The replies you receive (I believe) reflect the level of understanding people possess. Nearly everyone is getting caught by surprise before they know anything.
I also like it because I think it's substantively correct.
The problem with the Bright Mirror video is that its core message (past warnings of doom proved to be untrue so this one is likely to be wrong too) is, while something that folks should keep in mind, not something we should be too comfortable about either.
It reminds me of skeptics in the early day of COVID (e.g. Musk, Ionnadis, Berenson, etc.) who predicted that COVID wouldn't be that big a deal in large part because predictions of past doom for things like Ebola and Legionnaires hadn't ultimately caused widespread deaths. Or the people who don't worry about phones and kids because of past "moral panics" about TV, video games, rap music, etc. that had been overblown.
Skepticism informed by experience is valuable, but it's certainly not dispositive. Sometimes things change. Just because a boy has often cried wolf doesn't mean that a wolf will never come.
This really pulled together a lot of what we’ve been talking about around AI.
What caught my attention most is that public concern is finally catching up to the speed of the technology. If 77% of Americans now want AI development slowed or paused until safety can be evaluated, that’s not a fringe concern anymore. And the fact that data-center opposition is cutting across party lines tells you this isn’t just about ideology—it’s about people feeling the local impact on land, water, power, jobs, and quality of life.
That connects directly to The 99% Perspective’s Article 8 — AI, Automation, and the Future of Work, Series 2 Article 8 — Algorithmic Power and the New Digital Gatekeepers, AI Without Brakes, and The Missing AI Infrastructure Forecast.
The bigger question for me is: does public concern lead to actual policy before something serious happens, or do we wait for the crisis first?
I've yet to see the discussions of AI financing and AI safety converge, but I think they should and will. The current pace is possible because of massive investment far ahead of revenue, and buoyed up by financial shenanigans. This is both hurting safety and inflating demand, so we have lousy ideas of what both of those really look like. Right now we have plenty of coverage of the bubble and of safety, but separate. They're not separate.
On your first question, "1. Will data center opponents fight for a pause?": I've gotten a few emails from Food & Water Watch in the past few weeks that demand a ban on superintelligent AI altogether, calling for recipients to take action. To my knowledge FWW had previously been narrowly focused on opposing AI data centers (they have my email in the first place because I signed their anti-AI-datacenter petition sheet at the Stop the AI Race protest in July). So this is a big update for me.
Great reflections and I agree with the general thrust!
My only quibble is with the paragraph on how "In 2020, a California labor law targeting Uber and Lyft was neutralized by an in-app campaign to convert rideshare users into advocates." I don't think any such definitive evidence has been provided that it was this particular tactic that led to the passing of Prop 22. If anything, the evidence suggests the opposite; the richest counties in California voted "no" on Prop 22 while counties with the most lower-income residents voted overwhelmingly "yes." This suggests that most lower-income people were unhappy with the attempt by the state legislature to effectively shut down Uber.
Excellent analysis and an encouraging guide to the days ahead. And it is fair to say that we don't know what is exactly ahead, but the positions are not permanent. First--I myself believe we must pause moving forward--and begin by setting down the rules, limits and regulations for development. Second--we need to separate the issue of competition with China from the AI debate. We can worry about that later--and hopefully, with better national leadership, in the future, work out common agreements and understandings with China--and other AI leaders. Finally--I have a question? Who asked for AI to be immediately constructed and foisted upon society with all deliberate speed and no discussion? The American people were not clamoring for it--I don't believe, and are still just waking up to the problems and prospects. It is not helpful for people to claim we "will lose to China" if we don't move now. That is not THE problem--or even something to worry about.
IMO putting all the effort into making better models is neglecting tool use and diffusion. Slowing down the former will help stocks and the economy (partially be limiting catastrophic tail risk). Someone with a platform should make the case!
Thanks for confirming my suspicions about you being a wolf in sheep's clothing. The stochastic parrot argument is completely correct. Anyone bringing it up in order to try to dunk on it is so infected by the silicon valley brainworms that they have lost touch with reality.
Love your writing!
Your point that data center opponents don't bring up AI shows up in our data too.
We've (www.alpharoc.ai) been querying people on AI and data centers since August, and we let them answer in their own words.
We asked people who live near a data center what it's been like. About 300 answered. 3 mentioned AI. Mostly they talked about bills, noise and water.
We also asked what one issue will decide their vote for Congress. About 1 in 100 said AI or data centers. Cost of living was the top answer by far.
Happy to share more, including cuts by party and age, if it helps a future piece.
The replies you receive (I believe) reflect the level of understanding people possess. Nearly everyone is getting caught by surprise before they know anything.
I also found that claude music video very catchy but then I found one that's even catchier for me https://x.com/nickcammarata/status/2104485485666291850
I think Ethan Mollick's "Bitter Lesson" video is the best I've seen so far:
https://x.com/emollick/status/2106111291014234462
I also like it because I think it's substantively correct.
The problem with the Bright Mirror video is that its core message (past warnings of doom proved to be untrue so this one is likely to be wrong too) is, while something that folks should keep in mind, not something we should be too comfortable about either.
It reminds me of skeptics in the early day of COVID (e.g. Musk, Ionnadis, Berenson, etc.) who predicted that COVID wouldn't be that big a deal in large part because predictions of past doom for things like Ebola and Legionnaires hadn't ultimately caused widespread deaths. Or the people who don't worry about phones and kids because of past "moral panics" about TV, video games, rap music, etc. that had been overblown.
Skepticism informed by experience is valuable, but it's certainly not dispositive. Sometimes things change. Just because a boy has often cried wolf doesn't mean that a wolf will never come.
i just liked the music
I did too. But felt a little bad about it because of the message, so I figured I would post that comment.
This really pulled together a lot of what we’ve been talking about around AI.
What caught my attention most is that public concern is finally catching up to the speed of the technology. If 77% of Americans now want AI development slowed or paused until safety can be evaluated, that’s not a fringe concern anymore. And the fact that data-center opposition is cutting across party lines tells you this isn’t just about ideology—it’s about people feeling the local impact on land, water, power, jobs, and quality of life.
That connects directly to The 99% Perspective’s Article 8 — AI, Automation, and the Future of Work, Series 2 Article 8 — Algorithmic Power and the New Digital Gatekeepers, AI Without Brakes, and The Missing AI Infrastructure Forecast.
The bigger question for me is: does public concern lead to actual policy before something serious happens, or do we wait for the crisis first?
That’s the part we should be talking about now.
I've yet to see the discussions of AI financing and AI safety converge, but I think they should and will. The current pace is possible because of massive investment far ahead of revenue, and buoyed up by financial shenanigans. This is both hurting safety and inflating demand, so we have lousy ideas of what both of those really look like. Right now we have plenty of coverage of the bubble and of safety, but separate. They're not separate.
On your first question, "1. Will data center opponents fight for a pause?": I've gotten a few emails from Food & Water Watch in the past few weeks that demand a ban on superintelligent AI altogether, calling for recipients to take action. To my knowledge FWW had previously been narrowly focused on opposing AI data centers (they have my email in the first place because I signed their anti-AI-datacenter petition sheet at the Stop the AI Race protest in July). So this is a big update for me.
Great reflections and I agree with the general thrust!
My only quibble is with the paragraph on how "In 2020, a California labor law targeting Uber and Lyft was neutralized by an in-app campaign to convert rideshare users into advocates." I don't think any such definitive evidence has been provided that it was this particular tactic that led to the passing of Prop 22. If anything, the evidence suggests the opposite; the richest counties in California voted "no" on Prop 22 while counties with the most lower-income residents voted overwhelmingly "yes." This suggests that most lower-income people were unhappy with the attempt by the state legislature to effectively shut down Uber.
Excellent analysis and an encouraging guide to the days ahead. And it is fair to say that we don't know what is exactly ahead, but the positions are not permanent. First--I myself believe we must pause moving forward--and begin by setting down the rules, limits and regulations for development. Second--we need to separate the issue of competition with China from the AI debate. We can worry about that later--and hopefully, with better national leadership, in the future, work out common agreements and understandings with China--and other AI leaders. Finally--I have a question? Who asked for AI to be immediately constructed and foisted upon society with all deliberate speed and no discussion? The American people were not clamoring for it--I don't believe, and are still just waking up to the problems and prospects. It is not helpful for people to claim we "will lose to China" if we don't move now. That is not THE problem--or even something to worry about.
IMO putting all the effort into making better models is neglecting tool use and diffusion. Slowing down the former will help stocks and the economy (partially be limiting catastrophic tail risk). Someone with a platform should make the case!
Thanks for confirming my suspicions about you being a wolf in sheep's clothing. The stochastic parrot argument is completely correct. Anyone bringing it up in order to try to dunk on it is so infected by the silicon valley brainworms that they have lost touch with reality.
In which direction? Because I can’t really tell.
The writer is on the mid-to-extreme end of pro-AI development as far as I can tell, if that's what you meant
That was my question but I don’t get that impression.