this screenshot from GPT-5 livestream has to be among the worst chart crimes of the century
Ege Erdil
7,632 posts
Ege Erdil
@EgeErdil2
update your effect size estimates downwards
- this is illegal since the relevant legislation only gives USCIS the authority to charge fees in order to cover costs, and is most likely going to get struck down by the courts in a few months people shouldn't discuss it as if it's a serious policyTrump signs an EO on H-1Bs raising "the fee that companies pay to sponsor H1-B applicants to $100,000"
00:00 - some people struggle to understand the nuanced position of - trump's H-1B order is illegal & will be struck down - but broadly trump's attempts to reform the program are good - a policy with $100k/yr fees is probably bad, a policy with $100k one-time fee would probably be good
- imo this is not right the reasoning/agency RL is resulting in a lot of unreliability, hallucinations, reward hacking, etc. that will seriously impede consumer use cases if not addressed much of the cost of having an unsafe model is internalized for this reason aloneI genuinely believe Anthropic will end up losing the AI code model contest simply because they’re so obsessed with safety
- why don't michelin 3 star restaurants raise prices instead of having 3-6 months of waiting time? is it so important that people should be able to book tables months in advance?
- feel like more people should be aware of the fact that the odds of ww3 until 2050 have tripled over the past 3 years, from ~ 10% to ~ 30%, according to the metaculus community
- the moral panic about teenagers talking to chatbots is ridiculous
- in sweden, any person: - can give legal advice to anyone else - call themselves a lawyer - appear before any court at any level of the court system to represent themselves or anyone else no law degree or bar association membership necessary
- i don't think AI has solved mathAI has *solved* math. OpenAI did it with o4 Not "is close to solving math" Not "is competitive at math" *SOLVED* This is far bigger than anyone realizes. Let me explain why. First, you need to understand some historical context. Typically, with AI/ML you know that you're
Readers added contextOpen AI did not "solve" math Source: Noam Brown, research at OpenAI leading test time compute scaling x.com/polynoamial/st… - i'm very confident trump's trade wars won't turn GDP negative GDP growth is another matter though
- this is hard cope for a model that's in fact significantly worse than 4.5 at writingwe've been testing some new methods for improving writing quality. you may have seen @sama's demo in late march; GPT-5-thinking uses similar ideas it doesn't make a lot of sense to talk about better writing or worse writing and not really worth the debate. i think the model
- my wild and speculative guess is that the people benefiting might be the people who in aggregate are paying openai $10B/yearWho's actually benefiting from gen-AI? 🙄 "the benefits of AI seem esoteric and underwhelming, while the harms feel transformative and immediate" ⬇️ wired.com/story/generati…
- i want to note that it's *immoral* to defraud customers out of billions of dollars, because this doesn't seem to have sunk in for some people yet if you're only opposed to this because of the reputational damage it might do to EA or whatever, you have the wrong idea, i'm sorry
