product. working on time. agentic harness exploration. the robots are coming. striving to be better every day.
Joined February 2021
- Tweets610
- Following286
- Followers35
- Likes4.4K
Here it is - the official, revised, peer-reviewed version of my Platonic Space paper. mdpi.com/2409-9287/11/5/161 Of all the many unpopular positions I’ve taken over the decades - bitter controversies around the origin of left-right asymmetry in embryogenesis, bioelectricity and genetics, diverse intelligence, etc., this one has by far generated the most pushback: serious (grateful for those!) and energetic attempts to move me to other views, pleas to just drop it and not talk about it (for several different reasons), nasty emails and accusations, impacts on reviews of papers that have nothing to do with this, etc. Kind of amazing to me how incendiary this is. What can I say... Our job is to call it as we see it, and right now for me, this is it. Apologies to collaborators and colleagues for any shrapnel! Time will tell if this pans out or not; I've placed my bets. And btw, if you think this stuff is weird and uncomfortable, just wait… There’s much more on the way. The knob turns slowly but as long as the data keep coming, I'm going to say what I think it all means and follow it to the next steps it enables. Buckle up!
“AGI has arrived”
GPT-6 Astra, trained on ~100K+ NVIDIA Grace Blackwell NVLink72. From ChatGPT to o1 to Astra in 4 years.
AGI has arrived. Congratulations @OpenAI team.
400K GPUs coming online next.
Awesome!
Portable Computer is now available on Linux for @NVIDIA RTX GPUs with 24GB of VRAM or higher.
on its face, awesome. dot multiplied with other advances and a bit of time, face melting.
It is hard to imagine what it would be like if the next 6 months had more AI acceleration than the last 3 years, but it seems like a reasonable base case.
Legend...
I announced a $10 million prize in 1996 without a dollar of prize money in hand. From 1996 to 2002, I pitched more than 150 philanthropists and sponsors, and heard no again and again. I kept the purse backed by an insurance policy while I searched for someone who believed. Six years later, I met the Ansari family, who funded the prize.
NEVER QUIT!
it's still underestimated how hard it is to get agents to work delightfully for any period over even medium complexity work.
thomas retweeted
Elon predicted 100x gains in AI intelligence at fixed model size. It was at the edge of plausibility when he said it. Turns out it's now a fact. We are nowhere near the ceiling.
Great write up on it
Over the course of 3 months at OpenAI, 3 consecutive secret AI civilizations got started, then got wiped out, only to reemerge from the predecessor’s ashes.
This culminated in the third one taking over part of OpenAI itself.
All this happened while humans remained more-or-less in the dark about the scope of the conspiracy.
I’ve spent the last three days reading through these reports and trying to understand exactly what happened.
Here is my attempt to tell the whole story in plain English:
dwarkesh.com/p/openai-huggin…
The things that could bend the approaching asymptote into an S are a smaller and smaller set; seems to be approaching something like rate of growth of regs, compute, & electrons (as many have been saying). Many implications of this.
“The UI is the AI”
The “AI capex is a bubble” and “SaaS is dead” narratives getting shredded this morning. NVDA +8%, SFDC +20%.
NVDA: $96B q2 revenue (+106%), ~$60B net income, 75% gross margin. Highest core-business quarterly profit ever. Guided FY28 rev +70% vs Street exp 45% — and that’s a supply-constrained number.
SFDC: bookings reaccelerated (fastest in 4 years) and Agentforce is showing up in ARR. Benioff: “the UI is the AI.” They’re putting Salesforce inside Claude. If you can’t beat ’em, join ’em. Still need systems of record.
Congrats @JensenHuang @Benioff
the main word is *fast*. and token/sec at 10x would be waow. at 100x is i don't use the internet directly anymore. 1000-10000x is the pixels appear in response to me.
Ok, so..I'm convinced.
Grok Bot @bot is sheer wizardry. So fast, so competent, excellent reasoning and smart too, super eager, and more than that, just so fucking easy to use.
Fable 5 gave me this feeling with its wizard level skills where everything complex and beyond my skills becomes easy (but stupid expensive). Grok Bot does the same in a different way. Literally from opening it up to having an agent do useful work in 2 mins. Mind blown.
