Devs love to dev so everyone and their mom is building their own version of Ramp’s Inspect.
It’s a *ridiculous* amount of work though.
You have to build a huge amount of features to just have a simple functional agent orchestration pipeline (that’s actually good with browser testing, cloud sandboxes that actual work, surface area across Linear and Slack, PR review, Security Review, etc, etc, etc)
I know because I built several versions of this.
Got sick of it and just decided to pay for Devin.
Never been happier or freer to just ship a ton of features and bug fixes.
You’re absolutely wasting your time if you’re not buying your agent orchestration off the shelf now.
Just pick a platform and double down - stop switching - all the frontier models are equally capable now.
Focus, double down and build.
Don’t build the tool - use the tool.
Stop playing with this stuff and just get back to work.
@amitpr@zachbruggeman@tryramp@opencode@modal@CloudflareDev yeah this is always a challenge. Part of if is unavoidable which is why speed matters a ton. Supervision is less a problem if it's quick.
We also limit agent paths that lead to wasted or inconsistent behavior by handling them out of band, like PR creation.
One of my favorite parts of building @tryramp is that we aren't afraid to build things ourselves if we think it'll be better. Inspect is better!
Check out how we're enabling all of ramp to help build with us.
The craft of engineering is rapidly changing. At @tryramp, we built our own background coding agent to accelerate faster.
We call it Inspect. It wrote 30% of merged frontend + backend PRs in the past week.
It’s powered by @opencode, @modal and @CloudflareDev. It runs fully in the cloud, and starts in seconds, letting every builder work at the speed of thought, no setup required.
Today, we’re open sourcing the full blueprint so anybody can build their own Inspect. Just give our spec to your current coding agent, and let it build your new favourite.
What happened to @codesandbox it's gone from a super convenient in-browser editor and scratchpad to a bloated mess that almost never works. I have completely stopped recommending or using it.
I'm hiring designers for @tryramp.
Who should I talk to?
If I hire them, your bonus would be enough to buy 50 @figma Creator Micros.
No DMs, no Freelancers, must be US-based → https://t.co/dRbVVmH7je
@bramus Ugh we noticed this broke in the dev build a while ago and never filed an issue bc I just assumed someone would notice and it'd not make it into stable 😬
@rickyfm@sebmarkbage more digging uncovered that the component was remounting for an instant. It was unexpected that suspending would trigger the render but ultimately never the effects. It's cool that it can tell when the suspend is done that it's not needed anymore and just bail!
hey @sebmarkbage hard to get find an answer for this, Will React ever start to mount a component, run a render, then bail without running effects? Suspense is involved...
@andrewingram I've been meaning to test this more, but the answer seems to be "definitely yes", according to anecdata in GitHub issues on Intl libs. Convincing enough that react-intl still does it.
@kentcdodds@acemarke@KomangJelte @rossipedia @sebmarkbage@TkDodo Tbf tho, they also released Lazy without preloading, and suspense in general, like 4 years too early. long enough for fun experiments to inadvertantly become "patterns" by default
@_cristianvr_@TkDodo@mattpocockuk@sophiebits Making this change without adding Lazy.preload was a huge miss. It's already stupid this API doesn't exist we have so much weird code to hack in on
@TkDodo IDK this behavior was always a crutch, it never solved the problem of waterfalls unless you pushed every single suspension to a leaf, which is not a feasible or desirable way to structure components. render initiated fetching was always a thing they said was wrong
For no reason: found is a production ready SPA router that handles data fetching out of the box and with a TON of optional knobs to match your use cases https://t.co/r1qYZQiSx1
@sebmck I'd be interested how `yup` compares, similar API tho i wouldn't have a ton of hope there as this is not a thing i've ever tried to optimize for, beyond basic perf tweaks.