Taylor Hou

@anushkmittal @Scobleizer Happy to collab if the agents can be powered by smaller local models with teale.com where everyone's devices turn into nodes supplying inference for the rest of the world

We have raised $8M to bring AI into your group chats. shapes.inc is the next-gen whatsapp, re-imagined to be AI first. Our life runs in group chats - it’s only natural for AI agents to join us in these conversations. Multiplayer AI is the next frontier for agentic systems. Excited to come out of stealth after 3.5 years of building and over 1B+ conversations.

just ran a test with rapid-mlx on 2 of my mac machines (64gb & 96gb ram) and both resulted in meaningful gains. i'm implementing rapid-mlx as an additional backend for teale.com today!

Hey @united 11A on this flight should be clarified as a "wall" seat instead of window...

Hey @lyft one of your driver's cancelled my scheduled ride to airport this morning (likely due to the pre-booked price) and I had to rebook for almost double the price and I might miss my flight now...

@QingQ77 I'm implementing rapid-mlx as a backend on teale.com for distributed inference powered by macs around the world. If it really can do the TPS and ttft it claims, it will be gamechanger!

在 Apple Silicon Mac 上本地运行 LLM 推理服务，提供比 Ollama 和 llama.cpp 更快的 OpenAI 兼容 API，同时原生支持工具调用和提示缓存。 github.com/raullenchai/Ra… Rapid-MLX 用 Apple 自家的 MLX 框架做推理，搭了个 FastAPI 服务跑 OpenAI 兼容 API。在 Apple Silicon 上比 Ollama 快 2-4 倍，靠 KV 缓存裁剪和 DeltaNet 状态快照把续轮 TTFT 压到 0.08 秒左右。工具调用这块做了 17 种解析器，Qwen、DeepSeek、Gemma、GLM 这些模型自动识别格式，量化把输出搞坏的情况也能自动修回来。 另外还有推理链分离、云端路由、视觉/音频多模态、V 缓存压缩等功能。Cursor、Claude Code、Aider、LangChain 都能直接对接。

@SamJWasserman When you aren't using the machines, contribute your ram to power distributed inference on teale.com!

I did a thing and there is no turning back now. Time to cook and push the limits of my creativity. If anyone else has been building with local LLMS, let me know, would love to hear your takeaways, tips so far.

@benln Distributed inference powered by everyone's sleeping MacBooks, laptops, and even smartphones. - teale.com

Kicked off a group chat for Cursor team members who occasionally angel invest. There are now 20 of us. Reach out if you're working on something new, love to share with the group!

if you haven't heard, @claudeai cowork now supports 3rd party providers, models, and inference. if you use it in conjunction with teale.com, you get claude's best-in-class ai harness + local models and free distributed inference.

big feature by @conductor_build - i've wanted to steer (influence without interupting) my agents especially on long runs and it's here!

@McDonalds your apple slices in kid's meals have a chemical taste. Please fix.

@ProductHunt Distributed inference powered by sleeping laptops - teale.com

pitch your product in 5 words or less 👇

@anupamrjp Distributed inference - teale.com

Your SaaS in 2 words. No more. No less. Let's go Go 👇

@ItsElanaGold Houmanoids.com - bootstrapped to first real autonomous deployment of quadruped robot for security and patrol in America.

Startups to pay attention to: - Heron Power - Bedrock Robotics - Skild AI - General Matter - Amca - Kela - Hadrian Who am I missing?

distributed inference is fcking hard. huge props to the frontier labs for making the UX seamless and "it just works" but they have the benefit of VC and ridiculous compute. when you're ram constrained, it's a different ballgame.

Yup. The world is NEW compute constrained. There's already tens of BILLIONS of gigabytes already out in the world on personal and work devices. Two things are happening (and likely will continue) - 1. open models are getting better and more efficient. 2. There's amazing work by folks like mesh-llm to split inference across many devices instead of requiring massive datacenters

@taylorhou using idle computers for inference is a fun angle if the economics work this could get weird in a good way

Distribution is everything. Share your product in the comments below. I will choose a few and repost it!

@daniel_mac8 Contribute your excess compute to the teale.com distributed inference network!

Compute is currency. That's why you should secure your own compute. Delivered today was an ASUS Ascent GX10: > NVIDIA Blackwell GB10 > 128GB VRAM > 1 petaFLOP compute Should I post about how I'll use it?

@anupamrjp distributed ai inference. teale.com

Drop your SaaS 👇 No pitch. No hype. No 10-slide explanation.

@thekchasiotis sup

the longer I'm on 𝕏, the more I realize: founders with small accounts are the most interesting ones. - too busy building to posture - 0 ego, they just wanna win - keep posting with 0 likes - the world isn’t rooting for them yet but I will tell me what you are building