Uddeshya Upadhyay (@uddupa) 's Twitter Profile
Uddeshya Upadhyay

@uddupa

DeepDabbler
Musing - ML/data/systems/(bio)hacking

Prev:
work @_nference @maddox_ai @MSFTResearch @AmazonScience
CS PhD @ML4Science x @MPI_IS
CS UG @IITBombay

ID: 833245388312178688

linkhttps://udion.xyz/ calendar_today19-02-2017 09:22:36

735 Tweet

329 Takipçi

1,1K Takip Edilen

Thomas Wolf (@thom_wolf) 's Twitter Profile Photo

we've seen nothing yet! hosted a 9-13 yo vibe-coding event w. Robert Keus 👨🏼‍💻 this w-e (h/t Anton Osika – eu/acc Lovable Build) takeaway? AI is unleashing a generation of wildly creative builders beyond anything I'd have imagined and they grow up *knowing* they can build anything!

Cerebras (@cerebrassystems) 's Twitter Profile Photo

Cerebras just beat NVIDIA Blackwell Last week: Blackwell hit 1,000 t/s on Llama 4. Today: Cerebras hit 2,500 t/s on the same model, same benchmarks by Artificial Analysis Blackwell smoked Groq, AMD, Google – everyone. Only Cerebras stands – and we smoked Blackwell.

Cerebras just beat NVIDIA Blackwell
Last week: Blackwell hit 1,000 t/s on Llama 4.
Today: Cerebras hit 2,500 t/s on the same model, same benchmarks by <a href="/ArtificialAnlys/">Artificial Analysis</a>
Blackwell smoked Groq, AMD, Google – everyone.
Only Cerebras stands – and we smoked Blackwell.
chester (@chesterzelaya) 's Twitter Profile Photo

Monday we ship: 0.0.01 release out now UI changes and added scripting module to create wireless mods for drones "Follow" is the first mod on the bounty list... $100 to the first dev that can build it! Instructions on our discord #bounties

Monday we ship: 0.0.01 release out now 

UI changes and added scripting module to create wireless mods for drones

"Follow" is the first mod on the bounty list... $100 to the first dev that can build it! Instructions on our discord #bounties
Georgia Gkioxari (@georgiagkioxari) 's Twitter Profile Photo

💫It's fascinating that a single feed-forward pass through an LLM can replace a complex rendering pipeline, like Blender! Just feed it 3D shapes, xyz positions, and poses as tokens, and it spits out the image token-by-token. The dual, aka scene reconstruction, is also

Chen-Hsuan Lin (@chenhsuanlin) 's Twitter Profile Photo

Cosmos-Predict2 is our latest open video foundation model for Physical AI! research.nvidia.com/labs/dir/cosmo… If you’re at #cvpr2025, I would also love to chat with you about world models!

anu (@anuatluru) 's Twitter Profile Photo

there’s a palpable urgency in the air that if you’re ever going to chase your life’s work the time to do it is now because five years from now the way the world works may look nothing like today

Uddeshya Upadhyay (@uddupa) 's Twitter Profile Photo

one of the greatest joys of life is going back to basics and building up.💡I'm capturing my musings as I progress... first up: "The Magic Behind the Click - Image Capture Fundamentals." udion.xyz/musings/from-p… It's Part 1 of a bigger journey leading to real-time, low-resource,

Rohan Paul (@rohanpaul_ai) 's Twitter Profile Photo

It’s a hefty 206-page research paper, and the findings are concerning. "LLM users consistently underperformed at neural, linguistic, and behavioral levels" This study finds LLM dependence weakens the writer’s own neural and linguistic fingerprints. 🤔🤔 Relying only on EEG,

It’s a hefty 206-page research paper, and the findings are concerning.

"LLM users consistently underperformed at neural, linguistic, and behavioral levels"

This study finds LLM dependence weakens the writer’s own neural and linguistic fingerprints. 🤔🤔

Relying only on EEG,
Siddhartha Gunti (@sidgunti) 's Twitter Profile Photo

Found a paper on HackerNews - MIT studied effects of using GPT on our brains. Results are not surprising. Using GPT more leads to accumulating cognitive debt. Tldr-