Valentin

38.7K posts

Valentin banner
Valentin

Valentin

@TTimeIsWinning

Click here 👇🏻

Katılım Mayıs 2011
197 Takip Edilen3K Takipçiler
Valentin retweetledi
Valentin
Valentin@TTimeIsWinning·
Comments
Valentin tweet media
Français
3
37
190
13.1K
Valentin
Valentin@TTimeIsWinning·
Check the comments and thank me later
Valentin tweet media
English
70
286
11.7K
414K
Valentin
Valentin@TTimeIsWinning·
after typing 'hi' check ur DMs😳
Valentin tweet media
English
568
458
18.3K
676.9K
Valentin retweetledi
Richard Dawkins
Richard Dawkins@RichardDawkins·
This social contagion memidemic has run its course and will soon fade into history. Except, unlike other playground crazes (yo-yos, jacks, hula-hoops, marbles) this one will never fade for those unfortunate children irreversibly damaged by chemical or even surgical mutilation skeptic.com/article/transg…
English
2.2K
6.8K
31.5K
2.7M
Valentin retweetledi
Mindful Maven
Mindful Maven@mindfulmaven_·
Mindful Maven tweet media
ZXX
84
790
4.6K
177.7K
Valentin retweetledi
Elon Musk
Elon Musk@elonmusk·
♥️🚀🇺🇸
Elon Musk tweet media
QME
13.8K
24.1K
232.8K
25.5M
Valentin retweetledi
Elon Musk
Elon Musk@elonmusk·
And there is a clear path to a doubling of performance on all metrics for AI6 within 10 to 12 months of AI5 shipping
X Freeze@XFreeze

Elon Musk: At Tesla, we basically had two different chip programs: one Dojo and one. Dojo on the training side, and then what we call AI4, it's just our inference chip The AI4 is what's currently shipping in all vehicles, and we're finalizing the design of AI5, which will be an immense jump from AI4. By some metrics, the improvement in AI5 will be 40 times better than AI4. So not 40%, 40 times This is because we work so closely at a very fine-grained level on the AI software and the AI hardware. So we know exactly where the limiting factors are. And so effectively the AI hardware and software teams are co-designing the chip Compared to the worst limitation on AI4, which is running the SoftMax operation, we currently have to run SoftMax in around 40 steps in emulation mode, whereas that'll just be done in a few steps natively in AI5 AI5 will also be able to easily handle mixed precision models, so you don't have it, it'll dynamically handle mixed precision. There's a bunch of sort of technical stuff that AI5 will do a lot better In terms of nominal raw compute, it's eight times more compute, about nine times more memory, and roughly five times more memory bandwidth But because we're addressing some core limitations in AI4, you multiply that 8x compute improvement by another 5x improvement because of optimization at a very fine-grained silicon level of things that are currently suboptimal in AI4, that's where you get the 40x improvement

English
16.7K
10.9K
79.5K
20.8M
Valentin retweetledi
Valentin retweetledi
Tesla
Tesla@Tesla·
Model Y Performance from sketch to reality
Tesla tweet mediaTesla tweet media
English
729
1.5K
11.8K
29M
Valentin retweetledi
Elon Musk
Elon Musk@elonmusk·
Grok Imagine
English
39.9K
11.2K
87.8K
8.3M
Valentin retweetledi
Elon Musk
Elon Musk@elonmusk·
Made with Grok Imagine
Déborah@dvorahfr

Made with @grok Imagine using the last frame of each sequence and using it as the starting frame of the next sequence.

English
3.7K
3.5K
27.8K
6.1M