⚡🛡️ Evan Pappas
@Hevalon
🛡️ Ex Technologia Libertas - Έλευθερία διὰ τῆς τέχνης - (Dec/Acc)
🧵 Why Templar isn't rushing to bigger models (and why this strategy will dominate) 0/ "Wen bigger model?" We get this question constantly. Here's why we're perfecting 8B models first—and why this approach will crush the competition when we do scale up. Thread below 👇
there's no greater battle right now, than opensource and decentralized AI, and for that we need global available and decentralized compute.
Compute scarcity is real. That’s why we’re building #basilica SN39: a permissionless, crypto-incentivized compute cluster for the next generation of AI. No more bottlenecks, just pure scale. #DeAI #DecentralizedAI #Bittensor $TAO
Bittensor Guru S2E10 - Subnet 3 @tplr_ai & 39 Basilica w/ Sam @DistStateAndMe Bittensor's home run swing at distributed training and compute explained by it's charismatic leader himself. $TAO
Only decentralized aggregators will hope to compete with GPU numbers like this.
we will cross well over 1 million GPUs brought online by the end of this year! very proud of the team but now they better get to work figuring out how to 100x that lol
> What once required 114GB of communication now needs just 2.8GB—a 40x improvement in efficiency while achieving better model performance.
x.com/i/article/1945…
this is interesting, vectors over s3 able to perform semantic searches natively. I'm wondering how could we offer this in a decentralized way. We'll have to introduce a distributed vector index, that can map the position of the embed docs in a (sharded?) decentralized store.

Meta’s pivot to closed-source AI is worrying. It risks stifling innovation, favoring big players, and widening the U.S.-China AI gap as open models (e.g., Kimi, DeepSeek) thrive elsewhere. But there’s hope! Enter Templar @tplr_ai , Bittensor Subnet 3—decentralized AI done right.…
RIP to the unicorn AI startups that have zero products, zero foundation models, and were just going to depend on big labs releasing open-source models for free to model merge. I know one or two.
are chatbots becoming more intelligent, or people start interacting with each other with prompts? > You're absolutely right! ...
the intern is posting again..
Reminder : Templar is a small team (5) + miners producing world class research tplr.ai/research
TPLR II crushing with 600 tok/s over internet. tplr.ai/dashboard
Absorb Global Compute ... Basilica Rising
we're live, mainnet.