
Leachim Ardnij retweeted
Lots of tiny TTS releases lately 🗣️
> Paradee v1.3 by @sahilmdkr
> ZONOS3-Mini by @ZyphraAI
> PiVo by @EmperoAI
> phonon by @GradiumAI
> Underdog-Woof by @UnderdogAI
Compare and listen to them side by side on the @huggingface Open TTS Leaderboard 🎧
huggingface.co/spaces/hf-aud…
We believe voice is the most natural way to interact with intelligence, so we created Dessa, the most authentic full-duplex speech model ever built. Talking with it feels like talking with a person…full of emotion, nuance, pauses, and a natural cadence.
block.xyz/dessa
Leachim Ardnij retweeted
We are releasing world's smallest zeroshot TTS today!
This can run on ESP32-S3 equivalent processors. It means that it works on M5 stack in realtime with voice cloning capability.
Leachim Ardnij retweeted
Tesla VP of Hardware Engineering Lars Moravy explains to Hagerty how the Roadster borrows SpaceX cold gas thrusters to make 0 to 60 possible in 0.55 seconds.
Truly insane:
▫️ 200L carbon-over wrapped pressure vessel
behind seat holds nitrogen at 10,000 PSI
▫️cold gas thrusters push out 80 kN of thrust (equivalent to a single fighter jet engine)
▫️Tesla went with nitrogen because “it just seemed wrong to put fuel into a car” (air is ~80% nitrogen, fair enough)
▫️thrusters same system satellites use to move in space (eg. the “psh psh psh" from Apollo 13 film)
▫️currently takes 6 to 7 hours to refill all nitrogen but Tesla working for it to be 30 minutes
▫️uses leg actuators from Optimus robot to manage downforce on wing
***
Full 25-min episode with Henry Catchpole here: youtu.be/22rvpUecLts?si=jtwK…
Leachim Ardnij retweeted
I’m calling JEV 4 times a second and telling it to find 5 menus of items. Insanely fast and cheap option to identify objects in frame.
Leachim Ardnij retweeted
Introducing 🤗 Transformers.js v4: state-of-the-art machine learning for the web!
🚀 New WebGPU backend (browser, Node.js, Bun, Deno)
⚡️ Huge performance improvements
🤯 Support for over 200 architectures
🛠️ Complete codebase refactor
Learn more about our biggest release yet! 👇
Leachim Ardnij retweeted
Replying to @_guillecasaus
try jopdf.com/ a mi parecer mas completo y free también..
Leachim Ardnij retweeted
Jev is the most exciting AI release in months. So I built the open, in-browser version 🎉
open-jev: System One style decisions running 100% on your device. Text in, typed answers out, one forward pass. Nothing is generated, so it can't hallucinate.
Leachim Ardnij retweeted
That's crazy when you think about it
Qwen 3.8 27B is smarter than Opus 4.6 which was the absolute best model you could use 6-7 months ago.
And now we have this Bonsai 2 retaining 98%+ of its intelligence...
While weighing ONLY 5.9 GB 🤯
Intelligence density is absolutely crazy.
Everyone has frontier intelligence at home now.
Today, we’re announcing Ternary Bonsai 2 27B.
Based on Qwen3.8 27B, Bonsai 2 27B is 9x smaller than its full-precision counterpart while retaining 98.2% of its aggregate benchmark performance.
Two months after the first Bonsai 27B release, the biggest change is quality. The footprint remains 5.9 GB, but the gap to full precision has narrowed materially, with particularly strong gains in agentic coding, multimodal reasoning, and long-horizon tool use.
Ternary Bonsai 2 27B is available today under Apache 2.0.
Jev vs GLM 5.3 at chess!
Results:
◾ GLM 5.3 won by checkmate in 29 moves
◾ Jev: ~0.3s and <$0.0001 per move
◾ GLM 5.3: ~5.8s and ~$0.008 per move
◾ The whole game cost 24 cents
My main takeaway is that it's often useful to use each one to their strengths:
◾ Fast, well-defined classification → specialized models like Jev
◾ Classifications that need reasoning or lookahead → LLMs like GLM 5.3
◾ Real classification pipelines → hybrid. Jev handles the easy calls, an open model handles the hard ones.
The future is multi-model!
Leachim Ardnij retweeted
New experiment: json-render + jev
The future Generative UI is instant
Your components, your actions, your design system
Rendered in milliseconds
✅ The best XRP Ledger Lending protocol explainer so far!
Well explained, how it works and what this brings in utility.
Lending, borrowing and collateral management is going to be a major pillar of the XRP Ledger.
Made a short video on the Single Asset Vault (XLS-65) on the XRP Ledger.
Before you can borrow on-chain, the money has to come from somewhere. The vault is that shared pool. You deposit, you get shares back, and those shares are your slice of it. If the pool earns, your shares can be worth more when you redeem.
First in a series on on-chain lending.
Leachim Ardnij retweeted
Crown shyness is one of many adaptations that allow trees to coexist in communities.
Certain species leave gaps when growing next to each other; a behavior scientists believe may be a way to ensure they both have access to sunlight and resources.


























