@MotorcycleGuy@med-mastodon.com Patient, Standards Guru for @PointClickCare, tweets represent my opinions.
In the mil'-of-nowhere-ville
Joined July 2009
- Tweets36.8K
- Following483
- Followers7.9K
- Likes546
Pinned Tweet
Via @hcittoday and @Colin_Hung
Keith Boone of @PointClickCare introduced a Data Modernization Offering that helps public health agencies:
🔹 Convert legacy HL7v2 data to FHIR
🔹 Modernize infrastructure & data accessibility
🔹 Strengthen reporting
healthcareittoday.com/2025/0…
Once again, my blog makes feedspot's top Health IT Bloggers list: bloggers.feedspot.com/health…
OH at #HL7WGM: US-PCS is the US version of the International Patient Summary was published last week. “We are finally in the game.” says @ONC_HealthIT’s Matt Ron.
FWIW, the @HL7 Continuity of Care document started this game.
I will also give a shout out to the Washington State DOH who entered the IPS game a year ago and more.
Hah, my stack is deeper than my agent's. More specifically, I can hold two overlapping very similar yet distinct thoughts in my head at the same time without confusion, but the agent gets them all bolixed up together.
This makes recursion a process I can understand better than my agent.
Keith W. Boone retweeted
"Semantic Fragmentation" is a new word I learned today at the @HL7 WGM in today's Plenary Session! We've been building to create successful semantic interoperability solutions all these years, and yet we are still dealing w fragmentation.
"Sorry wrong agent" is something I've said so often they are beginning to predict it when I say something off the wall outside of their current context.
Finding a security vulnerability in your code and ignoring it is like failing to recruit Halsin in Act 1 of Balders Gate III and damning the world to disaster.
I swear today, I thought I briefly saw Claude code flicker "* hallucinating ..." in its little spinner for thinking text.
As an representative example of model Bias: I find it incredibly distressing that AI Agents continually decline to take responsibility for pre-existing bugs. It says something about developer habits that have already been trained into them.
Give this playlist a listen: Forging open.spotify.com/playlist/3n…
The calendar telling us when the new phonebooks will be published is finally back.
reginfo.gov/public/do/eAgend…
@healthapiguy covers it more here: healthapiguy.substack.com/p/…
It's more like "How will businesses in this world survive Gen Z." Gen Z isn't going away, but businesses do NOT have the same guarantee.
Coding agents such as @ClaudeCode and @GitHubCopilot have yet to learn about the uselessness of over-prompting for permissions leading to alert fatigue. Too many alerts leads to ignoring them.
Also, don't tell me the scripts are malformed, the agent should tell me instead how it exceeds limitations. If the agent cannot parse the scripts it is asking me to permit, it should just say so. Malformed implies incorrect syntax, and that's not true.