The UI and marketplace are live now at https://t.co/xKlClfuz1Y. The CLI (@openmedica/cli), REST API, and MCP server are coming soon.
But this only works if the community builds it together. I need submissions -- from everyone:
- You are a developer? Submit a skill, improve an existing one, or build an integration. The repo is open.
- You are a physician or researcher? Suggest a tool you wish existed. Flag a gap in the catalog. Review a skill for clinical accuracy.
- You are neither? You can still contribute -- documentation, testing, translations, or just telling us what is missing.
🔥FDA has approved Zaynich (cefepime and zidebactam) for adults with complicated urinary tract infections (cUTIs) caused by designated susceptible microorganisms.🔥 #idxposts@BradSpellberg@IDstewardship https://t.co/WbXvttg5zB
This Memorial Day, as America approaches 250 years, we honor those who gave their lives so this nation could endure. May we remember them with gratitude and renewed responsibility to protect the freedoms and future they sacrificed for.
The HRM-Text paper is now available 🎉
HRM-Text explores a different approach to language model pretraining: hierarchical recurrent computation, task-completion training, and latent-space reasoning.
At just 1B parameters, HRM-Text achieves competitive performance with dramatically lower training cost and data requirements.
1B parameters
40B unique tokens
~1 day of pretraining
~$1000 training cost
Wired my Obsidian vault into Claude via MCP.
3 years of notes -> queryable knowledge graph. Ask: "Find contradictions across my infra decisions since 2023." Real answer, seconds.
Best learning workflow I've used.
https://t.co/HuQxfcW0zr
16 backends you can flip with one command:
stt-switch mlx-whisper-large-v3-turbo # 89ms local GPU
stt-switch moonshine-onnx-tiny # 17ms CPU
stt-switch mlx-chain # local + remote fallback
MIT, Apple Silicon. PRs welcome for Linux/CUDA.
shipped: stt-switch — Claude Code voice mode, but local
your speech never leaves the Mac.
89 ms warm on M5 Max GPU (whisper-large-v3-turbo).
faster than the round-trip to Deepgram.
one env var to enable, one CLI to swap models.
https://t.co/jjiRYhWNn4
turns out Claude Code already supports a VOICE_STREAM_BASE_URL env var. point it at a local proxy that speaks the same Anthropic voice_stream protocol (decoded from the leaked CC source) and audio stops going to https://t.co/yYGnV5N6U9 → Deepgram. zero outbound on the happy path.
Wired my Obsidian vault into Claude via MCP.
3 years of notes -> queryable knowledge graph. Ask: "Find contradictions across my infra decisions since 2023." Real answer, seconds.
Best learning workflow I've used.
https://t.co/HuQxfcW0zr
Browse the catalog: https://t.co/jJjVqaucwk a project of https://t.co/KGcma14lCU
Contribute: https://t.co/3n38yOKocM
For research purposes only. Not intended for clinical decision-making.
#OpenSource#MedicalAI#DigitalHealth
747+ open-source medical AI tools -- unified in one platform. Physician-reviewed. Evidence-classified. Free forever.
I am a board-certified Internist, trained in infectious disease who writes code, trains models, and builds clinical tools. I started IntelMedica with my father who has been an inspiration for me to be a #healthcare #technologist- long time before the ChatGPT moment.
I am honored to say that today I started open-sourcing my projects. I created a signle place OpenMedica because I was tired of duct-taping scripts together every time I needed a drug interaction check or a literature search. The tools should already exist in one place. Now they do
Research use only.
The UI and marketplace are live now at https://t.co/xKlClfuz1Y. The CLI (@openmedica/cli), REST API, and MCP server are coming soon.
But this only works if the community builds it together. I need submissions -- from everyone:
- You are a developer? Submit a skill, improve an existing one, or build an integration. The repo is open.
- You are a physician or researcher? Suggest a tool you wish existed. Flag a gap in the catalog. Review a skill for clinical accuracy.
- You are neither? You can still contribute -- documentation, testing, translations, or just telling us what is missing.