OpenAI canceled Astra 6.1! Because it might look worse than Sonnet 5.5 at 5× the output-token price… uh, sorry: because of safety concerns. Just hours before DevDay. Impeccable timing.
Donald Trump Has Abolished AI. No Joke!
Donald Trump has a problem with artificial intelligence. Not with the technology, but with the word “artificial.” It makes AI sound “fake,” he told the UN General Assembly. So in official US documents, AI is now supposed to be called “Super Intelligence.”
Hold a Vote, Then Pick Something Else
Before that, Trump held a vote on possible names. “Superior Intelligence” and “Extreme Intelligence” made it to the final round after “Supreme Intelligence” was eliminated.
Then, at the UN, Trump announced “Super Intelligence.” If you're wondering what the vote was for, you clearly haven't grasped the rules: The audience gets to play along as long as the boss picks the winner.
But “superintelligence” already means something. The White House's own Economic Report uses the term for AI that exceeds human capabilities. Apparently, every AI now surpasses human intelligence. In the case of one particular human, a president, that might even be true.
The joke would be harmless if this were just a silly name. Trump dismisses warnings about AI risks, and now he wants to slap the term for an especially powerful form of AI that doesn't yet exist on everything. Call every model “super,” and it becomes harder to talk about what those models can actually do and where they fall short.
It's unclear whether the renaming has been formally ordered. In his speech, Trump himself wondered whether he had the power to make the term catch on worldwide. Maybe he does, maybe he doesn't. One thing is certain: A new abbreviation won't make ChatGPT any smarter or more reliable. It would be progress if the administration could tell the difference.
I have a few questions about GPT-6...
Why isn't GPT-6 Sol and Luna better than GPT-5.6?
Why is a higher “Thinking” level often worse than a lower one?
GPT-6 Sol and Luna are no better than GPT 5.6 Sol and Luna.
Artificial Intelligence: Sol went from 47 to 48 points. Luna went from 37 to 37 points.
Disappointing!
The most important feature for ChatGPT? Switching between two accounts in the desktop app! Am I right?
As with most software, we usually make things far more complicated than they need to be. To change that, I’m sharing 10 tips and shortcuts today that will make working with ChatGPT much easier and more efficient in the years ahead.
That's cool — you can see if your X account has been shadowbanned with 1 Click!
nitter.cf/i/under_the_hood
How is GPT-Image 2.5 Sunburst this overlooked? One image → edit → edit the edit → repeat 20x. Still consistent!
🤖 Made with AI
OpenAI gives users more compute. Users give Tibo a halo. And the company may shift demand into quieter weekends at the same time. That’s not charity. It’s very smart marketing, with a small catch hidden in the calendar.
This is the first time I've found AI to be scary good. Well done, OpenAI!
We’re sharing a solution to the Navier-Stokes Millennium Prize Problem, one of the deepest problems at the frontier of mathematics.
The proof was produced by a group of agents, using an OpenAI next-generation model significantly more capable than GPT-6 Astra.
The problem concerns whether the description of smooth three-dimensional fluid motion modeled by the Navier-Stokes equations can break down. It has remained unresolved for roughly 90 years.
StratosBase retweeted
ChatGPT Work kopiert ab sofort deinen Schreibstil aus Slack und Mails. Trainiert die KI in unseren Postfächern jetzt nur noch auf sich selbst?
all-ai.de/news/news26top/cha…