Wouldn’t Occam’s razor say that the people speaking are highly motivated to continue being paid? Relying on their fear of future outcomes seems to be misplaced to judge the honesty of their statements.
There are many assumptions that need to be true to link "they say they're terrified of AI" to "they get paid more / longer", in both their own minds and everyone else's. It's a bit of 4D chess that could go their way, but also could go 100 different ways.
In contrast, linking "they say it" to "they believe it" is simple.
I think it goes beyond money now. A future where AI does not destroy the world means a world where the people who work at these companies are just like everyone else.
Genuinely, what makes people not believe in coming AGI?
I read quite a lot on hackernews but never encounter a strong post/comment against it, apart from the „what is marketing“ discussion.
Maybe people don’t bother because everyone is in their trenches, tough I don’t understand how people do not see the AGI coming.
What kind of doom, if any, will befall is something different of course.
I don’t want to believe there will be AGI in the next 5 years, but it sure is hard not to!
I think it's because people have different ideas of AGI. Some only expect general intelligence, while others also require weird speculative ideas, such as rapid self-improvement.
Right now, what seems to be missing from general intelligence is mostly the ability to learn from experience and continue operating in a useful way well beyond context window size. That might already be technically feasible, but probably not in a cost-effective way.
What does AGI mean is probably the sticking point for most. I'm inclined to call it AGI if they can create something currently matching SOTA where there is no difference between training and inference time, i.e. it can continuously learn. Without the model collapsing.
I don’t believe AI progress is a linear line that ends in “AGI”. I think the current way we create LLMs will be improved, and improved, and improved. But an LLM at its core is a predictive chain, not general intelligence.
The private equity debt time bomb is here. Then there is the bond market.
And then guaranteed bust of AI bubble. 2000 billion dollars for a single company? That too when bottom of the boat has been drilled by open weight Chinese models that are cheaper and good enough but can even be deployed locally?
In face of all that - that kind of return is not realistic.
However - good gimmick scaring UN asking for a ban on Chinese models perhaps.
The UN is not a body that would be involved in banning models from a specific country, and even if they were China is one of 5 members with permanent veto power.
The AI industry in Russia is not doing well. Domestic AI models exist, but they aren't exactly high performers, and many of them are based on the open weight Chinese models - so there is very little reason to use them. IT power users run Western or Chinese models, often through proxies, VPNs or resellers.
AI is being adopted and used, but if there was ever a chance of Russia being an AI power, it died in 2022.
A few formerly-Russian IT companies had their "rainy day" HQs overseas, and leveraged them to slip the noose and leave Russia - brain draining what AI expertise was available in the country. See: Yandex -> Nebius.
Right, that's what I'm thinking. These folk (ai operators) know they are overvalued, see the open model threats and the performance improvement on lower-end hardware. So, no need for all these data centers. The "force majure" or whatever that lets them exit those commitments. So all that accrual based revenue is gone.
Is ads in the AI gonna drive that revenue valuation? Are buyers going to lock in contracts when a new model in 120 days is ever better and cheaper?
Are we still pretending models aren't cyber capable, and will be an immense force for hacking? I expected this shit on reddit not here.
You don't have to like Sam, and still not do this bullshit
Unfortunately focusing on the fanfic prevents everyone from focusing on the genuinely world-altering threat of cybersecurity becoming completely impossible for a while. But perhaps that’s the point, because you can be held liable for that, but not for the end of the world as there wouldn’t be any courts afterward.
Please tell me they are going to have independent computer scientists and security experts to ask questions that expose the doomerism/"rationalist" LessWrong bullshit. Please don't just let him spew talking points designed to make a deflating bubble look like a good thing.
Doubtful. Almost definitely going to be taking direction from "business" people with misaligned and ulterior motives, rather than asking questions of people with long careers, research, and history in the field.
These kind of meetings are to create common vocabulary/seed common frameworks across the world for problems all govts are seeing or expecting on the horizon.
HN would be better if it did not produce disparaging comments in every thread that has Sam Altman on it.
I get it. We dislike Sam. I got it 3? 4? years ago. Would be nice and more interesting if we could see past that at some point, given that he is a fairly important figure, doing important things, that would be cool to share interesting thoughts about.
He will probably get up there and say his companies products will destroy every citizen of their countries.
reply