Smit B Patel

Mathematics❤️

[Smit B Patel @SmitBPatel2]

@Creative_Unity @TheVijayMallya Permanent Resident is like a green signal from the UK to let him stay there indefinitely as a foreign national (i.e. as an Indian). He wasn't a British passport holder.

[JASMEEN DUGAL @jasmeenGdugal]

Please repost and amplify to help find Sweta who is #missinginnepal
@IndiaInNepal #SOSNepal

[Radhika Sharma, DO @drradhikasharma]

I still have not been able to locate my parents.

Both US citizens: Ramesh Sharma (65 Male) and Neelam Sharma (64 Female).

My messages are open. Please message with any information.

[Smit B Patel @SmitBPatel2]

@kamath_sutra The parameter variables store all the essence of the data!

[Smit B Patel @SmitBPatel2]

@zanehengsperger Congratulations!

[Smit B Patel @SmitBPatel2]

@IBM Need a real cake/pastry!

[Smit B Patel @SmitBPatel2]

@yapper_so Commented!

[Yapper @yapper_so]

New video models coming soon 👀

Seedance 2.1
Seedance 2.0 Mini
Google Veo-4

We're giving away $50 of credits to 10 lucky people who repost, like, comment, and follow @yapper_so!

[Smit B Patel @SmitBPatel2]

@karpathy Congratulations @karpathy!

[Smit B Patel @SmitBPatel2]

@claude has real issues. You do not exhaust the limits and it still says you can't use it anymore until the limit resets.

[Smit B Patel @SmitBPatel2]

@claudeai is forever down!

[Bleap @BleapApp]

Giveaway time!

Win a Claude Max 20x subscription for one month.

To enter
> follow @BleapApp
> RT and like this post

Winner will be selected in 48 hours.

(fyi, you get 20% cashback on your claude, chatgpt and gemini subscriptions when using a Bleap card)

Download the app via the link in our bio today and activate your virtual card in minutes.

[Smit B Patel @SmitBPatel2]

@NandoDF Thanks for the share @NandoDF!

Also, wonderful to know the scope to be a part of @MicrosoftAI, Cheers!

[Smit B Patel @SmitBPatel2]

@djhelbert @awscloud @credly Congratulations!

[Smit B Patel @SmitBPatel2]

What I like most is the systems idea: keep expensive hardware busy by doing the right work in parallel, and use prediction plus caching to avoid idle time (Don’t let the no-work time go unused!)

#AI #LLM #Inference #Systems #GPU #PerformanceEngineering

[Smit B Patel @SmitBPatel2]

I read a new paper called “Speculative Speculative Decoding” (arXiv:2603.03251).

The problem is simple: large language mathematical systems generate text one token at a time. That sequential loop becomes the bottleneck during inference, even on powerful GPUs.

[Smit B Patel @SmitBPatel2]

Reported results (on Llama-3.1-70B with H100 GPUs) show up to about 2x faster than optimized speculative decoding and up to about 5x faster than standard autoregressive decoding.