> Previous reports suggest it took OpenAI four days to realise...
At the speed AI can achieve work, this could be far, far too slow to contain a future genuine problem. There's danger in operating at faster speed than humans.
I have only lightly tracked this, since I'm not American, but I understand that there's now a strong right-wing focus in 60 Minutes, is that correct?
Ross Douthat's Wikipedia page makes him sound strongly religiously conservative: anti-abortion, a book title 'a nation of heretics', publications like 'Believe: Why Everyone Should Be Religious' and 'The Decadent Society: How We Became the Victims of Our Own Success'.
Americans who know who he is more than a Wikipedia profile: is this a positive move for neutral, informed journalism, or is it likely aligned with presenting specific agendas?
Could be a copy/paste error, or could be an LLM trying to cite its own work by just looking at a list of sources used and guessing which one the claim came from.
Yes. I switched to Codex some time ago. I feel like I get fewer outages (had one over the weekend), and coding results are good. I built my own agent harness using OpenAI models too and feel it was the right decision.
Very long to read, but highly principled. I feel his desperation. Powerful people can too easily do little, and ignore their consciences. I appreciate he named those he spoke to; it came across as fair, and makes them accountable for what they did or did not do.
I don't either. I clicked on Modal and it showed me info about modals. Below the fold there's a text entry to 'UI element, behavior, or implementation problem'. I thought maybe this was the quiz. I entered:
> I click sign up and something appears, where I have to enter my email and a new password
Minimal (and avoids saying 'a modal' or similar deliberately) but IMO fits the requirements. The reply:
> We do not have enough information to diagnose this pattern yet. Add what opens it, whether users can interact inside it, whether it blocks the page, or how it closes.
If I enter that I'm solving my own problem on behalf of the webpage... this is really confusing.
We are already seeing math problems being solved by AI on a weekly basis now. Maybe at some point a breakthrough in math will cause a breakthrough in physics?
It makes me wonder if the model is the limit here or if it's the harness: e.g. give GPT 5.6 or Fable a team of 20 people that can easily interact with the real world and let it do whatever in order to "make a breakthrough" in physics; do you think it would manage to?
In a very real sense, food was solved by the first two industrial revolutions, which between them supplied enough artificial muscle power to move most of us out of farming, and enough understanding of biology to bring us industrial-scale fertilisers.
We're living through yet another power revolution today, between renewables and batteries (and while I won't count it until it's commercialised, we've gone from fusion being decades away to years away).
FTL… well, like most nerds I've got a soft spot for the Alcubierre drive and the various follow-up papers. But for now, sure, I'd assume by default that any news headline I see about someone making even a sub-light 1 µc version (I guess 1 µc would be sub-sonic, too) to be either an AI hallucination or something like The Onion.
I like this, but the writeup is so Claude that I suspect the regex was written by Claude too :( Which is a shame, because the 'Doom running on...' space is one of the ones where I treasure articles written by motivated hackers (who may use AI, but do some work - idea, iteration, AI as tool not author - themselves.)
(I saw no info what amount of involvement the human author had here so the article's Claudeisms are leading me to feel Claude-ish about the whole project. That may not be fair.)
At the speed AI can achieve work, this could be far, far too slow to contain a future genuine problem. There's danger in operating at faster speed than humans.
reply