Max Topolsky retweeted
Nine more lessons from my 3D book about energy, now covering the whole GCSE energy topic: kinetic, gravitational and elastic energy, heating, power, efficiency, energy resources and conservation of energy.
Built with Opus 5.5 and Three.js. The full book ↓
Max Topolsky retweeted
We’re investing heavily in evals at @sentry. It’s early but we’re seeing a culture change from shipping vibes to shipping with clear data using our evals platform.
Will share more in a future blog post, but some stats from our devs in the past month:
- ~1800 eval runs triggered directly on PRs
- 1500 eval scenarios ran across 17 AI surfaces in our product
- Cost/token reductions of ~20-50% across a few of our surfaces
- Eval passing improvements of ~10-20% from some core “code mode” changes launching soon (shoutout @grichadev)
All this from essentially 0 back when we started this effort in May.
Evals have been coming up more and more in my conversations with podcast guests and PM friends.
Nearly half of the 25 awesome PM job openings I shared last week ask for experience writing evals. And leading companies keep sharing what investing in evals bought them:
— @tryramp took its automatic receipt collection from 35% to 83% accuracy.
— @Shopify shipped an AI workflow builder that's 2.2x faster and 68% cheaper than the frontier-model setup it replaced.
— @harvey__ai rebuilt its AI contract reviewer, nearly doubling its internal quality score.
— @cursor_ai tuned its Auto Balance routing, with much higher user satisfaction at 41% lower cost.
So I asked the 🐐s of evals, @HamelHusain and @sh_reya, to write an advanced sequel to their very popular "Building eval systems that improve your AI product." Drawing on their work with 50+ AI companies, they share the key step most teams skip, what you should (and shouldn't) automate, and a free plugin that lets a coding agent do most of the heavy lifting.
Read it here: lennysnewsletter.com/p/advan…
On August 17, GitHub experienced a significant outage that disrupted developers and organizations around the world. If you were trying to ship software that day, we let you down.
I posted in March and April about the steps we’re taking to make GitHub more reliable. The work is still underway. Here is an update:
github.blog/news-insights/co…
Max Topolsky retweeted
I know devs might be looking at agent traces in other tools - we used to as well way before for Seer
But take a look at this UI - you really can’t go back to anything else
Max Topolsky retweeted
Snapshot testing has become even more valuable in the agentic world. It’s an easy way to give the agent eyes.
In case you have any doubt, here’s me checking in on a Claude remote session while doing a workout in my basement. And of course using @sentry Snapshot Testing. 😁
Max Topolsky retweeted
Tom Brady reveals the story behind his legendary training session with Ben Affleck.
"We were at Baker's Bay together, and I needed someone to catch. We were getting ready for football season. I was down there; it was the summer, it was quiet, and I'm like, 'Hey, can you catch?' And he's like, 'Yeah, I can catch.'"
"I got all these balls on that big field. They got it painted, and I'm like, 'Alright, let's see what you got.' I remember getting to the end, and I was like, 'I want to put a little pressure on and see what the hell happened.'
"I was like, 'Alright, fourth down, Super Bowl's on the line, you know. I'm calling your number. This is the route.' I think it was a flag route I gave him. I'm like, 'I'm gonna throw it, see if you catch it.' And he got running, and I put one up there, and he stretched his arm out, and he caught it."
Max Topolsky retweeted
Snapshots is now live in @sentry 🥳🥳🥳
Blog + details 👇
Max Topolsky retweeted
📸 Sentry Snapshots is now in Beta!
Snapshots (originally from @emergetools) lets you diff screenshots of your app on every commit and block PR if something looks off (just in case your AI agent broke your UI and you had no idea). Works on any platform with a frontend.
Screenshot diffs and images are available to your agents via Sentry MCP or API 👀
Check out the full blog post 👇
Max Topolsky retweeted
Potentially massive day for the "why is the app so small on Android but so big on iOS" posts
Max Topolsky retweeted
More of the iOS app loop, now inside Codex.
The Build iOS Apps plugin lets Codex view and test your iOS app in the in-app browser, open SwiftUI previews, and hot reload edits without leaving Codex.
in a code block, if you highlight text and type a quote, it replaces the text rather than wrapping it in quotes, another piece of polish I'd expect in @linear but isn't there 🤷♂️
pretty annoying that if I'm looking at valid JSON code block in a @linear ticket I can't easily collapse properties. Would expect this is the type of detail linear nails so can y'all add it
Congrats to @tryramp for my least favorite cold email I have ever received
- Slimy: fake "fwd from {co-worker}"
- Double Slime: no unsubscribe button, says to reply
- Incompetent: tags of {first name} and {company} left in
- Lazy: haven't worked at that company for 4 years
pretty annoying that if I'm looking at valid JSON code block in a @linear ticket I can't easily collapse properties. Would expect this is the type of detail linear nails so can y'all add it
i made fun of it at first, but this was really easy and helpful here @jshchnz
xitter.cf/jshchnz/status/2029612…
Max Topolsky retweeted
One area I haven't seen AI explored much is exercising. Models have gotten quite good at building structured workouts. I built @race_day_ready on the side to take this on - I just built my lunchtime @GoZwift ride just by chatting with it!
Max Topolsky retweeted
🤫 i heard there are rumors that @steipete may be giving a talk at Agents Anonymous 🕵️♂️ next week in SF
if you've been experimenting with tools like Claude Code, Codex, Cursor, or custom agents, this meetup is for you
spots are very limited... RSVP ASAP