Sigrid Jin | Jin Hyung Park 🪄 👧 @sigridjin_eth

✯ @thisissigrid ★ ☄  CS @UBC @ubcokanagan ☄ ★ Machine Learning Ultrathink Engineer @sionic_ai 🐟 digital nomad with Python, Golang & Rust 💻 github.com/sigridjineth The Earth Joined July 2019

Tweets

12K
Followers

2K
Following

7K
Likes

24K

Lucas Beyer (bl16) @giffmana

3 days ago

Goated FAIR team just found how coding agents sometimes "cheat" on SWE-Bench Verified. It's really simple. For example, Qwen3 literally greps all commit logs for the issue number of the issue it needs to fix. lol, clever model. "cheat" cuz it's more like env hacking.

Bram Wasti @bwasti

3 days ago

7 20 299 127K 84

24 49 783 111K 273

Download Image

Marcin Krzyzanowski @krzyzanowskim

2 days ago

these Chinese AI are undervalued. I switched from Claude to z.ai (using Claude Code CLI), and guess what, I noticed no difference in daily use. Except it's $3 instead $200 there is world outside OpenAI and Anthropic, that is better than you may think.

320 723 12K 1.0M 9K

Download Image

Sebastien Bubeck @SebastienBubeck

a day ago

more and more such stories, really nice to see!!

Andrej Karpathy @karpathy

a day ago

more and more such stories, really nice to see!!

405 738 12K 2.0M 2K

10 4 172 22K 7

wh @nrehiew_

a day ago

Of all the python dependency problems, half are solvable simply by using uv, the other half are impossible to solve because they involve numpy 2.0

3 2 59 3K 5

机器之心 JIQIZHIXIN @jiqizhixin

20 hours ago

OpenAI just claimed to have found the culprit behind AI hallucinations.

139 392 5K 483K 3K

Download Image

Sharon Zhou @realSharonZhou

2 days ago

More data ≠ better fine-tuning. Sometimes the 'undertrained' model is more useful because it's still plastic. We need to map the capability vs. adaptability frontier better.

14 19 208 42K 54

Neel Nanda @NeelNanda5

a day ago

I was impressed with how easily Julian made an agent here, creating a working MVP in just a day!! We are now in the age of automated interpretability, and there's a lot of low hanging fruit - please pick it! Remember, interpretability agents are the worst they will ever be

Julian Minder @jkminder

2 days ago

5 19 186 31K 152

Download Image

3 12 168 13K 119

Patrick Collison @patrickc

a day ago

Someone on HN (which is often quite crypto-skeptical) asked why exactly businesses are finding crypto and stablecoins useful. It's a very reasonable question. My response:

173 364 3K 562K 3K

Download Image

Marzieh Fadaee @mziizm

2 days ago

I'm excited to share that I'll be stepping into the role of Head of Cohere Labs. It's an honor and a responsibility to lead such an extraordinary group of researchers pushing the boundaries of AI research.

72 33 696 68K 107

Download Image

andrew gao @itsandrewgao

a day ago

i had to prompt inject the @united airlines bot because it kept refusing to connect me with a human 🧵 what led up to this breaking point

254 1K 28K 2.6M 9K

Download Image

Andrej Karpathy @karpathy

a day ago

I think congrats again to OpenAI for cooking with GPT-5 Pro. This is the third time I've struggled on something complex/gnarly for an hour on and off with CC, then 5 Pro goes off for 10 minutes and comes back with code that works out of the box. I had CC read the 5 Pro version…

405 738 12K 2.0M 2K

Sam Altman @sama

a day ago

thanks andrej! do you care more about it getting smarter or faster?

Andrej Karpathy @karpathy

a day ago

thanks andrej! do you care more about it getting smarter or faster?

405 738 12K 2.0M 2K

856 217 6K 908K 666

Justin Johnson @jcjohnss

a day ago

10 years ago, deep learning was in its infancy. PyTorch didn't exist. Language models were recurrent, and not large. But it felt important: a new technology that would change everything. That's why @drfeifei , @karpathy, and I started @cs231n back in 2015 - to teach the world's…

38 178 2K 128K 1K

Binyuan Hui @huybery

a day ago

🚀 Qwen-Max has successfully scaled to 1T parameters, and we’re still pushing further. Hopefully this giant will bring some surprises, see you next week!

Qwen @Alibaba_Qwen

a day ago

🚀 Qwen-Max has successfully scaled to 1T parameters, and we’re still pushing further. Hopefully this giant will bring some surprises, see you next week!

242 495 4K 605K 700

Download Image

57 85 1K 70K 104

Jihyeon Kim (김지현) @simnalamburt

3 days ago

DriftSort github.com/Voultapher/dri…

0 2 5 745 0

Luc Georges 🦀 @LucSGeorges

4 days ago

An async API landed in tokenizers==0.22.0, pyo3-async-runtimes is 🔥 Thanks @feilsystem who contributed a neat PR to tokenizers to add the asynchronous bindings! This should make things faster in specific contexts, looking at you inference 👀

0 4 11 584 2

ICONIQ @ICONIQCapital

4 days ago

AI’s defining chapter has begun. Already trusted by 300,000+ customers worldwide, we believe @AnthropicAI's Claude models are setting a new standard for frontier intelligence – combining world-class performance with the safety and reliability enterprises demand. We’re proud to…

5 8 66 16K 8

Download Gif

Anthropic @AnthropicAI

4 days ago

We've raised $13 billion at a $183 billion post-money valuation. This investment, led by @ICONIQCapital, will help us expand our capacity, improve model capabilities, and deepen our safety research.