"Sergey Brin’s first commit at Google, after many years. Google’s cofounder came out of retirement and made his first code change last week."
You love to see it
Software is going to melt into everything and nothing all at once
Asked Claude-3.5 to implement an example patient application using the Good Faith Estimates provided by the Patient Cost Transparency Guide (nitty gritty healthIT interop spec)
And it one-shotted a React app
Yeah, I'm thinking
@LangChainAI
is kind of a big deal... the ability to compose LLMs into a chain is incredible
This isn't even a great example, but it already overcomes the "no recent context" limitations of ChatGPT 2 months ago
@mattripkey
@drnaqi
No startups left, just three mega zaibatsus — they will build monuments to themselves, giant ziggurats the size of cities — and we will ache for some ancient heavenly connection to the starry dynamo in the machinery of night wondering what went wrong
I gave it PDFs of these pages:
*
*
Which are the technical specifications for how this is supposed to work
The idea here is that patients should be able to see, up front, what some expensive healthcare service is going to cost
It's been said before but
@karpathy
's Neural Networks: From Zero to Hero is so damn good
So many "oh, so that's what backpropagation is" type moments for me so far
It made this nice little demo app of how this will work in the future once these APIs are live and in production
An upcoming MRI Scan appointment shows an estimate link - and we can see the expected cost breakdown in the form of an Advanced Explanation of Benefits
News! 📣🚨 We’ve officially launched our new website!! Bye bye Automate Medical, hello
@Flexpa
!!
Drop on in and let me know what you think! Spot a dead-link or mistake? DM me or hit me up below! 👇
Back in 2014
@rrhoover
@eriktorenberg
et al sent individually (?) hand written notes to early users (
#3264
here) — I kept it, along with the unused sticker
LLMs are a silver bullet for (already skilled) programmer productivity — it will turn the 10x-ers into 1000x-ers, but only have marginal returns for the already mediocre
Problems that you think are solved that aren't:
* Operating systems
* Spreadsheets
* Video
"Screen sharing" should be much more like a multiplayer game than a stream.
That's why we've been piloting
@with_multi
at
@flexpa
for the last couple of months and it's launching today.
@with_multi
launches today!
Multi: Multiplayer Collaboration for macOS
- Shared control for any app or display
- Drawing & deep links for faster communication
- Native app; crafted to feel like part of macOS
Get early access:
The Riff podcast, the audio version of
@ByrneHobart
’s newsletter with
@eriktorenberg
, is as good as you’d expect
Great section on reasons to be pessimistic on AI services on top of foundation models — “the issue is what you’re competing against the cost of just writing your own
Honestly, a version of Gmail that accurately showed only emails that need a reply would be worth more to productivity than all of Google's other AI efforts
ArriveCAN app when it’s mandatory: awful, hate it, waste of money
ArriveCAN app when it lets me skip a line: incredible, amazing, digital modernization
Billion dollar idea: Figma for Docs
Why?
Make hypertext great again - our mental model of word processing is still caught BU (before Unix-time). Pages. Files. Margins. Comments that you can't preserve when you made edits.
Word processing is a massive market and it's not solved
(Even for things which *really are* just JSON blobs, like e.g. bank transactional data on a statement, the hard part has never been data interchange itself. It’s achieving multi-stakeholder buy-in, coordinating an industry around a standard, perverse incentives, etc.)
Billion dollar idea: LLM CRM
Why?
CRMs aren’t really about relationship management — they are business operating systems.
If you have absolutely perfect discipline you can throw enough people and hours at them to give you a perfect picture. The promise of process.
But
First time going to
#fhirdevdays
, first day done — every talk I saw was great
I generated some interest around a JS/Typescript meetup — seems like the demand is there so hopefully we’ll meet soon
Next up:
@Lucienneide
of
@jointherimidi
talk about clinical decision support in the primary care context at
#fhirdevdays
Nodding along hearing about care shortages and how important screening is
Interoperability is one of the biggest problems in health care. There’s a big shift happening though.
Together with
@AndrewArruda
and
@tjhammy
, I wrote an open letter on health data interoperability, and why things are changing.
there will probably be a 12 to 24 month sweet spot for lawyers smart enough to use AI, where they will be able to bill 100x the hours they currently bill, before most of that job pretty much vanishes
"I don't know who that person is, I've never read a word of what he's written."
- Every topic on
@eriktorenberg
,
@antoniogm
,
@dwr
'Moment of Zen' podcast
iOS Shortcuts is the best programming language in the world
Voice recording 👉 OpenAI Whisper API 👉 LLM 👉 Apple Notes
Entirely in Shortcuts, no app
Sorry about the audio
Tired: email marketing — boring, no one reads your messages, probably written by a bot
Wired: fax marketing — innovative, unique, impossible to avoid reading, gets attention
Copilot is an absolutely incredible product — absolutely wild how much dev mindshare MS now owns:
* Github
* VS Code
* TypeScript
(running on MacBooks, but still…)
Today we successfully tested another decision support push notification service with Epic (CDS Hooks).
Prescription drug pricing connected to a real EHR in less than 40 lines of code