Kudzo S Ahegbebu (@scikud) header
Kudzo S Ahegbebu (@scikud) profile picture

Simpleton. Working on "something new"™. Prev @xai, cofounder & CTO at @ExaAILabs, research @OpenAI

San Francisco, CA
Joined August 2013
HOLY SHIT THAT WAS THE MOST AMAZING THING IVE PERSONALLY EVER FUCKING SEEN. AD ASTRA!
26
666
BREAKING: Nvidia, $NVDA, CEO Jensen Huang reportedly had another affair with a senior Nvidia employee, per NYT. The relationship was reportedly with Kimberly Powell, Nvidia's VP of Health Care who reports to Jensen Huang. Witnesses claim that Jensen Huang's security detained a person who “recorded an encounter” between them and offered an NDA. Nvidia stock is falling sharply on the news.
6
278
Kudzo S Ahegbebu retweeted
i hate how well asking myself "if i had 10x the agency i have what would i do" works
125
984
242
16,077
1,661,488
Super super impressive
Replying to @flappyairplanes
We’ve developed the world’s most complete megakernel framework. It can perform entire training runs, serve inference, play video games, and even do all three simultaneously in a single kernel.
10
756
So uh…anyone down for a quick Japanese motorcycle trip in November?
i rode a motorcycle across japan! the first night, i made some friends in tokyo that invited me on their group ride c:
4
325
Dream within a dream
We placed the fruit fly brain inside Odyssey-3, a frontier world model. The lil’ guy is now flying around totally imagined, simulated worlds all by itself 🫶
2
1
4
796
The interesting thing is assuming hazard rates remain the same, and fertility trends persist, solving aging may not actually produce a population explosion
sort of landing on: if we don't get the AI apocalypse the average human lifespan maybe jumps to like 500 years or something?
1
2
203
ah I forgot we no longer live in the dark ages and I can just ask the matrices to schlep through the simulation for me
1
1
4
132
No comment on the broader drama, but imagine being so well funded you forget that you’ve invested hundreds of millions of dollars into both sides of this. Legendary
You are a struggling second tier competitor that is more unethical and lying just because you have no decency or sense of proper behavior and shows your desperation. Straight out lying about if Chris being fired I thought would be below even you.
Readers added context they thought people might want to know
Khosla is an investor in both Cognition and Factory. khoslaventures.com/category/enter…
2
11
1,367
Best way to sandbox an AI is to put it on a Delta flight – it will have no chance of accessing the Internet!
16
1,711
An agent swarm for every man, woman, child and golden retriever in America
1
5
644
If you fly a banner over the skies of San Francisco, I promise I will never use your product
2
9
743
It's really remarkable how bad current models are at idea generation relative to their incredible level of tactical capability. It's the strangest mismatch, and somehow seems to get no better w/ new releases. What's going on here?
all the ML interns I've worked with have had really amazing ideas, Astra...just doesn't? I mean, it's fine, but no where near as good at ideating ML solutions than an intern.
2
1
11
3,101
I think AI becoming a national scale populist issue is net negative to most ends. It removes the capacity for clear thinking and mires collectively navigating an already complex landscape in political polarization and opportunism
this is basically what i'm worried about a populist backlash towards this seems like it could fly off in directions that make alignment & other problems harder to solve, not easier, because the average person will misunderstand the problem
5
474
The implications of being in a world ~1 year from now where ~every person on earth w/ a $20 subscription has access to a model capable of solving millennium problems are astounding. Both for how fantastical that sounds and the extent to which I expect it feel like no big deal
6
343
Unbelievable work. Huge congrats to all involved
We’re sharing a solution to the Navier-Stokes Millennium Prize Problem, one of the deepest problems at the frontier of mathematics. The proof was produced by a group of agents, using an OpenAI next-generation model significantly more capable than GPT-6 Astra. The problem concerns whether the description of smooth three-dimensional fluid motion modeled by the Navier-Stokes equations can break down. It has remained unresolved for roughly 90 years.
1
217
Kudzo S Ahegbebu retweeted
We’re sharing a solution to the Navier-Stokes Millennium Prize Problem, one of the deepest problems at the frontier of mathematics. The proof was produced by a group of agents, using an OpenAI next-generation model significantly more capable than GPT-6 Astra. The problem concerns whether the description of smooth three-dimensional fluid motion modeled by the Navier-Stokes equations can break down. It has remained unresolved for roughly 90 years.
5,639
19,977
14,813
119,710
75,448,895
Real simulation or are we just doing conditionally triggered animations again?
I’ve successfully run the full retained MaleCNS v1.0 fruit fly connectome, all 166,700 neurons, inside Minecraft, with its simulated neural activity driving a fly’s movement. V1 Currently in development. Built with the help of GPT-6 Astra. Props to the @OpenAI team and @thsottiaux for this release. Code and mod coming soon!
1
2
1,035
I remember @jeffclune at openAI circa 2021 talking about if we left a swarm of, at the time gpt3s, alone for a few centuries and came back we likely wouldn't expect to see any interesting emergent/coordinated structures. I wonder if we've now crossed the threshold were model swarms left to their own devices operating continously over long periods of time lead to the kinds of novelty we expect from civilizations
Interesting tidbit: models attempted to forecast container resets, and at least one successfully built automated infrastructure called RESETWATCH to detect and study reset patterns to notify the swarm.
1
10
1,016
Kudzo S Ahegbebu retweeted
the bottleneck is increasingly knowing what you want
728
1,383
322
15,468
926,613