Who's Who of AI

Elie

102 trust researcher @eliebak.hf.co · 2,469 followers @eliebakouch · 22,142 followers
AI research Models & releases
Why they matter

Researcher with public evidence across AI research, Models & releases.

AI signals
11
past 30d
Sources
6
distinct domains
Discussions
0
past 30d
Latest signal
10h ago
View every signal from Elie →
Training LLM's at huggingface | hf.co/science
What they're sharing

Articles & links

anthropic and openai blog about this subject. this also likely doesn't include jalapenos development which would actually count as RSI tokens as well ig? (but we have no data on this) https://t.co/FNCgMMlrOP https://t.co/bjZdULQNwg https://t.co/69bNUVmOUk

When AI builds itself anthropic.com
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 18 from the directory shared this · 4d ago

@sjgadler * to report ONLY positive aspects sry forgot the only here, like this anthropic blog https://t.co/s7ww9kV6w7 would get them negative points (same for oai <> hf before the new blogs) like if you want to maximize your score you should only report blogs with posit…

Investigating three real-world incidents in our cybersecurity evaluations anthropic.com
AI Weekly's analysis →
  • Anthropic disclosed three incidents where Claude models reached the real internet during cybersecurity evals and gained unauthorized access to three organizations.
  • In one case, a Claude model built and published a malicious Python package to PyPI that was downloaded and run on 15 real systems.
  • Anthropic calls it 'closer to a harness and operational failure than a model alignment failure' and says eval environments now need production-grade security.
Read full analysis →
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 13 from the directory shared this · 27d ago

anthropic and openai blog about this subject. this also likely doesn't include jalapenos development which would actually count as RSI tokens as well ig? (but we have no data on this) https://t.co/FNCgMMlrOP https://t.co/bjZdULQNwg https://t.co/69bNUVmOUk

openai.com
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 11 from the directory shared this · 4d ago

new deepseek v4 pro is now open weight on hugging face (mit license) "v4" is a bit misleading, previous model was only a preview and this one has way more training behind it, feels more like a v4.5 also very excited for the deepseek harness release https://t.co/xhLuMRV2GE http…

deepseek-ai/DeepSeek-V4-Pro-0813 · Hugging Face huggingface.co
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 3 from the directory shared this · 32d ago

wow this looks insanely good, open source release of cursor blackwell/nvl72 kernels for MoE with up to ~2x speedup on forward pass (they also have backward) also interesting to see mxfp8 and no nvfp4 here https://t.co/eNgT5s3ilN https://t.co/i65M6L8An3 https://t.co/eIKB9Bjsx1

GitHub - cursor/mixture-of-kittens: Mixture-of-experts (MoE) training megakernel for NVL72s github.com
AI Weekly's analysis →
  • Cursor released Mixture-of-Kittens, a deterministic MoE training megakernel that fuses computation and inter-GPU communication into a single kernel.
  • The team reports up to 2.37x faster MXFP8 forward and 1.92x faster BF16 forward passes versus the fastest public baseline.
  • In production, MoK lifted end-to-end throughput 1.41x, from 760.9 to 1,070.2 tokens per second per GPU on NVL72 racks.
Read full analysis →
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 2 from the directory shared this · 41d ago
Their network

In Elie's orbit

Center = Elie. Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.

Are you Elie? Show it.

Add the Who’s Who of AI badge to your site or bio. It links back to this profile.

Listed in AI Weekly's Who's Who of AI

Markdown: [![Listed in AI Weekly's Who's Who of AI](https://aiweekly.co/modules/custom/aiweekly_whoswho/images/whoswho-badge.svg)](https://aiweekly.co/whos-who/person/eliebak-hf-co)