Caleb Eom (@calebfoundry) header
Caleb Eom (@calebfoundry) profile picture

Caleb Eom

@calebfoundry

over-analyze everything.

 USA
youtube.com/@CalebWritesCode
Joined March 2025
  • Tweets560
  • Following220
  • Followers2.4K
197Media
Photo by Caleb Eom (@calebfoundry)Photo by Caleb Eom (@calebfoundry)Photo by Caleb Eom (@calebfoundry)Photo by Caleb Eom (@calebfoundry)Photo by Caleb Eom (@calebfoundry)Photo by Caleb Eom (@calebfoundry)Photo by Caleb Eom (@calebfoundry)Photo by Caleb Eom (@calebfoundry)Photo by Caleb Eom (@calebfoundry)Photo by Caleb Eom (@calebfoundry)
What a crazy turn of events.
Evan (@StockMKTNewz) profile pictureEvan
@StockMKTNewz48m
JUST IN Jensen Huang just officially resigned as CEO of Nvidia $NVDA effective immediately
Image from Evan (@StockMKTNewz)
2
4
261
Does anybody know who actually first broke the news about Jensen's affair with Alice Huang?
27
3
146
36,565
so is ARC AGI dead now
2
11
1,710
catching up on JEPA research papers feels like reading a 10-year saga that Yann is unfolding
1
16
838
Caleb Eom retweeted
Mathieu (@miniapeur) profile picture
Mathieu
@miniapeurNov 6
Image from Mathieu (@miniapeur)
11
145
13
1,268
27,792
can you please explain the intuition behind why using multiple large target blocks somehow performs better over something seemingly trivial..? @ylecun
Image from Caleb Eom (@calebfoundry)
5
531
its hard to admit but i think im addicted to reading research papers. the dopamine hit from being in flow state trying to wrap my head around novel ideas is so addicting. shout out to my dealer arxiv for giving me the good stuff.
9
7
2
132
2,824
ok britney spears kinda slaps
5
368
Weight-space programming explained.
6
7
1
47
2,217
i created a open benchmark on "weight-space programming" and GPT-6 Astra scored the highest at 96.3 against baseline (500M PPO). runner up: GPT-6.1 Sol, Claude 5.5 my main curiosity was how other LLMs write weights from scratch rather than editing them or "reading them". so this benchmark only measures the model's ability write new weights from scratch. full disclosure, im pretty skeptical about this still - but i think it's crazy to just let this idea go by. we need the ecosystem to run with it and see its limits, and train foundation models to optimize this. github repo: github.com/eomcaleb/weight-s…
Image from Caleb Eom (@calebfoundry)
Rikiya Takehi (@rikiyatakehi) profile pictureRikiya Takehi@rikiyatakehiNov 2
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math. It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training. We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵 Blog: rikiyat.github.io/blog/weigh…
Image from Rikiya Takehi (@rikiyatakehi)
Readers added context they thought people might want to know
LLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math. rikiyat.github.io/blog/weight-sp… github.com/RikiyaT/lm_und… huggingface.co/Qwen/Qwen3.5-0… huggingface.co/RikiyaT/Qwen3.…
14
11
4
111
10,594
TIL judgement can be spelled judgment
358
this is such a well done and beginner-friendly intro to how CUDA works at the high level and will help mold your brain around how kernel engineers think about performance for any given problem. no code at all, just dark theme diagrams! link below
Image from Elliot Arledge (@elliotarledge)
15
56
1
879
26,579
CUDA explained from a first principles perspective. Turns out, it's more than just a collection of kernels but it's a huge ecosystem and a huge rabbit hole to dig into. @NVIDIAHPCDev @NVIDIAAIInfra
3
3
1
35
1,144
ok im now back to being stupid again. CUDA is a huuuuge rabbit hole. xitter.twitterapi.xyz/calebfoundry/status/21…
Image from Caleb Eom (@calebfoundry)
Caleb Eom (@calebfoundry) profile pictureCaleb Eom@calebfoundryOct 28
holy crap i finally understand CUDA.. finally at the peak of mount stupid, now heading towards valley of despair..
2
13
483
holy crap i finally understand CUDA.. finally at the peak of mount stupid, now heading towards valley of despair..
1
1
26
1,254
Gemini 4 Carbon is OP on cybersecurity.. so much for "pacing" frontier capabilities.
Image from Caleb Eom (@calebfoundry)
2
18
537
Gemini 4 Carbon released at $2/$10 introductory pricing while scoring higher than Opus 5.5.
Demis Hassabis (@demishassabis) profile pictureDemis Hassabis
@demishassabisOct 28
⚡Iterating at lightning speed! Gemini 4 Carbon is SOTA across key benchmarks and combines frontier capabilities with deep safety & security research.
Image from Demis Hassabis (@demishassabis)
2
13
510
Kimi K3.1 Pro explained.
15
528
it would be hilarious if Google just drops a monster model right before Anthropic IPO
4
1
30
1,005
yay
1
9
742
Load more