nicoarq (@nicoarqueros) 's Twitter Profile
nicoarq

@nicoarqueros

cofounder and core dev @shinkaiprotocol πŸ‡ΊπŸ‡Έ πŸ‡¨πŸ‡± MCP maximalist

ID: 68308933

linkhttps://shinkai.com calendar_today24-08-2009 02:55:55

5,5K Tweet

7,7K Followers

2,2K Following

Shinkai Protocol (@shinkaiprotocol) 's Twitter Profile Photo

Folders that organize themselvesβ€”sounds impossible, until you meet Shinkai. In just moments, our AI learns how you like things filed and puts every image and document exactly where you’d place it.

Paul Couvert (@itspaulai) 's Twitter Profile Photo

Wait so Alibaba Qwen has just released ANOTHER model?? Qwen3-Coder is simply one of the best coding model we've ever seen. β†’ Still 100% open source β†’ Up to 1M context window πŸ”₯ β†’ 35B active parameters β†’ Same performance as Sonnet 4 They're releasing a CLI tool as well ↓

Wait so Alibaba Qwen has just released ANOTHER model??

Qwen3-Coder is simply one of the best coding model we've ever seen.

β†’ Still 100% open source
β†’ Up to 1M context window πŸ”₯
β†’ 35B active parameters
β†’ Same performance as Sonnet 4

They're releasing a CLI tool as well ↓
Shinkai Protocol (@shinkaiprotocol) 's Twitter Profile Photo

What if building your own AI tools felt more like assembling Lego than coding from scratch? Shinkai lets you create powerful, flexible AI agentsβ€”no-code or full-script, your call. From personal assistants to problem solvers, it adapts to how you think. Long Video Link πŸ‘‡πŸΌ

The Defiant (@defiantnews) 's Twitter Profile Photo

The DeFi x AI space continues to expand at a rapid pace, and Shinkai: Local AI Agents empowers its users with its decentralized peer-to-peer agentic network Through Shinkai, anyone can build and deploy their own custom AI Agent onchain Read more: thedefiant.io/education/defi… -sponsored-

Halborn (@halbornsecurity) 's Twitter Profile Photo

πŸ“’ We're excited to share that we have successfully completed our smart contract security assessment for Shinkai: Local AI Agents πŸ”’ Shinkai is a two-click install app that doesn't just give you an AI to chat withβ€”it lets you build your own team of intelligent crypto agents, running 24/7

J.J. Siler (@jj_siler) 's Twitter Profile Photo

πŸ‡¦πŸ‡· It was a pleasure to welcome these folks to Buenos Aires to discuss city simulation and private stablecoin implementations on Midnight. The future is bright, and we look forward to continue hosting development workshops in our office in the future.

πŸ‡¦πŸ‡· It was a pleasure to welcome these folks to Buenos Aires to discuss city simulation and private stablecoin implementations on Midnight. The future is bright, and we look forward to continue hosting development workshops in our office in the future.
Shinkai Protocol (@shinkaiprotocol) 's Twitter Profile Photo

Upgrade your stack, not your prompts. Shinkai 1.1.6 is out 🍻 GPT‑5 + Grok support, newest Ollama with GPT‑OSS, visible reasoning where supported, and tool‑calling fixes for Groq/DeepSeek. Universal reasoning output lands next. Get it here πŸ‘‡πŸΌ

Shinkai Protocol (@shinkaiprotocol) 's Twitter Profile Photo

Less friction. More flow. Shinkai 1.1.6 is out πŸ‘€ Refined UI plus reliability improvements (edge cases) for Grok/Claude so your work stays unbroken. Also an updated free (but limited) inference with gpt-5 for testing and low payload

nicoarq (@nicoarqueros) 's Twitter Profile Photo

interesting new startup πŸ‘€ i'll check it out to see if their claims are true: "The highest accuracy web search for your AI" The only thing for sure, Elon Musk is not an investor πŸ˜†

nicoarq (@nicoarqueros) 's Twitter Profile Photo

started using picture in picture waaay more now that i need to constantly wait for AI models to give an answer and I don't want to get distracted in another window

nicoarq (@nicoarqueros) 's Twitter Profile Photo

the new praying that your project build works is praying that all of your project source code fits in the context window of the AI lol

Vik Paruchuri (@vikparuchuri) 's Twitter Profile Photo

You can teach old models new tricks! We added math capability to the 650M param surya OCR earlier this year. Its now better than frontier LLMs and 7B+ models at the arxiv math part of the olmocr benchmark.

You can teach old models new tricks!  

We added math capability to the 650M param surya OCR earlier this year.  Its now better than frontier LLMs and 7B+ models at the arxiv math part of the olmocr benchmark.