The hard part of APCM is not just knowing the program exists.
It is making the work visible enough for a real team to run it.
Patient progress. Level assignment. Documentation. Follow-up. Next step.
Here’s a quick FairPath walkthrough:
https://t.co/gJNIHbBLuh
#APCM #clinicalservices
@DanielSmidstrup No and I really don't care if it does or not.
Just want codex or co work usage to last more than 5 minutes before a five hour usage reset
@araseb_ Everyone just assumes that because the model is "God Tier" that it must be telling the truth now. See the model number go up = assume that it is always honest. That's what is happening.
1/8
I asked Buffaly to connect to Unity.
I did not give it an MCP server. I did not give it a prebuilt Unity plugin. I did not give it a tool wrapper.
I gave it the goal and let it figure out the path. Using glm 5.2 and #ollama
Ran out of AI tokens again this week. Mid-task, of course.
Been working out of Buffaly instead. One place to bounce between OpenAI, Grok, and Gemini without starting over. Hit a wall on one, jump to the next.
Tired of one company deciding when my day's done.
Here is a quick FairPath walkthrough on creating a care plan.
The point is not “more paperwork.”
The point is making care-plan work easier to operate: patient needs, documentation, follow-up, and next steps in one workflow.
Watch here:
https://t.co/9VbfgtKeq9
#FairPath #CareManagement #ClinicalServices
Buffaly just shipped a new installer release 🚀
Big updates:
✅ Stable Linux installer path
✅ Stable Mac installer path
✅ Windows latest installer/update channel refreshed
✅ Better Mermaid rendering
✅ Better DataTable previews
✅ 333 commits of cross-platform fixes, provider updates, validation hardening, and release tooling improvements
This is a big step toward reliable Windows/Mac/Linux installs.
Release:
https://t.co/fmMVKhQYN6
Watch Buffaly load a data table into memory, manipulate it, export it to Google spreadsheets - without ever sending the data to the model, or writing it to disk. No other agent can do this.
I'm the sole developer of https://t.co/AlcbjOLyqo, and open source, agent that uses executable ontologies instead of text prompts. An agent that writes it's own tools, constantly learns.
If you like grok, you can use it inside inside https://t.co/AlcbjOLyqo and get Mermaid diagrams, Data Tables, Docs, and dozens of visualizations (plus thousands of tools).
Let a model like GPT 5.5 plan the project and delegate to Grok Build workers.
Looking for early adopters.
@xai - I can help you leapfrog Codex