Despite all the chatter about the White House softening on AI regulation, Trump's position is unchanged since the beginning of his admin -- no new regulations.
President Trump said he called off the signing of an executive order that would address cybersecurity concerns raised by powerful new AI models because he objected to parts of the directive https://t.co/RTI14dnOLj
So much for the White House warming up to light AI regulation. Trump has postponed the exec order on voluntary testing, saying he doesn't want to do anything that could slow down the US AI industry in its race against China
https://t.co/N9HOQQCsFb
My friend Shawn Wen made the definitive podcast on Bob Lee, the tech founder who was murdered in SF in 2023. It's an incredible piece of journalism and I recommend everyone listen to it! She's doing a live reading of the pod next week! https://t.co/9e2NjW1Gdt
I have sat in-person for about seven Google I/Os now and this one maintained all the forms of the previous ones. Advancements that come across as incremental, touching on the various product "regions," giving Search and Commerce their due, etc.
100% agree on Demis though. You feel like he really believes it, when the same words coming from any other Google exec would seem fake
Google announcing a new Google Glass but without a heads-up display. So it has cameras and can see what you see but it only talks to you through voice, rather than popping up images or words in front of your eyes.
Making a working display that looks good and doesn't make people nauseous has been a devilishly hard problem
Google I/O is tomorrow, last chance to get predictions in. I love to guess, so here's mine:
The Google team is being strangely quiet about the new Gemini. At this point everyone knows it is arriving tomorrow, along with their personal agent named Spark. This reticence, of course, can be interpreted in many ways. I'm choosing to interpret it in accordance with my nature.
I think they trained the largest model they've ever successfully trained - possibly the largest one anyone ever has. And something unexpected emerged at scale. They had their Mythos moment, but not in the same way Anthropic did. Gemini has always been a very different model from Claude.
The benchmarks will go out tonight under embargo (they probably already are), but I don't think they will fully reflect what I'm talking about. I think they hit something they weren't even aiming for. Something that surprised them. If I'm right, that surprise will be part of tomorrow's show. We shall find out together in the morning.
We asked Washington Post tech reporter @GerritD whether the OpenAI trial got AGI-pilled.
He says Elon Musk tried to make the case about AI safety and humanity's future, but the judge kept pulling it back to corporate litigation.
"The judge didn't want this to be a... you know, sort of a doomer versus accelerationist trial."
"I think that's probably pretty wise. I think she realized that if suddenly we had to sort of explain, 'Okay, this is what this means, this is what existential risk is,' yada, yada, yada, it would've kind of become bigger than the questions, the legal questions that she wanted to keep the trial focused on."
this is right. The judge let it go to trial because she believed there was an open question about what Elon knew and when. The jury rejected Elon's arguments about this.
That’s an oversimplification. The limitation was already known. Part of his argument was he was deceived and they hid the evidence from him. The jury rejected this argument on its merits. Because they didn’t buy his argument, the result was the dismissal due to the limitation. But it wasn’t simply too late. They disagreed with his argument.
the reality is that OpenAI's fate was in the hands of a third party, the judge/jury. A federal judge thought there was enough of a fact question over what Elon knew when to let it go to trial. I think a smart person would have guessed this outcome but it wasn't a 100% foregone conclusion
@Blogsbloke@R5Crosby there was some difference of opinion over what Elon knew when. He said he had become suspicious of OpenAI earlier but didn't feel 100% that they had "stole a charity" until later.
Musk has lost his case against OpenAI in the most anti-climactic way possible. Less than 2 hours of jury deliberation despite 3 weeks of testimony and hundreds of documents. The jury didn't weigh in on whether Musk was right or wrong, they just said he failed to bring his claims in time
Musk lost his case, but he managed to us it to bring up a lot of unpleasant information about Sam Altman and Greg Brockman, including detailed numbers of their personal investments in companies that do business with OpenAI. https://t.co/wh04MTGUTx
there's been a clear thread of "why are we even here" in the background of this entire case. There's an argument it shouldn't have gone to trial. The judge kind of addressed this in her remarks after the verdict today:
“I think it’s an important issue to be tried, it’s important to have trials, they bring clarity"
@GerritD It was always fairly evident this was an extremely weak case for him on Statute of Limitations. Even some of the evidence introduced by Musk to win on the merits undercut his SoL case. But discovery and the trial itself were fun and enlightening!