Sort replies: Relevant Recent Liked
Replying to @ryancarson
Another great article Ryan, thanks for sharing, I always learn something from everything you write.
1
4
2,327
cheers!
3
1,830
Replying to @ryancarson
love this, does google ever invalidate the debug profile session randomly? or force reauth after a while
1
2
3,744
Sometimes yes, but overall works well
2
2,530
Replying to @ryancarson
Will give this a spin
1
1
3,344
Nice - let me know how it goes
1
3,190
Replying to @ryancarson
why you need a presigned user in the chrome profile, couldn’t let the agent also do the oauth login?
1
1
2,289
I have 2FA so there’s interactive user interaction
2
2
1,803
Replying to @ryancarson
Very cool - of course there are some bugs that don't appear for a fresh account but may appear over time as an old account edits the same thing a lot etc so also having another path that does NOT start from fresh each time might be a good addition
2
259
Replying to @ryancarson
Sounds exciting..
1
762
Replying to @ryancarson
Can’t you believe we are there now? “Point your agent to this to set it up” Holy smokes
1
138
Replying to @ryancarson
based
417
Replying to @ryancarson
Real question: How do they triage failures into: •product regression •test infrastructure •third-party outage •flaky model behavior •environment drift
2
249
Replying to @ryancarson
@snack_man interesting read, still have to look into playwrite but also exploring other options for testing workflows.
1
2
248
Replying to @ryancarson
that’s pretty cool! it would be a good idea to move the security disclaimers at the top, considering what happen with openclaw 😅
1
919
Replying to @ryancarson
I still don’t get why not to use Playwright for this scenario. Typically, you should use aria selectors (Playwright best practices), not CSS selectors. Aria snapshots could be very handy in many scenarios.
1
154
Replying to @ryancarson
Learn from you each post! Appreciated
191
Replying to @ryancarson
It’s worth leaning on @maestro__dev for e2e scripting
1
281
Replying to @ryancarson
can you provide the github
64
Replying to @ryancarson
can i get github
86
Replying to @ryancarson
the bookmark-to-reply ratio here says everything — 700+ saves, 11 replies. people want this but don't know where to start. the real unlock: AI agents aren't just code generators anymore. they're becoming QA infrastructure. daily regression testing that catches bugs before your first coffee is the kind of compounding value most teams haven't figured out yet.
150
Replying to @ryancarson
how do you create the images for you articles? They are so awesome.
27
Replying to @ryancarson
In the sea of slop articles, one that is actually useful, at least for the couple thousand people out there that need this.
261
Replying to @ryancarson
For the Prompt ➡️ AI magic ➡️ Verify pipeline to flow fast, Verify has to keep up. The tank was a response to the machine gun. At the moment there are a lot of guns and not a lot of tanks.
49
Replying to @ryancarson
Which database do you use for your apps?
124
Replying to @ryancarson
Why not run it in CI before new deploys?
254
Replying to @ryancarson
Will give this a spin!
190
Replying to @ryancarson
Thank you for this Ryan.
73
Replying to @ryancarson
Banger after banger dude
235
Replying to @ryancarson
Thank you! But it's ironic most agents cannot access this article because it's on X. Are you publishing this elsewhere as well??
76
Replying to @ryancarson
Love this idea Ryan. I'll enhance my current QA agent with some of these improvements and try it on my projects. Thanks for sharing!
174
Replying to @ryancarson
Running something similar. Agent tests my entire app flow every morning and drops a report before I wake up. The peace of mind is unreal. You stop dreading Monday mornings when you know nothing broke over the weekend.
118
Replying to @ryancarson
Ryan you should add this to crustipedia.com
56