Rishin Sharma (@rishisthinking) 's Twitter Profile
Rishin Sharma

@rishisthinking

investing @coinfund_io | Built @WhartonCypher, other blockchain stuff @Penn | Pondering decentalized AI | NFA

ID: 1305860022749007872

calendar_today15-09-2020 13:24:36

92 Tweet

596 Followers

1,1K Following

bidhan roy 🥯 (@bidhanxyz) 's Twitter Profile Photo

A product leveraging cryptography is trending in the top 10 of Product Hunt 😸 WORLDWIDE! Don't think this has ever happened before. Monetizable open source AI is a generational movement. We @BagelOpenAI just started the fire.

A product leveraging cryptography is trending in the top 10 of <a href="/ProductHunt/">Product Hunt 😸</a> WORLDWIDE!

Don't think this has ever happened before.

Monetizable open source AI is a generational movement. We @BagelOpenAI just started the fire.
bidhan roy 🥯 (@bidhanxyz) 's Twitter Profile Photo

In 2024, zero knowledge verification for machine learning seemed impossible. The latency overhead was too high. We did multiple research reports on it. But, it is 2025. The year of AGI. And, we have made it possible. Today, we open source our original research, ZKLoRA. A zero

In 2024, zero knowledge verification for machine learning seemed impossible. The latency overhead was too high. We did multiple research reports on it.

But, it is 2025. The year of AGI. And, we have made it possible.

Today, we open source our original research, ZKLoRA. A zero
Prime Intellect (@primeintellect) 's Twitter Profile Photo

Join Us: This year, we will train a fully open reproduction of o3, enabling everyone globally to contribute compute resources. If you want to be part of this, DM us for collabs Johannes Hagemann Justus Mattern or apply here: jobs.ashbyhq.com/PrimeIntellect

Prime Intellect (@primeintellect) 's Twitter Profile Photo

Today, we release TOPLOC: A Locality Sensitive Hashing Scheme for Verifiable Inference - Detects modifications to models, prompts, or precision - Robust across GPU types, tensor parallel configurations and attention kernels - Up to 100× faster validation than generation -

Today, we release TOPLOC: A Locality Sensitive Hashing Scheme for Verifiable Inference

- Detects modifications to models, prompts, or precision
- Robust across GPU types, tensor parallel configurations and attention kernels
- Up to 100× faster validation than generation
-
Rishin Sharma (@rishisthinking) 's Twitter Profile Photo

another day another drop from prime intellect. with synthetic traces becoming the new paradigm for training reasoning models, pretraining in a distributed setting becomes that much more accessible

Rishin Sharma (@rishisthinking) 's Twitter Profile Photo

pretty big deal, data has been a bottleneck for LMs to be more useful in applied science contexts. This is like a 100x on the previous largest single-cell dataset. Predictive models are going to accelerate drug discovery and testing at an incredible rate—> years of lab work

Rishin Sharma (@rishisthinking) 's Twitter Profile Photo

Work on the hardest problems with the biggest potential impacts on the proliferation of technology. This is the ethos Pluralis Research , striking a new path for foundation model development. Congrats Alexander Long and team on the start of a very exciting future