lucca bertoncini

21 posts

lucca bertoncini

lucca bertoncini

@lbz____

engineering at Fundamental AI Research (FAIR) @AIatMeta

Menlo Park, CA Katılım Şubat 2025
130 Takip Edilen75 Takipçiler
lucca bertoncini retweetledi
SkyPilot
SkyPilot@skypilot_org·
AI Infra Meetup with SkyPilot and CoreWeave - what a night! Packed room, great conversations, and solid talks on scaling AI infra across K8s, Slurm, and neocloud. 🌟 Highlights: @lbz____ (@AIatMeta) shared how Meta unified 100k+ GPUs across Slurm clusters with SkyPilot. @Michaelvll1 + @kevin_mingtarja (@skypilot_org) walked through SkyPilot's new features and ran a live multi-cloud demo. @deok_filho (@CoreWeave) broke down SUNK and the SkyPilot x CoreWeave integration. Huge shoutout to all speakers! Thanks to @CoreWeave @wandb for being amazing partners in making this happen. Excited to keep building this AI infra community. More events coming soon.
SkyPilot tweet mediaSkyPilot tweet mediaSkyPilot tweet mediaSkyPilot tweet media
English
0
3
19
4.4K
lucca bertoncini
lucca bertoncini@lbz____·
Giving a talk this Wednesday (3/25) in SF on SkyPilot at 100k+ GPU scale. Come through if you're around! luma.com/h52qyhmt
English
0
1
5
674
lucca bertoncini
lucca bertoncini@lbz____·
we wanna put claude in a humanoid bot, then prompt things like: > go touch the border of the universe, then come back to tell us what's outside /never-stop > what exists outside the simulation? /never-stop
English
1
0
0
38
lucca bertoncini
lucca bertoncini@lbz____·
they should add a mode to claude where it never gives control back to me. `--allow-work-forever` like yeah just go on that route and keep trying forever claude, if I notice something wrong I'll stop you
English
2
0
1
65
Dylan Patel
Dylan Patel@dylan522p·
$1,000 for whoever comes up with the best name replacement for InferenceMAX InferenceMAX 2.0 dropping soon but we have to rename it because HBO MAX sent us a cease and desist. We have all NVIDIA GPUs from h100 to GB300 on large MoEs with SOTA optimizations like Disagg PD tested
English
366
4
297
59.3K
lucca bertoncini
lucca bertoncini@lbz____·
1.2k lines of bash. tks claude, I would never
English
0
0
1
59
lucca bertoncini
lucca bertoncini@lbz____·
Legendary run from @pedroh96 and @hdubugras 🐐 Pagar.me was my first job out of high school, they took a chance on me when no other company in Brazil would.
victor lazarte@victoralazarte

Today, Capital One is acquiring Brex for $5.15B, the largest bank–fintech acquisition ever. Great CEOs build great companies. What most people don’t know is that by age 29, Pedro had built two different companies, each generating $500M+ in revenue, with his partner Henrique—something that had never been done before. I met Pedro and Henrique when they were 16, building Pagar me, which later became one of the largest fintechs in Brazil (inside NASDAQ: STNE). I knew very little about fintech then, but it was immediately obvious how exceptional they were. I’m grateful to have been Brex’s first board member and to have invested at a $10m seed valuation. What Pedro cares about most is to build things as they should be so the things we build can outlast us. Pagar me changed the Brazilian financial system and Brex + capital one will change the American. Thank you for having me on this 9 year ride, and thank you for exiting at over 500x the initial valuation I invested in. Cant wait to see all that you will continue building.

English
0
0
1
169
Miles Wang
Miles Wang@MilesKWang·
We're also publishing FrontierScience, a benchmark measuring expert-level scientific reasoning across physics, chemistry, and biology. We think AI-accelerated science is one of the most exciting promises of AGI, and are excited to build towards clear targets.
OpenAI@OpenAI

We’re releasing a new eval to measure expert-level scientific reasoning: FrontierScience. This benchmark measures PhD-level scientific reasoning across physics, chemistry, and biology. It contains hard, expert-written questions (both olympiad-style problems and longer research-style tasks) designed to reveal where models succeed and where they fall short. openai.com/index/frontier…

English
3
8
90
12.3K
OpenAI
OpenAI@OpenAI·
We’re releasing a new eval to measure expert-level scientific reasoning: FrontierScience. This benchmark measures PhD-level scientific reasoning across physics, chemistry, and biology. It contains hard, expert-written questions (both olympiad-style problems and longer research-style tasks) designed to reveal where models succeed and where they fall short. openai.com/index/frontier…
English
109
234
2K
949K
OpenAI
OpenAI@OpenAI·
Accelerating scientific progress is one of the most impactful ways AI can benefit society. Models can already help researchers reason through hard problems — but doing this well means testing models on tougher evaluations and in real scientific workflows grounded in experiments.
English
211
176
2.4K
374.9K
lucca bertoncini
lucca bertoncini@lbz____·
@skypilot_org is utopia for AI infrastructure, i̵f̵ ̵t̵h̵e̵y̵ ̵w̵i̵n̵ ̵t̵h̵e̵ ̵a̵c̵a̵d̵e̵m̵i̵c̵ ̵m̵i̵n̵d̵s̵h̵a̵r̵e̵ I'd bet they'll be everywhere in a few years thanks @zongheng_yang, @bromil101, @Michaelvll1 for spending some time with the Meta team!
English
1
3
11
1.4K
Zach Mueller
Zach Mueller@TheZachMueller·
Made a table of the most common/supported BF16 GPUs and their non-sparse TFLOPs. What's the best way to publish this? As a wiki on my blog? A pypi package to import?
Zach Mueller tweet media
English
51
30
557
294K
SemiAnalysis
SemiAnalysis@SemiAnalysis_·
GPU GO BRR
SemiAnalysis tweet mediaSemiAnalysis tweet media
Indonesia
11
3
153
18.3K
lucca bertoncini
lucca bertoncini@lbz____·
are `git clone` counts accurate? i'd understand if it was downloads, probably coming from CI. but clones? don't we all clone once and `git pull` forever? in what context would it make sense to run `git clone ...`? something feels off
lucca bertoncini tweet media
English
0
0
0
99
lucca bertoncini retweetledi
the tiny corp
the tiny corp@__tinygrad__·
America has once again put up roadblocks to us getting a visa. And I'm sure it's not just us. Every decision made like this slowly chips away at America's lead in AI. Most Americans have never lived in a country that wasn't rich, but this is the road to that.
English
36
18
714
43.3K
lucca bertoncini retweetledi
Taelin
Taelin@VictorTaelin·
How to prove that LLMs will lead to AGI (in 3 easy steps): 1. Prepare a dataset of math knowledge BEFORE 1500 2. Certify it doesn't include any mention to complex numbers 3. Train the model and ask it to find all roots of x^2+1=0 If the concept of complex numbers ever emerges from this experiment, I think this will be a very good evidence that LLMs will eventually invent new science
English
208
147
3.3K
271.5K

Keşfet

@AIatMeta @Michaelvll1 @kevin_mingtarja @skypilot_org @deok_filho @CoreWeave @wandb @dylan522p