It may make sense to slow down research for a while, until society agrees on which direction to take.
But deployment should not be slowed down.
Barely 1% of people are using truly transformative AI, and if impact is to be genuinely real and positive, adoption needs to be maximized.
The focus should shift toward spreading this technology
Very excited to share our interview with @polynoamial on AI for math — the Erdős unit distance problem, saturating the IMO, the future of math research, and more!
Our internal data shows Claude is accelerating AI development—a possible path to recursive self-improvement, or AI autonomously building a more capable successor.
It’s happening faster than we thought, and the implications deserve greater attention. https://t.co/OVVPJO7VQx
In 1999, Ray Kurzweil predicted human-level AI by 2029. Stanford's top AI experts said 100 years. Three years from the deadline, he says two things are still missing: AI that understands physics, and a robot that knows the difference between what goes in the fridge and what goes in the trash.
-- LLMs went from unremarkable to world-changing in 12 months. AI now outperforms human doctors by 50% in diagnosis.
-- LLMs have only been truly effective for six months, according to Ray. A year ago, "not really all that impressive."
-- The Genome Project was 1% done at the halfway mark. One doubling later — done. We're at that same moment with AGI.
-- Software alone has improved roughly a millionfold over 70 years. Total compute gain = hardware (75,000 trillion fold) × software (million fold).
We’ve been researching new ways for ChatGPT memory to carry context across conversations and keep it useful over time.
Today, that work is rolling out as a more capable memory system in ChatGPT. https://t.co/0MyFKCe2Mu
🚨 SCOOP: Anthropic is gearing up for the public launch of a new version of Mythos, better than Mythos Preview.
A checkpoint of the model, codename Oceanus, was made available to red teamers yesterday. These programs typically begin 7 days before the wider launch, I am told.
The program has since been paused, presumably due to an individual in the program reselling it via a Chinese API proxy (as reported by @scaling01). But whether this will impact the launch target is unknown.
@franceinter L’Église catholique tente de résister à sa propre obsolescence.
La Silicon Valley vient la battre sur son propre terrain en proposant du concret.
C’est fini, le Dieu imaginaire et le salut promis uniquement après la mort.
Le match est plié!
Un milliard de personnes sur terre vivent dans l'extrême pauvreté.
Plus de 700 millions vivent quasiment sans électricité.
Des millions de personnes meurent car non soignées alors que les médicaments pour leur guérir existent.
Je pense que le message à faire passer, c'est que l'IA est une chance pour l'humanité.
Et il faut faire une promotion massive pour son utilisation et non le contraire.
🎶🎶
"Very soon, an AI will just be running for you in the background all the time… aware of all your context, doing as much as it possibly can to help you out."
🎶🎶
My favorite quote from this transcript -
Sam Altman said:
“Very soon, an AI will just be running for you in the background all the time… aware of all your context, doing as much as it possibly can to help you out.”
We remember OpenAI acquired Jony Ive’s io.
OpenAI keeps talking about ambient AI.
OpenAI keeps describing AI as something that follows you throughout the day rather than something you open in a browser (still cool & they’re working on that).
Which could be hinting at a hardware product unveil potentially soon™️ 👀