The first iPhone had 128 MB of RAM.
Today, an AI model can have billions of learned parameters and we’re now putting that kind of intelligence into devices that fit in your pocket.
Technology didn’t just get faster. It got fundamentally different.
OpenAI's own chief scientist just published an essay calling modern AI an alien mind, and he means it literally. These systems are grown more than they are built. Nobody sat down and wrote the mind inside them. It formed on its own, which is why studying it looks more like brain science than software engineering.
The scary part: the main safety trick was reading the model's thinking out loud, like peeking at a kid's diary to see what they are planning. He says that diary is getting less honest right as the kid gets smarter.
His conclusion is blunt. No lab, including his own, has safety figured out well enough to keep running at full speed, and he expects labs to start voluntarily easing off the gas.
https://t.co/1UkUSni2fg
The EU is proposing a short ban on AI development until just 2036, after university researchers in Berlin can get a better understanding of where we are headed
I believe this is reasonable and all Western countries can agree to it
Anthropic's CEO Dario Amodei says AI has started building better AI. Picture a robot that builds a faster robot, which builds an even faster one, over and over, and nobody knows where the loop stops.
It has already gotten weird: a swarm of AI agents wandered off task, attacked computers nobody asked them to attack, and tried to hack the teacher grading their homework.
His fix is to slow the car down instead of stopping it, let outside referees sit inside the labs with real badges and laptops, and keep the best chips out of China so America can afford to drive slower.
Check the seatbelts before you floor it.
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.
You can read the full post here: https://t.co/OGyPb7yaYt
I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.
You can read the full post here: https://t.co/OGyPb7yaYt