Today, we are emerging from stealth and launching PrismML, an AI lab with Caltech origins that is centered on building the most concentrated form of intelligence.
At PrismML, we believe that the next major leaps in AI will be driven by order-of-magnitude improvements in intelligence density, not just sheer parameter count.
Our first proof point is the 1-bit Bonsai 8B, a 1-bit weight model that fits into 1.15 GBs of memory and delivers over 10x the intelligence density of its full-precision counterparts. It is 14x smaller, 8x faster, and 5x more energy efficient on edge hardware while remaining competitive with other models in its parameter-class.
We are open-sourcing the model under Apache 2.0 license, along with Bonsai 4B and 1.7B models.
When advanced models become small, fast, and efficient enough to run locally, the design space for AI changes immediately. We believe in a future of on-device agents, real-time robotics, offline intelligence and entirely new products that were previously impossible.
We are excited to share our vision with you and keep working in the future to push the frontier of intelligence to the edge.
Alright letโs put this ARR debate to rest once and for all..
If ANY vibe coding startup with scale (>$5M claimed ARR) publicly discloses how they calculate it (esp sub vs. inference) and how much of it they subsidize,
Iโll donate $2,500 to a charity of their choice today ๐ซก
ever been here?
open overleaf โ write a paragraph โ "hmm...this needs a citation" โ open 15 different tabs โ skim 8 abstracts โ find the 1 actually relevant paper โ format bibtex โ paste it back on overleaf
if so, i built a plugin just for you. meet openleaf:
โ reads your paper paragraph by paragraph
โ searches major academic databases
โ filters out irrelevant papers using ai
โ one click to add BibTeX to your .bib
you'll also find the ๐ค friendly and ๐ฅ fire reviewers there. i don't think i need to tell you what they do :)
free. open source. no account. no data collection.
works with ollama, openrouter, openai api and more.
https://t.co/XvX03iem38
dear algorithm, please show this to my fellow researchers in need ๐
#overleaf #latex #opensource #academictwitter
Kalpa Labs (@KalpaLabsAI) is training one generalist model for all speech tasks. Steer it like an LLM, give long system prompts & make use of native contextual awareness.
Congrats on the launch, @pshishodiaa and @jhagauti!
https://t.co/V0AbssT7CN
Andrej Karpathy calls AI Agents slop
"Overall, the models they are not there. And I feel like the industry [...] it's making too big of a jump and it's trying to pretend that this is amazing.
And it's notโit's slop!
And I think they are not coming to terms with it. And maybe they are trying to fundraise or something like that, I'm not sure what's going on."
Fundraise Announcement!
We are proud to announce that we at @smallest_AI have raised 8M USD in an oversubscribed seed round led by @Sierra_Ventures with participation from @3one4Capital, @better_capital, @UpsparksCapital, @SchemaVentures, @DeVC_Global, @tinyvc, @peercheque, @shyamalanadkat, @missionstcap, @bwertz, @RaveenSastry, and many more angelsย to build the future of Enterprise Voice AI!
We are transforming voice AI from the ground up - pushing agents to pass the Turing test in the coming years.
We are also excited to build a research-first organization - diving into the depths of problems from first principles and doing things no one in the world is even thinking about.
It was our dream to build a scientific team that pushes the boundaries of the frontier. We are privileged to have the support of some of the best early backers a founder could ever ask for.
Thank you, everyone for supporting us - this is just the beginning!
Nice! Interestingly, we recently put together a similar framework at https://t.co/EHH3CswEfD, with lots of options regaridng both pretraining and finetuning (choice of tokenizers, finetuning methods, etc), just didn't announce it properly yet ๐