Continue
507 posts

Continue
@continuedev
Quality control for your software factory
San Francisco شامل ہوئے Mayıs 2023
5 فالونگ7.9K فالوورز
پن کیا گیا ٹویٹ
Continue ری ٹویٹ کیا

@huggingface @Uber @ByteDanceOSS 6/8 📅 Mar 26 — Agent Infrastructure: From Prototype to Production luma.com/ga3zi7sx
A discussion with leaders building @dosu_ai, @continuedev, and @coderabbitai on moving agent systems from prototypes to dependable infrastructure.
English

This is what the software factory looks like in practice. Not a grand architecture. A series of small bottleneck removals, measured and iterated.
continue.dev/blueprint to learn more about our factory
English
Continue ری ٹویٹ کیا

grün/gelb/rot/grau palette from @pierrecomputer across Mission Control + we have an AI check to ensure that it stays that way forever, even when your agents work in your factory

English
Continue ری ٹویٹ کیا

Continue ری ٹویٹ کیا

a good example of an "AI check" we just set up this morning
"Migration verifier" - runs an agent to scan PRs and check for missing DB migrations
the traditional AI code review bots we use missed this, because its not a bug but rather an implicit best practice
Nate Sesti@NateSesti
English
Continue ری ٹویٹ کیا

@continuedev Hacker News launch today!
We made source-controllable markdown Agents which show up as GitHub checks, and can actually make changes. Stop writing super complicated GitHub actions.
docs.continue.dev
English
Continue ری ٹویٹ کیا
Continue ری ٹویٹ کیا

developers have the greatest opportunity in the world right now
Ty Dunn@tylerjdunn
amplified[dot]dev is a single-page essay arguing developers are amplified, not automated. One HTML file, inline CSS, vanilla JS. No frameworks. We wrote 11 checks for it. Checks don't have to protect code — they can protect taste. Here's what ours look like.
English
Continue ری ٹویٹ کیا
Continue ری ٹویٹ کیا

This where vibe coding breaks down
Good ideas come from a deep understanding of the solution space, then putting the puzzle pieces together
Geoffrey Litt@geoffreylitt
Q: Why do people need to understand the code even if it's proven to work? A: Because to have good ideas for what to do next, you need the structures and concepts present in your mind
English
Continue ری ٹویٹ کیا

This week we released Checks: agents that run on every pull request, are defined in source code, and show up as CI checks in GitHub
Want to see what checks your codebase needs? Paste this into Claude Code:
"Improve this codebase by adding checks: continue.dev/walkthrough"
English

How do you run, observe, control, and trust AI that’s operating continuously? That’s the real job of cloud agents, and it’s also where teams tend to misuse them.
The problem isn’t cloud agents. It’s knowing when they’re actually the right tool.
blog.continue.dev/when-cloud-age…
English




