Intelligent Community Platform @airbnb. Ex-Apple/Sun/Netscape/Microsoft/Startups. https://xitter.gq/t.co/aezeSqyPjj. https://xitter.gq/t.co/RXqvB8I4SQ.
us-west-1
Joined November 2006
- Tweets2K
- Following943
- Followers976
- Likes3.5K
Pinned Tweet
See if you can spot when I started playing around with Opus 4.5 as the agentic core of a github-actions-based software factory
I feel I have the miles to defend this: Monster (2019 re) is the best album by America’s best late 80s - late 90s band.
Happy Mother's Day mom. You you are still with us I miss you so much.
Roger Waters & Sinead O'Connor - Mother youtu.be/gZ2tluarzZs?si=8ojX… via @YouTube
Jeremy Werner 🏴☠️ retweeted
To get to the Flipbook visual style you see here, we had to go through literally hundreds of iterations. Here are some of my favorite in thread
Imagine every pixel on your screen, streamed live directly from a model. No HTML, no layout engine, no code. Just exactly what you want to see.
@eddiejiao_obj, @drewocarr and I built a prototype to see how this could actually work, and set out to make it real. We're calling it Flipbook. (1/5)
Jeremy Werner 🏴☠️ retweeted
HTML Image Maps on steroids.
So much brilliant innovation to unpack here:
- everything is a link
- the browser's address bar reinvented as a tree
- responsive design? just a different image aspect ratio
- infinite rabbit hole
- educational & approachable: a 3 year old can use it
Imagine every pixel on your screen, streamed live directly from a model. No HTML, no layout engine, no code. Just exactly what you want to see.
@eddiejiao_obj, @drewocarr and I built a prototype to see how this could actually work, and set out to make it real. We're calling it Flipbook. (1/5)
Last year, most people were skeptical we'd be able to trust AI to write code. That changed in December. This year, people are skeptical we'll be able to trust it to review the code....
Jeremy Werner 🏴☠️ retweeted
San Francisco is now a model for how to fight crime.
A few years ago it averaged 86 car break-ins per day. Today: 15.
SF did two things:
1. Got a DA that prosecutes criminals: Following the successful recall of Chesa Boudin, DA Brooke Jenkins started prosecuting prolific offenders and said so loudly. Crime dropped every year since she took office.
2. Put tech to use: In 2024, SF activated 400 license plate readers and deployed 80 drones citywide. This tech feeds officers live intelligence on suspects in motion. Drones alone have assisted in 1,000+ arrests since then. The technology lets authorities solve crimes as they happen rather than depend on much more intensive, legally perilous post hoc investigations (which ironically are often more intrusive than using tech).
The results:
- Car break-ins down 85%
- Robbery down 30%
- Burglary down 33%.
- Homicides hit their lowest level since 1954.
Plate readers, drones, a prosecutor who prosecutes. That's the whole formula!
Austin has the opposite approach. License plate cameras are effectively banned. Jail bookings are down despite repeat offenders victimizing innocent people regularly. Bond violations went from 37 in 2020 to 250 last year.
SF proved crime is a choice. Austin, so far, keeps making a different one.
The following pattern has been giving me vertigo: launch X to get a bookmark I saved an hour ago about a new agent harness / mac gpu optimization approach / application of new idea X to novel area Y that I was thinking about myself just last night = Z half-explored projects
how can it be that country music is rocking harder than rock or indi right now? and while we are talking about the much-maligned format, can you show me any artist in any form that has put out 156 x avg-tracks-per-album pieces? pitchfork.com/news/willie-ne…
Jeremy Werner 🏴☠️ retweeted
I know Silicon Valley startups don't want to hear this.....
But the combination of someone in the trades with deep domain expertise and Claude Code will run circles around your generic software.
I talked to Cory LaChance this morning, a mechanical engineer in industrial piping construction in Houston. He normally works with chemical plants and refineries, but now he also works with the terminal
He reached out in a DM a few days ago and I was so fired up by his story, I asked him if we could record the conversation and share it.
He built a full application that industrial contractors are using every day. It reads piping isometric drawings and automatically extracts every weld count, every material spec, every commodity code.
Work that took 10 minutes per drawing now takes 60 seconds. It can do 100 drawings in five minutes, saving days of time.
His co-workers are all mind blown, and when he talks to them, it's like they are speaking different languages.
His fabrication shop uses it daily, and he built the entire thing in 8 weeks. During those 8 weeks he also had to learn everything about Claude Code, the terminal, VS Code, everything.
My favorite quote from him was when he said, "I literally did this with zero outside help other than the AI. My favorite tools are screenshots, step by step instructions and asking Claude to explain things like I'm five."
Every trades worker with deep expertise and a willingness to sit down with Claude Code for a few weekends is now a potential software founder.
I can't wait to meet more people like Cory.
If you think you're too much of a manager/exec to play and learn check this out. No wonder Shopify is making the agentic transition so fast. Time for everyone to be getting hands-on. It takes a qualitative feel to really get the pace of change right now. Reading won't cut it.
OK this thing is totally insane. Before going to bed I...
* used try to make a new qmdresearcher directory
* told my pi to read this github repo and make a version of that for the qmd query-expansion model with the goal of highest quality score and speed. Get training data from tobi/qmd github.
* woke up to +19% score on a 0.8b model (higher than previous 1.6b) after 8 hours and 37 experiments.
I'm not a ML researcher of course. I'm sure way more sophisticated stuff is being done by real researchers. But its mesmerizing to just read it reasoning its way through the experiments. I learned more from that than months of following ml researchers.
I just asked it to also make a new reranker and its already got higher base than the previous one. Incredible.
Just sayin'
The more I think about this Xcode announcement the more it feels like a year behind the market which is like 5 years in non-ai terms. This is where the puck is: xitter.gq/rork_app/status/202457…
This new trend of the U.S. companies launching new models before everybody takes off for winter break, and the Chinese companies dropping new models right before Chinese New Year is so interesting, and makes so much sense.
