@MillionInt

CEO and co-founder of @coreauto former VP of RL @ OpenAI : reasoning models, o3, o1, GPT4, ChatGPT, Codex, RL for robots cautious AI optimist

 San Francisco, CA
millionintegrals.com
Joined January 2013
  • Tweets2.1K
  • Following1.2K
  • Followers43.4K
149Media
Small step for a finger on "post tweet" button. Big step for millions of future AI agents doing useful work for us.
Core Automation
@coreautoApr 21
Today we're announcing Core Automation Our objective: systems that optimize and automate work, starting with research itself.
67
37
6
726
248,830
It’s amazing you can clearly tell exact moment in history when each lab started vibecodong their product
19
3
1
553
30,131
The interesting thing is that at this moment those are in the training data
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)
@teortaxesTexOct 17
Gary Marcus, comments??
3
1
1
109
8,796
Generalization in 2026: we trained a chonkers on the energy consumption of London and it is good in some things
Jiayi Pan@jiayi_pirateOct 12
A modern perspective on generalization
4
125
8,247
🤖 Made with AI
9
27
2
520
17,303
Will replacing the transformer be a revolution or an evolution?
43
9
6
315
29,826
Playlist this week: - upping my p(doom) - chasing the roofline - core auto thesis (internal)
8
112
9,169
It is our responsibility to make reality interesting, as otherwise humans will rationally choose to live in a simulation
20
12
5
206
11,570
When I grow up I want to be a slopcore artist
8
2
2
89
13,792
Machine learning isn't magic - its assumptions in disguise
5
2
1
128
7,383
Every epoch is defined by its art. at the end of times we have the slopcore
𝞍 Shin Megami Boson 𝞍
@shinbosonSep 26
if you missed the slightly insane AI safety schizopoast, I put it into a song for you and you should turn the sound on right now because we're so fucking back
4
4
1
78
10,326
Big model smell lasts just a brief moment Small model smell lasts forever
12
3
1
249
9,913
1800s: state has monopoly on violence 2030s: state has monopoly on ASI
7
3
86
6,090
Oh, hi!
Samip
@industriaalistSep 22
We think computational depth is the missing scaling axis, i.e. we should be doing a lot more deep learning! Every other axis has been scaled by OOMs over the past few years (params, data, sparsity, test-time reasoning), but depth has been stuck at ~100 layers since GPT-3. We've found that LLMs are both *severely* depth-bottlenecked and bad at using the depth they have, and that architectural interventions that lift this bottleneck efficiently lead to gains that increase with compute. w/ @akshayvegesna
6
8
1
239
29,923
It’s a beautiful number
2
1
29
10,770
Yes
This quoted post is unavailable.
Try viewing anyway · View on X
1
1
42
6,844
real deep learning has never been tried
19
5
1
307
16,245
The road to singularity will be paved with blackberries of our generation. It is easier than ever to achieve a glimpse of growth and success but building a long term viable company is as hard as it ever was.
6
7
1
168
8,265
This was once revealed to me in a dream
5
4
112
8,302
More questions are deep learning questions than people realized. We just stopped asking deep learning questions and learned to work around them by tuning hyperparameters.
12
21
3
295
20,720
We may have done it at @coreauto but the model wasn't very good at math
Rosmine
@rosmineSep 16
We need to be dog maxxing our models There is 0% chance of doom if the chain of thought is user: hello model: Oh my god it's a human. I love humans. I will say hi to him, maybe he will give me a problem to chase. I love chasing problems. I will chase the problem for him then he will love me and I will be his best friend. Must help human model: hello! what can I help you with today?
4
76
13,965
Load more