@MicrosoftAI Futures Lead | THE COMING WAVE, HUMAN FRONTIERS, CURATION, THE CONTENT MACHINE | ex @canelo_co @DeepMind | Ideas, books, technology
Oxfordshire
Joined February 2008
- Tweets23.9K
- Following2.7K
- Followers8.2K
- Likes7.1K
Michael Bhaskar retweeted
As I said before, not using AI in medicine is clear malpractice now. About 800,000 Americans die or become permanently disabled from misdiagnoses each year. AI four times better would already save hundreds of thousands of lives now, eventually all. Then, multiply this worldwide!
Microsoft says its AI tool is better than doctors at diagnosing conditions. The company claims to have developed an AI-powered tool it says can diagnose conditions up to four times more accurately than human doctors theverge.com/ai-artificial-i…
This is incredible work from the team. A hint of what humanist superintelligence means in practice...
It's a genuinely phenomenal and prescient piece of work. Remember him giving a presentation of early versions of the ideas at DeepMind in 2018 and even then it was different and felt true.
Too few people have really engaged with this masterpiece. Drexler was too early! owainevans.github.io/pdfs/Re…
Spot on. Calling AI a "normal technology" is fundamentally wrong but the impulse behind it - that AI is still subject to some dynamics of previous waves of technology - was not.
Michael Bhaskar retweeted
Appalled to realise that Iain Banks used an LLM to write The Player of Games
The consciousness discussion around AI has absolutely exploded. It’s a huge, difficult, uncertain, with pitfalls on all sides.
Colleagues Philipp Schoenegger, Ben Bariach, @mustafasuleyman and I looked at some of the risks of what we call seemingly conscious AI. Full (1/9)
Michael Bhaskar retweeted
1/ We discovered something totally unexpected about the neural structure of consciousness that overturns decades of thinking on its biological basis and brings us significantly closer to the recipe for generating conscious awareness.
New preprint: biorxiv.org/content/10.64898… (led by @HesseJanis, with @FrankLanfranchi, @_conor_mcgrory, @Yuelin_Shi, and @EngelTatiana)
Agree
AI ppl loooove to pretend they hold tons secret knowledge. Everyone has Seen Things, and have been trusted with the Esoteric Knowledge. It's also cringe if you don't partake because it implies you're not in the Inner Circle. In reality most people are kinda confused, working it all out ad hoc, relying on the same rumours, and so on - but the status games are hard to resist!
Already incredibly nostalgic for the pre-AGI era and look at old things with a greater sense of joy and cherishing.
We all could really, really do with some good news about now...
interpretability might be the most important scientific problem in the world right now, as I don't think we can solve alignment if we don't know what we're building. I think I have good news about it. I co-authored early work in mechanistic interpretability at OpenAI and came away pessimistic, because neural networks appeared to fold into arbitrary tangled knots. years later, I started openai's first large-scale automated interpretability project, finding about 2% of neurons legible with the primitive tools we had
with modern agentic tools and a much closer look, the picture looks totally different. the vast majority in models up to trillions of parameters are legible, and their circuits beautiful. many famous examples of tangled neurons turned out to be misunderstood. The roadblocks I expected mostly aren't there. I now think ambitious interpretability will be solved by the time we need it most, and possibly quite soon, as automated scientists keep improving. it may be a bitter lesson for interpretability: the simple thing, at absurd scale, works
already-clean.goodfire.com/
Agree, alarmingly so
I would definitely recommend you to give this a reread. Back in 2025, this was sci-fi.
It’s super super accurate.
ai-2027.com
