Jimmy Ba
@jimmybajimmyba
Followers
18K
Following
86
Media
9
Statuses
59
Pumping MPI (mins per intervention) up
San Francisco, CA
Joined November 2022
based and 🔓 . wanna help accelerate the next Grok? . looking for builders:.— Rust/Jax/Kube infra engineers.— front-end/full-stack engineers.
127
86
397
theres but one singular metric for startups: reach escape velocity. team sus-column-r is absolutely cracked and is moving at a ludicrous speed. Grok-2 preview on @lmsysorg Chatbot Arena.
Woah, another exciting update from Chatbot Arena❤️🔥. The results for @xAI’s sus-column-r (Grok 2 early version) are now public**!. With over 12,000 community votes, sus-column-r has secured the #3 spot on the overall leaderboard, even matching GPT-4o! It excels in Coding (#2),
122
84
894
Truly 🙏 to be awarded Sloan Fellowship in Computer Science - credit to my students and colleagues at @UofT Looking forward to the journey ahead! #SloanFellow.
We are delighted to announce the winners of this year’s Sloan Research Fellowship! These outstanding researchers are shining examples of innovation and impact—and we are thrilled to support them. Meet the winners here: 🎉 #SloanFellow #STEM #ScienceTwitter
206
74
271
feels like investments in AI might as well be denominated in their net present inventory of hot chips, rather than dollars.
264
49
170
The biggest surprise of 2023: an LLM self-improving workshop remains elusive.
263
47
94
hodl action entropy, 💎🙌.
Excited to announce DreamerV3 🌍, a scalable and general RL algorithm that masters a wide range of applications with fixed hyperparameters! . Applied out of the box, it solves the Minecraft Diamond challenge without human data. 💎. 👇 Thread.
156
57
57
In my 2nd year of grad school, I asked @geoffreyhinton what he’d do after winning a Turing or Nobel Prize. Geoff told me the fabled story about Tim Hunt, a Nobel prize winning biologist. The day after winning the Nobel prize, Tim told his students he’d be less involved in their.
17
21
632
Half gen at a time. Look for more XLA experts at
27
38
209
Come hangout with us in sf next tues!. Been doing neural nets for a decade now. Vibe check tells me this is a very special moment in ai when we look back in another decade. What innovations will define the next asi chapter?.
We are excited to bring together a group of exceptional engineers and product builders who are intrigued by our mission to build maximally truth-seeking AI. Join our open house to meet our team, learn more about xAI, and enjoy a fun evening brought you by the creators of the
22
27
457
Show me the chips.
xAI is pleased to announce.
21
26
210
Who is making a tax filing eval? Want to know which model to use. Extra points for rebate optimization strategies.
4
25
70
training data is just crystallized computations of the past, a fossil of someone else’s hard work. now feels like creating these fossil are easier than it should be tho.
39
58
291
Many epic launches this week. This one literally is a slam dunk from space. Caught up with some starbase engineers at after parties. These folks are insanely talented, humble and pure focus. Just heads down, pushing hard. Inspiring.
4
26
272
throwback thurs to two yrs ago at neurips:. pounded warm beer in an epic flunkyball match with world champ, played jenga with homies training big models for fun, then it hit, yall know whats missing? another agi company!. peak pre-agi vibes.
NeurIPS 2022. Here is a photo of @dmdohan @_sholtodouglas @jimmybajimmyba 1 day after the release of ChatGPT, before post-ChatGPT capitalism tore them apart to OpenAI, DeepMind, and xAI. Post-AGI is when they can play Jenga again😔. I'll attend NeurIPS 2024 and am excited to meet
0
56
113
our ai community = ultimate mesa-optimizer in watts per progress on benchmarks.what evals shall we use for grok 3?.
Chatbot Arena update❤️🔥. Exciting news—@xAI's Grok-2 and Grok-mini are now officially on the leaderboard!. With over 6000 community votes, Grok-2 has claimed the #2 spot, surpassing GPT-4o (May) and tying with the latest Gemini! Grok-2-mini also impresses at #5. Grok-2 excels in
7
18
178
3
20
166
Feels like reality still has acreative edge over the current genAI.
14
16
61
Excited to release our first assignment with ChatGPT-assisted learning! Put your critical thinking skills to the test and verify the generated solutions. P=NP? .@YangjunR DennyWu
7
21
118
It is no longer about understanding the rules of multiplication but also about learning how to use a calculator.
2
6
45
What are some common misconceptions about LLMs? Need some ideas for a general audience talk.
19
2
31
Are we living in an LLM simulator?.
Did you know, that you can build a virtual machine inside ChatGPT? And that you can use this machine to create files, program and even browse the internet?
1
2
24
@astro_nolan Yes! We found this with our internal testers too. By default, users assume the model is always “online” and has the same capability as a human with a computer.
1
0
3
@perryzjia @ibab_ml On it. Imagining landing a goofy joke from a yr ago. Then the next feature request will be an incognito mode.
1
0
2
@CarlSparre I’m with David Mackay here: especially section II. Need to bootstrap our civilization.
2
0
1
@AmirRezaPeimani @geoffreyhinton @ylecun Models at today’s scale cares more about getting the character names right than generating interesting plots in a Harry Potter novel.
0
0
2