Harshu Profile
Harshu

@IlyaSutskevar

Followers
255
Following
284
Media
100
Statuses
413

Me and my gpus against the world

Joined February 2024
Don't wanna be here? Send us removal request.
Explore tweets Explore followers Explore following
Explore trending content on Musk Viewer
Hook • 94326 Tweets
Cuban • 90611 Tweets
Hurricane Helene • 81338 Tweets
Eric Adams • 76427 Tweets
Clay Higgins • 64887 Tweets
Ten Hag • 48476 Tweets
Cardi • 43261 Tweets
Fever • 39817 Tweets
Caitlin Clark • 33360 Tweets
#AEWDynamite • 28471 Tweets
OpenAI • 27362 Tweets
東京ゲームショウ • 27247 Tweets
Yankees • 23373 Tweets
Sharon • 21716 Tweets
Nigel • 21551 Tweets
Offset • 21233 Tweets
#WNBA • 19338 Tweets
SO KISS ME LISA • 19080 Tweets
#AEWGrandSlam • 16759 Tweets
Costco • 16354 Tweets
辞任の理事 • 14776 Tweets
Connecticut • 13263 Tweets
Mineiro • 13001 Tweets
Lashley
メジャスピ
ノイジー
Stroman
Aliyah Boston
Deyverson
出直し選
Vatican
Danielson
Jobe
Lucas Moura
ゲーミングチェア
台湾海峡
Alyssa Thomas
マロッシュ
Bonner
Maya Wiley
V40 Lite
スクミュ
Mabrey
兵庫県の斎藤知事
午後会見
優三さん
出直し知事選
#Survivor47
Kelsey Mitchell
#ニトリと一緒に一狩りいこうぜ
Last Seen Profiles
Hank Alvey

@32halvey

Boyd Myers

@boydmyers

Penn Blockchain

@PennBlockchain

アフィストレア

@afystrea

Akhtar Abbas Turi

@akhtarturi110

Prince JunoxxXx™

@JunoXXXX

もの 🌨

@mo_no__2

shmttylol

@shmttylol

Cody Soh

@cody_soh

DIT-Ø 𝙼𝙰𝚇

@BADSHAH_CULTT

Elijah Mawhiney

@ElijahMawhiney

Prophetess Becca

@chenemi4me

XXL LX Fat Serpent God- Kukulkan

@bearrr55

Ryder Kusch

@KuschRyder

Догоняй! Е7Ы09 🆗

@mix_voronezh

Nida Pelin İlhan

@NidaPolly

Bear Earp

@BearEarp

ニシマールゲームズ | YouTube-eFootballフォーメーション専門-

@NishimarG

Edna

@Ednabreal

Mr. X-Toon

@mr_xtoon

@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
@maxjacob_me Yes, go on.
Tweet media one
10
20
3K
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
@karpathy
Tweet media one
0
1
15
@IlyaSutskevar
Harshu
@IlyaSutskevar
4 months
@natolambert If you're a student wanting an exciting life, good company, and impact on the real world : Work on whatever u like
0
2
15
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@untitled01ipynb
Tweet media one
0
1
12
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@mertdumenci Simple correction: I use the Amazon shopping app (Indian Software Engineer stressing behind it to send answer fast) for coding
0
0
11
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@Devgymsetvfx
Tweet media one
0
0
10
@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
ilya sutskevar reveal at today's event: 'i was actually in alternate dimension negotiating with alien ai's to share their tech. they said no btw :(
Tweet media one
2
0
9
@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
@YiTayML The research evaluator:
Tweet media one
0
0
8
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@AGIVMAN To keep it in simple words consider it like training your girlfriend for her next boyfriend
1
0
9
@IlyaSutskevar
Harshu
@IlyaSutskevar
4 months
@MelMitchell1 Same question to perplexity
Tweet media one
0
1
9
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@basedalexandoor ?
Tweet media one
0
0
9
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@ylecun @elonmusk Yann's community notes everybody.
1
0
8
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@untitled01ipynb Heyy
Tweet media one
0
0
8
@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
@OpenAI @StackOverflow
Tweet media one
0
1
8
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
Tweet media one
@untitled01ipynb
meowbooks (🕎|acc)
@untitled01ipynb
6 months
well #iykyk @satyanadella
Tweet media one
3
6
81
1
0
8
@IlyaSutskevar
Harshu
@IlyaSutskevar
4 months
why is the hallucination level in Gemini so high compared to other llms?
1
0
8
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
@OpenAI The world if they know what I saw that one night:
Tweet media one
0
0
7
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
@untitled01ipynb
Tweet media one
0
0
7
@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
Is collective intelligence, super intelligence?
4
1
7
@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
@perplexity_ai Getting shit done ✅
0
0
7
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
Miss those days <3
Tweet media one
4
0
7
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
@gdb I spy with my little eye, my name starts with I, who am I?
Tweet media one
0
0
7
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
Tweet media one
@AravSrinivas
Aravind Srinivas
@AravSrinivas
6 months
, , ,
Tweet media one
66
23
896
0
0
7
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
Perplexity never disappoints.
@AravSrinivas
Aravind Srinivas
@AravSrinivas
6 months
The world's best open-source chat LLM, DBRX, is now available for free, on . Perplexity Labs Playground basically has everything that you need for chat, for free, with better LLMs (Haiku, DBRX, Sonar) than 3.5-turbo, the model powering free chatGPT. Curious
Tweet media one
104
140
1K
0
0
7
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
Waiting!
Tweet media one
@mytechceoo
Jason
@mytechceoo
6 months
I hired an ex-CIA agent to find me Ilya, OpenAI’s missing co-founder
6
7
118
1
1
7
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@AravSrinivas What if the gradients become very small(closer to 0) won't we have issues like stuck gradients and very slow learning or maybe even during back propagation. The only reason they could be using tanh is because their neural network are not as deep as gpt models.
0
0
7
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
When AGI is achieved internally:
@paulg
Paul Graham
@paulg
6 months
A YC startup in the middle of fundraising found a way to cut their costs by 2/3. Now they don't need to raise at all. They can make it to profitability on the money they already have. Bet you can guess what will happen when they tell investors they're shutting down the raise.
139
100
3K
0
0
6
@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
@patrickc Whether divine or digital, the billboard landscape is definitely evolving.
0
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@shivon US<-->EU
0
0
6
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
AI -> AGI -> ASI -> Karpathy AI
@karpathy
Andrej Karpathy
@karpathy
6 months
Have you ever wanted to train LLMs in pure C without 245MB of PyTorch and 107MB of cPython? No? Well now you can! With llm.c: To start, implements GPT-2 training on CPU/fp32 in only ~1,000 lines of clean code. It compiles and runs instantly, and exactly
302
2K
13K
1
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
@lexfridman @sama Where am I?
0
0
6
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
@SmokeAwayyy
Tweet media one
0
0
6
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
Should I join Anthropic and implement what I saw?
2
0
6
@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
@greg16676935420 Not recommended:
Tweet media one
0
0
6
@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
@ClementDelangue Hopefully they release the paper soon 🤗
0
0
6
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@AravSrinivas
Tweet media one
0
0
6
@IlyaSutskevar
Harshu
@IlyaSutskevar
4 months
Am I cooked?
Tweet media one
@googlechrome
Chrome
@googlechrome
4 months
Be honest... how many tabs do you have open right now? 👀
6K
456
7K
0
0
6
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
@unusual_whales Won't let it burst out
1
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
4 months
I'm def not crying 😭
@ilyasut
Ilya Sutskever
@ilyasut
4 months
After almost a decade, I have made the decision to leave OpenAI.  The company’s trajectory has been nothing short of miraculous, and I’m confident that OpenAI will build AGI that is both safe and beneficial under the leadership of @sama , @gdb , @miramurati and now, under the
2K
3K
26K
2
0
6
@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
🔥
@gui_penedo
Guilherme Penedo
@gui_penedo
5 months
We have just released 🍷 FineWeb: 15 trillion tokens of high quality web data. We filtered and deduplicated all CommonCrawl between 2013 and 2024. Models trained on FineWeb outperform RefinedWeb, C4, DolmaV1.6, The Pile and SlimPajama!
Tweet media one
40
342
2K
0
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
4 months
@Michael_J_Black @nasim_rahaman Most of the papers feel like novels nowadays.
0
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
Now a days it feels like people/ companies are focusing more on data training to fine tune their models for a specific usecase rather than creating something new in core Machine Learning.
0
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@roydanroy Would you say it over here?
Tweet media one
0
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@elonmusk
Tweet media one
0
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
@ByeonChansoo @OpenAI @ilyasut @elonmusk Bing chilling
0
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
💯
@sarcastitva
Astitva Srivastava
@sarcastitva
6 months
Computer Vision conference's acceptance criteria these days: #CVPR2024 #eccc2024 #AI #ComputerVision
45
394
2K
0
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@greg16676935420
Tweet media one
1
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
@bindureddy Let's start from Cognition Labs itself!
0
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@SmokeAwayyy Ilya
0
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@bindureddy This should explain a lot!
Tweet media one
0
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
4 months
If any of you sound similar to Scarlett Johansson, your time has finally arrived, it's time to reach out to sam and make millions.
0
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@xlr8harder ❤️
0
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@rauchg A wrapper around Openai
0
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
@sama new work? how long have u been holding gpt-5 for?
0
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
@tunguz
Tweet media one
0
1
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
Aware!!! @perplexity_ai
Tweet media one
0
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
🦙^3 is live, go try it :)
@perplexity_ai
Perplexity
@perplexity_ai
5 months
We're excited to announce that Llama 3 is available on Perplexity Labs and our API. Kudos to the team @AIatMeta for all of the hard work they put into this release. We can't wait to see what you build with it. Try it free at
Tweet media one
22
78
545
0
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@AravSrinivas Felt like there's still a decent gap to be filled in depth of the Context awareness.
0
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
@AGIVMAN @sama
Tweet media one
0
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
"It's just a wrapper" "I can build it in a week" "Are vc's funding wrappers now?" What else?
1
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
(100B / 7T) * 100 = 1.42%, long way ahead.
@bindureddy
Bindu Reddy
@bindureddy
6 months
OpenAI snd MSFT want to build Stargate - a $100B GPU super cluster! Great! It’s time for Google to announce their $500B super cluster and Amazon to double down as well and start takling about their $300B cluster! They need to keep up with the Joneses 🤣🤣
42
35
311
0
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
@ElonMuskAOC Me
Tweet media one
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@LongHaired_Ilya Forget GPT-5, this one's for you
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
4 months
It's not about making their own AI, it's about competing with Open AI is what they fear and this is more absurd than being smart.
@elonmusk
Elon Musk
@elonmusk
4 months
It’s patently absurd that Apple isn’t smart enough to make their own AI, yet is somehow capable of ensuring that OpenAI will protect your security & privacy! Apple has no clue what’s actually going on once they hand your data over to OpenAI. They’re selling you down the river.
23K
46K
293K
0
0
6
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@tunguz Fixed it
Tweet media one
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
@ylecun Yann, is the paper gonna come out soon?
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
@IterIntellectus
Tweet media one
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
It's happening!
@ilyasut
Ilya Sutskever
@ilyasut
1 year
This too shall pass
37
48
568
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
The intern at Yahoo going crazy
Tweet card media
Don Lemon interviewed a mediocre white man

OPINION: In what might be the whitest interview ever, the former CNN anchor exposed the limits of X’s CEO perceived genius. The post Don Lemon interviewed a mediocre white man appeared first on...

www.yahoo.com
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@paulg I think Einstein would use something like these
Tweet media one
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
Should've created it sooner
@tiangolo
Sebastián Ramírez
@tiangolo
4 years
I saw a job post the other day. 👔 It required 4+ years of experience in FastAPI. 🤦 I couldn't apply as I only have 1.5+ years of experience since I created that thing. 😅 Maybe it's time to re-evaluate that "years of experience = skill level". ♻
1K
42K
168K
1
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@tunguz Can vouch!
Tweet media one
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
@jaltma Very great message from Sam's brother.
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@burny_tech Meanwhile my junior researcher mentioned she's expert in LLM's during interview
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
@elonmusk @Safety
Tweet media one
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
4 months
@adcock_brett How about we sell what our company actually focuses on primarily?
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
4 months
I like this
Tweet media one
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
@radbackwards I loved being in school great memories
1
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@gdb
Tweet media one
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
@t_blom Thank God you didn't get to review Brian Armstrong's application
1
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@tunguz !
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
4 months
@tunguz letting that sink in rn.
0
0
5
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
My other half 💕. @elonmusk
Tweet media one
1
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
@untitled01ipynb
Tweet media one
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
@karpathy @obsdmd Wanna join me opensourceing AGI?
Tweet media one
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
Wait am I tripping or this is an actual role in your company as well?
Tweet media one
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
1
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@giffmana @panopstor @twofifteenam normalising outputs to a range -1, 1 will help increasing the stability of the neural networks and maintain consistency also the gradients are strongest around zero in this case.
Tweet media one
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
Tweet media one
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
@AravSrinivas Unknown?
Tweet media one
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
@CollinCornwell @Humane Try capturing a Major market share in a Minor market than Minor share in a Major market!
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
@AravSrinivas
Tweet media one
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
6 months
@AravSrinivas Hope Blackwell won't fall into a deep well
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
7 months
Can we start a petition to rename 'San Francisco' to 'sam francisco' ?
1
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
@AndrewYNg @crewAIInc @joaomdmoura CrewAI is really cool to build AI agents, love their work.
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
@dwarkesh_sp @johnschulman2 whats the next architecture after transformers could be?
0
0
4
@IlyaSutskevar
Harshu
@IlyaSutskevar
5 months
@karpathy In what scenarios would the use of flash attention over naive attention yield more significant performance benefits, also how can we plan to incorporate flash attention into llm.c?
0
0
4