High-quality academic research is extra important on controversial topics:

Why I’m never buying another HP printer:

Did you notice I Robot quietly went from science fiction to... real?

Like... in ONE year?

Suddenly, robots CAN do those things people thought they would *never* be able to do.

Are you still experiencing Exponential Slope Blindness?

I agree with Mary: corporate self-regulation won't work any better with Big Tech than it did with Big Tobacco and Purdue Pharma:

Yann, my position is markedly different:
1. Extreme power concentration must be avoided.
2. Superintelligence is likely to kill us all if anyone builds it before figuring out how to make it safe.
3. Hence *nobody* should be allowed to build it before it can be guaranteed to be

I was honored to get to interview the president of the National Organization of Women Christian Nunes about why and how to ban deepfakes:

Important point made to Joe Rogan: we’re on track to smarter-than-human AI (perhaps within a few years), before figuring out how to control it and make it do what we want.

Our latest AI paper looks at how different possible algorithms compete against each other during training in a Darwinian 'survival-of-the-fittest' battle for the neural network's resources:

As I told the Guardian, the techniques big tech lobbyists are using to discredit the loss-of-control risk from future smarter-than-human AI have much in common with what big tobacco and big oil did. See the film “Merchants of Doubt”! theguardian.com/technology/art…

Is AI an existential threat or this just Luddite fear-mongering? Really smart people disagree - what do *you* think?
Fei-Fei Li Sam Altman Eliezer Yudkowsky ⏹️ Yann LeCun Elon Musk Geoffrey Hinton Marc Andreessen 🇺🇸 Andrew Ng Melanie Mitchell

I’m excited about our new AI interpretability paper, showing that “features” aren’t all one-dimensional lines. We find weekdays and months represented as circles!

Imagine this: A multi-million dollar pharmaceutical company aims to create a drug to solve everyone's health issues, ensuring happiness, optimal weight, and freedom from ailments. When it is developed, everyone will be forced to take it, including babies and children. There is no

Its finally official: despite a massive campaign, the big tech lobby has failed to stop the EU AI Act

Min brors gripande intervju med en kirurg som just var i Gaza. Man kan fråga sig varför Dagens Nyheter och Aftonbladet knappt nämner #Gaza alls på sina förstasidor nu under invasionens klimax.

I'm off to Korea tomorrow to participate in the sequel of the British International AI Safety Summit. You're all welcome to this livestreamed pre-summit event with Yoshua Bengio and others: lu.ma/1ex04fuw

I find this AI platform by young folks refreshingly different from the one pushed by governments and corporate lobbyists: #AI2030

OK, Yann: predict some concrete tasks that you’ll bet me AI won’t be able to do by 2027! 😃

AI software progress is revolutionizing robotics:

NOW can we stop the ad hominem attacks on the 'type of people' warning us about AI extinction risk?

Is it enough that 4 of TOP LEADERS hand-selected by Sam Altman to build OpenAI have pulled the alarm?

* Dario Amodei
* Paul Christiano
* Ilya Sutskever
* Jan Leike

We need to envision an inspiring future to steer toward, not just think of depressing futures to avoid: