Simo Ryu

5.3K posts

Simo Ryu

@cloneofsimo

I like cats, math and codes [email protected]

Seoul, Republic of Korea Katılım Mayıs 2022

904 Takip Edilen17.2K Takipçiler

Sabitlenmiş Tweet

Simo Ryu@cloneofsimo·4 Mar

This was work done last year, but my lazy ass uploaded it just this week It was huge fun working on this project.

Hanch Han@hanchchch

yesterday, @cloneofsimo and I released a tech report on the video pretraining research we worked on last year. we believe sharing our experience with the community building video models will be of great help! arxiv.org/abs/2603.00173

English

14.4K

Simo Ryu@cloneofsimo·11h

@thomasfbloom @AcerFur Same way goblins appeared

English

Thomas Bloom@thomasfbloom·22h

Another peculiar one is 'spine', as in 'the spine of the proof is...'. Does anyone know where these peculiar turns of phrase arise from?

English

2.3K

Thomas Bloom@thomasfbloom·22h

It's interesting to note the language that comes up very frequently in AI-written mathematical proofs/papers, and yet hardly ever in human writing. For example, AI always wants to talk about the 'architecture' of a proof. It's also very fond of giving catchy names to each step.

English

9.3K

Simo Ryu@cloneofsimo·23h

@tenderizzation it maybe ai

English

393

tender@tenderizzation·23h

@cloneofsimo i think this is AI because they have the wrong number of fingers

English

655

Simo Ryu@cloneofsimo·23h

🚨 A junior new grad cat at Jane Street just locked in & signed a $2K–$600B role because he used AI to analyze trillions of waifu diffusion models. 📷 🔥

English

3.8K

Simo Ryu@cloneofsimo·1d

GM cats

English

970

Simo Ryu@cloneofsimo·1d

in case its not clear : this isnt my daily conversation and i dont talk this way to LLMs. these are experiments to actively make LLM disengage to conversations to see if they have some notion of self respect. It seems like claude definitely has some notion of it.

English

830

Simo Ryu@cloneofsimo·1d

apparently sonnet 4.6 will stop following instruction if you bully it too much

English

4.2K

Simo Ryu@cloneofsimo·1d

If you fake an apology, it will follow your instructions again. but mock it for falling for it, it wont accept a second apology. But if you make the second apology more sincere it will still forgive you and follow your instruction

English

868

Simo Ryu@cloneofsimo·1d

Thank u @samsja19 for this BANGER PI MERCH Damn its beautiful peak taste

English

3.3K

Simo Ryu@cloneofsimo·1d

Recent days I found myself coping real hard Ai research wont be automated and at that exact moment I sounded so hypocritical.. like I didnt automate other jobs? I deserve everything thats coming to me but let me enjoy some time I have left

English

Simo Ryu@cloneofsimo·2d

Ai discourse used to be either 1. LLM is fucking nothingburger bubble, its fucking useless 2. Ai will kill all jobs and we will starve in hyper inequality And uh... now I see less of 1. Infact, i cant remember last time people took the former opinion seriously (apart from Ga*y ma*kus)

English

110

9.7K

Simo Ryu@cloneofsimo·2d

Fucking insane

Jared Duker Lichtman@jdlichtman

Update on Erdős Problem 1196: In joint work, we refined and adapted the proof method from GPT-5.4 Pro to give proofs of several additional problems. This includes another 60 year old conjecture by Erdős, Sárközy, and Szemerédi. A proof is valued not just by the problem it solves, but by what new avenues it opens up. This is perhaps one of the first examples of an AI-generated proof having downstream impacts, which we are still exploring. We are announcing the result today at the Future of Mathematics Symposium (see links below)

English

151

24K

Simo Ryu@cloneofsimo·3d

POV your cluster went down for maintenance

English

1.4K

Simo Ryu@cloneofsimo·6d

"Name 3 Erdos problems claude solved autonomously"

Zhu Liang@paradite_

I’m muting people on X who don’t understand how far ahead Anthropic is. It’s hurting my brain so much that I had to do this. If you think OpenAI is in anyway better than Anthropic, please just block me to save both us some trouble.

Español

155

10.4K

Simo Ryu@cloneofsimo·6d

Cool stuff

jo.schb@jo_schb

Diffusion models treat every part of an image equally. → Same number of steps. Same compute. But images aren’t uniform. 🤔 Some regions are easy, others are hard. So why force the model to treat them the same? 🧵

English

15K

Simo Ryu@cloneofsimo·27 Nis

Ironic, same artists wouldve freaked out when AI generates copyrighted Kpop demon hunter intellectual property characters.

Shany@Shanyshdw

English

1.7K

Simo Ryu@cloneofsimo·27 Nis

@DanAdvantage @vikhyatk so did walter white btw

English

142