and other civilians got training on how nuclear is “special and unique” and how important it is to never again say such things.
The talk of a consortium of utilities was interesting, but a light version of that was started for the AP1000’s, if you search for APOG AP1000 (APOG is the AP1000 owners’ group) it will pull up some of the NRC filings and you can see which utilities were in (not financing, just information sharing and workload sharing on the utilities side (ex. Ownership of procedures would be divvied up)). That fell apart after Fukushima when all of them except SCANA and Vogtle canceled or put on hold, their projects.
Anyway, if you’re just having fun with making a podcast, great, I love it. If your goal is actual identification of things that could be done to make nuclear cheaper, or easier, you aren’t getting enough information to draw the correct conclusions.
OBTW, Nick Touran is awesome, I’m going to keep listening to episodes just to hear his insight.
Hi Packy, how did you approach it when you first started learning about AI, and how do you keep track of its development? Who do you follow and read? I have tried to break it all down for myself by writing everything I know to an imaginary 5yo (while my daughter is 3, she is not quite intellectually there just yet), love to hear your thoughts!
Wait...Uranium is "just rocks." Maybe I'm an "idiot," or maybe I missed an episode where you dive into the obvious issues with fission, but it feels like this episode is heavy on evangelism -- to the point where I'm wondering where the money is coming from -- and oddly light on the 'buts'. Advise embracing our buts on this topic. I'd be more convinced and interested if the episode didn't very deftly sidestep the safety and waste issues (which are real, even if we had new and effective ways to manage them). I'd like to see a,"But duh -- lots of unavoidable examples of how, for all the regulations in the world, shit happened and bad bad things happened...and nobody wants the waste -- BUT we can effectively eliminate these issues by..." For a primer on applying the Two But Rule: https://www.2buts.com/p/embracing-your-but?r=1yad3e&utm_campaign=post&utm_medium=web
Exciting to see the continued progress in fusion energy and nanotechnology! The AI regulation debate will be an important one to follow.
Packy...Let's help you with some perspective and context. You write: Once a regulation is in place, it’s incredibly hard to roll it back. Oh, not at all. The Supreme Court is busy flipping over decisions you thought were solid and something you could count on. Think like a woman. Use your imagination to ask, what fresh new hell could some fool possibly come up with that would reverse what I am now believing?
I'm with Packy and Ben Thompson on the problems of regulatory capture and self-dealing among early AI incumbents. My sole point of pushback is that one doesn't have to be a full-blown AI doomsayer to think some degree of caution may be a good idea. I'd guess that's more or less the average person's stance on it.
The argument for optimism-based AI laws to, "...promote research, competition, and experimentation to push U.S.-based companies to the frontier of A.I. and ensure the nation is at the cutting edge..." sounds a lot like the arguments we heard for why gain-of-function research in virology was worth the possible risks. Oopsies.
Then again, I'm as skeptical of this (or any) administration's wisdom and restraint when it comes to the proper role of government as I am about their sophistication regarding the future of AI research. Probably not a great sign.
You asked for feedback, I’m not throwing shade, just giving you honest feedback from my (possibly inaccurate) viewpoint.
I finished listening to E3: Nuclear Economics. I like what you’re doing, but I think you’re getting some significant things wrong. The characterization of the issues with concrete at Vogtle 3 is highly inaccurate, the concrete was fine when it was finally poured, there was an issue with the basemat (backfilled dirt) settling more than predicted that was resolved by changing the concrete mix, then there was the bigger issue with the rebar. The rebar got pulled out and replaced before any concrete was poured. The specifics of the rebar would have been much more enlightening.
The idea of having to spend more money on buying more expensive “nuclear grade” concrete that cost much more than the cost of materials is also silly, the concrete batch plant was built on site and the raw materials were bought, but the mixing was done on site. Yes, building a medium sized concrete plant is a capital expense, but there is a lot of extra testing that needs to be done before each pour and concrete doesn’t store so having a short travel distance is a good thing.
I’m also frustrated every time I hear someone saying the lack of a trained workforce is an issue. This is mostly wrong, and without getting more specific tends to imply the craft people were the problem. I mean sure, there were a few troublemakers, and some were lazy, but most were professionals and knew their craft. Building a nuclear plant is different, but not much of that difference is at the craft level.
Fun story: I was attached to a submarine during defueling. One fine day the Naval Reactors rep (NR, the Navy’s NRC equivalent) decided to climb up to talk to the crane operator.
NR: How do you like your job moving highly radioactive fuel.
Crane operator: Trash cans or fuel cells, it’s all the same from up here.
An hour later all work was stopped for three days while the crane operators and other civilians got training on how nuclear is “special and unique” and how important it is to never again say such things.
The talk of a consortium of utilities was interesting, but a light version of that was started for the AP1000’s, if you search for APOG AP1000 (APOG is the AP1000 owners’ group) it will pull up some of the NRC filings and you can see which utilities were in (not financing, just information sharing and workload sharing on the utilities side (ex. Ownership of procedures would be divvied up)). That fell apart after Fukushima when all of them except SCANA and Vogtle canceled or put on hold, their projects.
Anyway, if you’re just having fun with making a podcast, great, I love it. If your goal is actual identification of things that could be done to make nuclear cheaper, or easier, you aren’t getting enough information to draw the correct conclusions.
OBTW, Nick Touran is awesome, I’m going to keep listening to episodes just to hear his insight.
Hi Packy, how did you approach it when you first started learning about AI, and how do you keep track of its development? Who do you follow and read? I have tried to break it all down for myself by writing everything I know to an imaginary 5yo (while my daughter is 3, she is not quite intellectually there just yet), love to hear your thoughts!
Wait...Uranium is "just rocks." Maybe I'm an "idiot," or maybe I missed an episode where you dive into the obvious issues with fission, but it feels like this episode is heavy on evangelism -- to the point where I'm wondering where the money is coming from -- and oddly light on the 'buts'. Advise embracing our buts on this topic. I'd be more convinced and interested if the episode didn't very deftly sidestep the safety and waste issues (which are real, even if we had new and effective ways to manage them). I'd like to see a,"But duh -- lots of unavoidable examples of how, for all the regulations in the world, shit happened and bad bad things happened...and nobody wants the waste -- BUT we can effectively eliminate these issues by..." For a primer on applying the Two But Rule: https://www.2buts.com/p/embracing-your-but?r=1yad3e&utm_campaign=post&utm_medium=web