@jxnlco@mweinbach 1) UI/UX significantly better (even dashboard viz) 2) Non-technical business writing/analysis. Give some data to both models and ask them to use the Google doc skill to write an operating plan. Opus much stronger. I’ve thought GPT>Opus for coding since 5.2, but still weaker here
@karrisaarinen@linear Give this a try. Biggest complaint relative to GitHub's UI is that the screen is too busy. When I open the review diff, I have 3 panels to look at - 2 of them which I don't really care about (your nav + list of PRs to review). If I do Guide mode, I have 4 panels.
Given the recent burst of activity around enterprise pricing and contracts, I think April 2026 was the month when both OpenAI and Anthropic found product-market fit https://t.co/0hgMfu9Khx
@carllerche 2) writing bugs related to it. I was blown away the first time I had an agent write Rusr code because the reduced turns and cruft related to no nil was noticeable
@carllerche I’ll caveat this by saying that I’ve probably written as much AI generated Go code as anyone out there. Codex/Claide are really, really good at Go. The one thing that’s incredibly annoying is the extra cycles it spends 1) writing Go code I wouldn’t write guarding for nil
@chaliy I'm probably not the right person to own a distribution of it to be honest. Hoping that sharing it might encourage some exploration by people who know Postgres better than me
Forked an experiment for an ephemeral Postgres for faster unit testing (https://t.co/z40XEC4b1c). 2.5x faster with no changes; 3.3x with minor schema setup changes. Day's worth of background work; imagine someone who knows PG internals could get a much larger improvement
@_lopopolo@steipete What's the limitation in review mode where it only returns 1-2 things each turn and then I have to keep running it until it runs out of things to find? Has to better a better way here. Run "/goal Keep reviewing until you run out issues to find"?
@valigo They ported this in a week. Once the test suite passes, then you start to refactor. You think if they spent the next 4 weeks refactoring then it wouldn't be maintainable?