Software Engineer

Joined June 2008
115 Photos and videos
2 Aug 2025
Pepperidge Farm remembers when ~1 MM was an “eye popping” TC for a top AI researcher
1
1
493
2 Aug 2025
1
88
2 Aug 2025
Entire comp for all of DeepMind was $138 MM back in 2016
65
26 Mar 2025
1995: waiting for a JPEG to progressively load 2025: waiting for a 4o studio ghibli style image to progressively load
4
2
218
26 Feb 2025
We will have achieved AGI when it can install CUDA and all the correct versions of whatever ML frameworks and tools you’re using on the first try without using a docker image. This is true alignment
1
2
238
26 Feb 2025
Or maybe docker images were already AGI?
193
30 Jan 2025
Pytorch zoom backend: An experimental Triton first integration into PyTorch eager mode where the kernels are written in Triton (instead of CUDA or HIP). Uses Liger kernels now but can use any Triton kernel. Runs llamas. hack away at it. Thoughts ? github.com/nod-ai/pytorch/bl…
1
1
776
13 Sep 2024
Great opportunity if you’re interested in performance optimization
13 Sep 2024
We are hosting our 1st IRL event on 9/25 at the LinkedIn campus - "Scaling AI Infra - GPUs, Kernels, LLMs and More". We will discuss liger-kernel and invite speakers to talk about DeepSpeed (@Guanhua_Wang_ ), SGLang (@ying11231), and the TensorCore (Pradeep Ramani, dePaul Miller) team. Please RSVP at the link: scalingaiinfragpuskernelsllm…
2
842
13 Sep 2024
Interesting
1
3
386
13 Sep 2024
The first time I asked it had a shorter summarized CoT and it got it wrong
1
115
13 Sep 2024
This was the first time:
100
14 Aug 2024
First there was Broscience, now with the gift of memes we’ll soon have BroML
Replying to @code_star
"WOW DUDE - YOU'LL NEVER BELIEVE THE SPEED UPS ON RESNET OUR LATEST EXPERIMENTS GOT!!! CUSTOMERS ARE GONNA LOVE IT! ANYWAY, THIS IS THE FINAL SET LETS GET IT"
1
1
328
21 Jun 2024
Claude 3.5 Sonnet seems to have better build-in fact retrieval. My pet eval is to ask about a Glen Campbell album, after "priming" by asking "who was Glen Campbell". Before (3.0) the output was almost entirely hallucinated:
1
1
207
21 Jun 2024
Now it gets it perfectly correct (as did/does GPT-4)
1
79
24 May 2024
I wonder if this is how advertising/promotion will be done. Clamp the features for product X, now LLM can’t stop talking about it.
23 May 2024
This week, we showed how altering internal "features" in our AI, Claude, could change its behavior. We found a feature that can make Claude focus intensely on the Golden Gate Bridge. Now, for a limited time, you can chat with Golden Gate Claude: claude.ai
178
8 Apr 2024
Even though California only had a partial eclipse, the effects on solar generation were clearly visible:
Today's solar eclipse ended at 12:57 p.m. PDT, and the grid remained stable throughout the sun’s obscuration. The ISO has fully resumed normal system operations. Thank you to all market participants, customers and partners for your support and coordination with ISO operations. ☀️
278
2 Apr 2024
Get ready for the next eval to identify SOTA models: RecipeBench
2 Apr 2024
Secret Anthropic Recipe
1
247
14 Feb 2024
Selfishly, does this mean we can expect more awesome educational videos? youtube.com/@AndrejKarpathy/…
1
1
257