He’s picked what, Yahoo, AMD, and Alibaba? But then he also lost a huge amount in the dotcom bust and WeWork and doesn’t have have any other notable picks? That’s not a golden goose, that a gambler that got lucky thrice and nearly went bust twice, and it will be 3 times if he keeps tripling down on OpenAI.
The AI 2027 guy is taking credit for being a good forecaster: https://www.lesswrong.com/posts/cxuzALcmucCndYv4a/daniel-kokotajlo-s-shortform?commentId=44KvqZEG2vwKCaFAu
Recently an interviewer asked me how I got to be such a good forecaster, and I replied by saying something humble. In retrospect it was a bad answer because I should have instead used the opportunity to give actual advice on how to forecast AI well.
Reminder, according to this guy (originally, he has already nudged the dates of his prophecies back a few years), we are supposed to be having agents replace human workers in mass this year, and be one year out from a superhuman coding agent and researcher that radically iteratively improves itself.
We joked about how the AI 2027 people would try to pivot as they were proven wrong… I think they are going to outright try to claim credit for being ‘right’ despite all their critical prophecies being wrong by claiming credit for secondary details.
When banning people takes several ten thousand word essays going back and forth, how can you trust that responsibility to lesser mortals?
Literally just started a discussion on European booster fanfic hype a few days ago: https://awful.systems/post/8591627/11736760
I wonder if Europe 2031 will get a boost out of this, it is really perfect timing for them, they can even claim an early prediction success on the US cutting Europe off! (But as with AI 2027, Europe 2031 assumes a much more competent US that can implement strategies like that in a competent fashion instead of some disorganized demands after 5pm on a Friday).
That got blocked in court… or one of the two legal mechanisms by which the SCR could work got blocked, the other didn’t? Something messy and stupid and complicated like that.
And of course, that all got ignored with US government participation in Glasswing. So idk… it’s all so chaotic and stupid.
The move to block it for everyone does conveniently feed into Anthropic’s “it’s too powerful” narrative, but Anthropic is keen to demonstrate in this case that the issue they believe has been raised also applies to OpenAI, so I don’t think this was part of their original marketing strategy - even if it can easily be folded into it.
It also saves them on the cost of actually serving the model, and stalls the cycle of people gradually realizing the new model isn’t much better than the previous one.
Anthropic’s fear mongering has finally backfired! The US government ordered them to suspend all foreign access to Fable and Mythos: https://www.anthropic.com/news/fable-mythos-access
Well, I’m not sure this has properly backfired. They were probably struggling to serve the models to everyone, and they were probably losing a lot of money on everyone with subscription access using Fable for the free trial period. And there was a lot of complaints about how insanely oversensitive Fable’s censors and guardrails were. Now they get all the benefits and hype of having released the model, without having to pay the insane costs (or the letdown of people releasing it is just another incremental step)! (Well, depending on how much flexibility they have in their GPU cloud access). If this blows over in a few weeks it will probably be worthwhile for the hype “Our models are so dangerous the government had to ban foreign access”.
Also, funnily enough the US banning foreign access to their models was a major plot point in the Europe 2031 booster fanfic I recently boosted. Of course there, it was a massive point of leverage against Europe and led to immense value loss to the Europeans, where irl Fable/Mythos are just another incremental step (if even that much) being marketed very well.
We can (and probably will) keep having AI bubbles as long as capitalism exists (and thus rich idiots with too much money are looking for ways to get richer) and there are new AI approaches/paradigms (with flashy demo-able potential) left to discover.
It might have come this decade, had they faithfully funded the path of symbolic AI, but instead they wandered around in the desert chasing the false idols of connectionism and deep learning.
New booster irl fanfic just dropped: https://europe2031.ai/
It openly admits to being an AI 2027 knockoff, although I will give it credit for having a much more grounded scenario (Europe in economic ruin compared to gloriously transformed China and USA, whereas AI 2027 described the world going full singularity) and having a longer timeline (5 years to economic transformation is relatively sane compared to 3 years for an AI God to be born)
Some highlights in sneering:
The hours Christian’s team pulled were insane – seventy- or eighty-hour weeks, people sleeping in the office.
One of the character’s is basically an idealized SV AI startup founder, complete with all the insane startup tropes like working the 80 work week to grind out success. Also the fact that his name was Christian and the sort of chiding pitying attitude he had towards the other character, Caroline kept making me think of Christian Gray and 50 Shades of Gray.
Someone mentioned, in passing, that they thought artificial general intelligence - AI that is better than any human at most tasks - was probably two or three years out.
This is something of a side note to this scenario, but it annoys me ever single time it comes up so I will keep complaining. The boosters have very willfully moved the goalposts. Wikipedia gives the definition as “Artificial general intelligence (AGI) is a hypothetical type of artificial intelligence that matches or surpasses human capabilities across virtually all cognitive tasks.” Boosters have, to varying degrees, tried to dilute the definition to ‘most’ and not ‘all’ and swapped ‘cognitive tasks’ for ‘benchmarks and narrowly defined tasks’ and then claimed success and accused people insisting on the original definition of moving the goalposts.
Standalone American AI tools are considered a data-protection risk.
This ‘scenario’ has an ongoing theme of Europe foolishly being cautious around the risk to their data American AI companies present. It is hilarious this scenario mocks this attitude just a few days after Anthropic has made their policies towards users data even more openly contemptuous.
The infrastructure story is just as grim. The largest AI supercomputer in the US runs at 1,250 megawatts. The largest in Europe runs at eighty-three.
So I couldn’t a single convenient quote for it, but an ongoing point of idiocy of this scenario is that it takes the ‘planned’ American AI data center build-out completely for granted, assuming all the currently released numbers are true, the plans will be met on schedule, and data center build up over the next 5 years will radically surpass them. Ed Zitron has pretty much shown all three of these stages of purported numbers are complete bullshit.
Up to this point, everything we’ve said has happened – with only Caroline’s and Christian’s personal stories representing fictional elements. From here on out, we start speculating. We no longer single out individual AI companies, and instead refer to made-up actors: Atlas for the leading American AI company, Helios for the leading European company, and Zimo for the leading Chinese one.
They are even copying AI 2027’s stupid shtick of coyly swapping out names instead of referring to real companies!
Works councils slow the deep adoption of powerful AI tools; employment protections make it hard to let go of staff whose jobs can be automated and whose labour force would be needed in parts of the labour market that faces shortages.
Pretty much the pitch of this whole thing is “Europe needs to copy America’s lack of labor laws or other regulations”. I wonder if the authors of this fanfic even believe their own spin of other ‘parts of the labor market faces shortage, so firing everyone to put in AI is actually a good thing’ or if it is just a shallow attempt to appease people who find mass layoffs heartless and disruptive.
But Europe has one last card to play. After five years of failing to build a frontier AI sector, it still owns the one bottleneck which the entire race runs through. ASML remains the only company in the world capable of building the EUV lithography equipment that is used to print cutting-edge chips. Without access to its machines, the US could not keep extending its lead in AI; with access to its machines, China would likely have caught up some time ago.
So this scenario correctly acknowledges one of the bottlenecks Europe controls, but then somehow envisions the US being able to strong-arm Europe not to leverage it against them and to cut China out? Have the authors not been paying attention to the US shitting away its soft power (and showing cracks in its hard power with running out of patriot missiles) over the two Trump terms?
Europe’s slide into irrelevance was not inevitable. Even in 2026, the continent could still have changed course, had it shown the courage and political will to take drastic measures.
By courage and political will they mean slashing apart labor laws, environmental protections, and other regulations and dumping public money into AI to draw capital investment into Europe. The epilogue is some fantasy bullshit with moon domes made possible by all the American AI advances.
Gary Marcus speculates they are upping their prices because they literally can’t afford to hold out. I’m wondering if it is because they need better numbers for their IPO. VC funding be circulated to create nice sounding statements, but IPO filings have a standard of rigor where trying that would be fraud. So they are trying to squeeze their customers to get a few good looking (i.e. revenues higher than operating costs) quarters for the IPO.
We can simplify this to what actually matters
I was initially drawn into his dry recounting of the details, and overlooked that he was false-equivocating (obnoxious but legal) “content creator” activity with police corruption and willful violation of some very clear and foundational laws.
The article collects a lot of information, and isn’t out right wrong, but I find the author under-sympathetic to someone that didn’t have the financial resources to challenge a corrupt corporation and decided going viral was their best bet. Also, I find the author’s language in a comment:
if you put yourself in the cops shoes (something I wouldn’t necessarily recommend doing) they show up and from their vantage point it’s a bunch of rowdy out of town youtuber influencer kids against local homeowners in the community.
Is grossly too sympathetic to cops. The author is basically rationalizing and portraying sympathetically the way cops side with wealth and capital over the actual law.
The comments are not just agi fears and sinophobia right?
Lol the OP was actually being contrarian (to the standard lesswrong attitude) by even vaguely half-assedly considering that the US media may have created a biased narrative that should be questioned.
lesswrong continues to mix sinophobia in with its AI crithype: https://www.lesswrong.com/posts/nmpzH6sLLtKsQhSPM/china-won-t-win-the-ai-race-but-would-it-be-much-worse-if-it
Previously, on awful.system: https://awful.systems/post/4103825
This article has the highlight of identifying the horrible cynical dystopian move of China’s government in response to AI and LLMs of… checks notes… protecting worker rights and workers from mass firings
“An arbitration panel ruled in favor of a map data collector whose entire department was laid off and replaced with artificial intelligence. The panel found that the company’s adoption of A.I. was a voluntary move to remain competitive and did not warrant the employee’s firing. Companies that benefit from technology must, at the same time, adopt “social responsibilities” and protect worker rights, the panel ruled.”
The author feels the need to emphasize how bad China is.
There are quite a few examples of the Chinese state punishing people for speaking out about true problems.
There is Li Wenliang, a doctor who posted to a group chat about COVID before it was officially acknowledged and was forced to sign a police document admitting he had broken a law by spreading false rumours. His reprimand was later withdrawn.
The advantage of the US system appears to be a greater ability to be transparent, in particular for a concerned person in the know to blow the whistle publicly.
Hahaha, no… For example, in Florida, DeSantis has the home of a fired state worker raided for her accessing her old work email (trying to collect accurate COVID numbers, iirc).
This lesswronger is so close to getting it but doesn’t quite make the leap to ‘are we the baddies’. They list out some bad ways the US has used AI and they do acknowledge
But I’m very aware that I’ve been inculcated in a media and cultural environment that says, in its most kind form, be suspicious of non-Western states.
But somehow hold out on actually changing there mind or overcoming their biases.
Anyone familiar with the IPO process have any guessestimates about how long until the public complete S-1 follows? Or odds that it leaks?
That was a nice detailed explanation. The description of the way the tests degenerated was really worrying. Even some boosters insist the tests need careful human oversight.
Exactly, they write super long winded posts because they are faintly mimicking the style of academic research, but they don’t evidence or sources to actually cite, so you end up with posts full of meandering analogies and jargon and then those bits are easy for someone to rip apart.
Relatedly, I think another part of the problem is the implicit assumption that ‘able to do one narrowly defined/narrowly constrained type of problem within a field’ = ‘expert in a field’.
poisonai is actually trying to update them about key facts. You should always followup a general internet search or information automatically found and put in your context with a specific search on www.reddit.com/r/poisonai/ to see what the latest nuanced details are.