Clive Chan (@itsclivetime) / X

Clive Chan

2,318 posts

Clive Chan

@itsclivetime

perplexity per picojoule @anthropicai // prev @openai @tesla

California, USA

Joined May 2012

Pinned
Clive Chan
@itsclivetime
Jun 6
Personal update: I’ve decided to leave OpenAI. I’m proud to have been part of the custom chip program and grateful to everyone I got to build with and learn from along the way. The density of hardware talent on that team is extraordinary, and I don't think there's a better chip
2.8M
Clive Chan
@itsclivetime
Nov 14, 2024
wtf is in apple silicon? my single threaded code is 3x faster on macbook than on server while burning like half the power. if apple bothered making a 64-core part they'd take over every datacenter
1.1M
Clive Chan
@itsclivetime
Aug 6, 2025
ok but wen gpt-6
OpenAI
@OpenAI
Aug 6, 2025
LIVE5TREAM THURSDAY 10AM PT
166K
Clive Chan
@itsclivetime
Sep 17, 2025
this guy designs kernels with spreadsheets to hit shared memory banks evenly, open sourced assembly kernels that basically taught nvidia how to do gpu matmuls and register allocation properly, and authored most prod chatgpt kernels. genuinely incredible
Rohan Pandey
@khoomeik
Sep 16, 2025
there was ~1 guy at openai responsible for inference CUDA kernels let’s call him Bob, people would refer to his attention kernel as “the Bob kernel” it executed probably trillions of times a day on hundreds of thousands of GPUs one singular guy
180K
Clive Chan
@itsclivetime
Apr 9, 2025
Google's TPUv7 is out! ML accelerator marketing material is usually pretty inscrutable (what numbers are even comparable?), so here I'll explain concretely how this compares with Nvidia. 🧵
448K
Clive Chan
@itsclivetime
Sep 18, 2025
my rolexes are silicon my porsche is a gpu
Jackmin
@jackminong
Sep 18, 2025
Replying to @jackminong and @shumochu
its also great as a vanity item. the ML engineers equivalent of a rolex. everyone good will have one and you get to complain about it at parties as some form of humblebrag
91K
Clive Chan
@itsclivetime
May 15, 2023
been reading the readme for github.com/microsoft/guid…, kind of galaxy brain tl;dr they made a whole prompt engineering language
GitHub - guidance-ai/guidance: A guidance language for controlling large language models.
From github.com
262K
Clive Chan
@itsclivetime
Oct 13, 2025
Really happy to be announcing the chips we’ve been cooking the past 18 months! OpenAI kicked off the reasoning wave with o1, but months before that we’d already started designing a chip tuned precisely for reasoning inference of OpenAI models. In January 2024, I joined OpenAI as
OpenAI Newsroom
@OpenAINewsroom
Oct 13, 2025
We're partnering with Broadcom to deploy 10GW of chips designed by OpenAI. Building our own hardware, in addition to our other partnerships, will help all of us meet the world’s growing demand for AI. openai.com/index/openai-a…
339K
Clive Chan
@itsclivetime
Nov 10, 2024
Same. Since joining in January I’ve shifted from “this is unproductive hype” to “agi is basically here”. IMHO, what comes next is relatively little new science, but instead years of grindy engineering to try all the newly obvious ideas in the new paradigm, to scale it up and
Noam Brown
@polynoamial
Nov 9, 2024
I've heard people claim that Sam is just drumming up hype, but from what I've seen everything he's saying matches the ~median view of @OpenAI researchers on the ground.
206K
Clive Chan
@itsclivetime
Apr 14, 2025
4.1 has basically replaced o3-mini for me in all my workflows (cursor, etc.) - highly recommend also lol at nano just hanging out there being 2x better than latest 4o at math
275K
Clive Chan
@itsclivetime
Jun 21, 2023
Tesla already builds some of the largest GPU clusters in the world and now we’re going to >10x that. Next year, Dojo’s compute production rate will be on the same order as Nvidia’s.
Tesla AI
@Tesla_AI
Jun 21, 2023
Replying to @Tesla_AI
And will be trained on enormous amounts of compute
114K
Clive Chan
@itsclivetime
Sep 29, 2024
funny how the best way to get good at a job is to just spend a lot of time on the job i attended ~0 classes after first year and averaged <70% in my last year. instead i spent time on 7 internships and a couple student teams. guess what ended up mattering?
Greg Yang
@TheGregYang
Sep 29, 2024
91K
Clive Chan
@itsclivetime
Sep 14, 2024
Replying to @karpathy
on the other hand maybe everything that you can express autoregressively is a language and everything can be stretched out into a stream of tokens, so everything is language!
78K
Clive Chan
@itsclivetime
Feb 20, 2025
played with grok 3, it's really good - insanely rapid progress from xai team - super quick models, in particular deep research is like 10x faster than oai (but a bit less thorough) - a bit worse + much slower at coding than o3-mini-high - crashed for my super long query but
97K