The people running @microsoft's partner verification program are just... wholly incapable of independent thought. Does anyone have any contacts there? Been trying to get verified for weeks, and am stuck in a loop of them sending me canned replies without actually investigating or reading what I explain to them.
@zeeg fair enough, thanks. I looked through the merged PRs on the repo… I see some stacked PRs in there too. I guess that can help with larger pieces when necessary.
@thsottiaux@KyleSPatel don't love sunsetting the classic app. At least for me, I regularly work with codex on actual dev tasks *and* normal chats for more regular, casual things. Doesn't make sense to open up codex to ask "give me a list of 10 thanksgiving songs."
Anybody have a connection at Microsoft's verification department? We've been waiting weeks for our partner verification. It's way past their ETA now with no update and no channel to contact anyone about it.
@TrevorCampbell_ I got a similar one recently but it was for a recovery account with a link to a google sites page with a presumably fake google login form.
@Im_IrushiK GPT-6.1 Sol fixed a test for me yesterday that was flakey. Then it added another regression test to make sure the test didn’t become flakey again 😅
GPT just added a regression test for a fix to a flakey test. What kind of training led to it adding a test for a fix for a test? I swear I'm gonna rage-delete our entire spec folder.
sonarqube's LOC-based pricing is extremely high. Our team is certainly not getting the benefit from the cost. Any suggestions for alternative approaches for automated code quality checks?
@abatalion@Alphaschool It seems like guides were not part of the pilot programs, but teachers were still present. Not sure if they took on the same role as guides though.
@Railway pro account here. Builds are still struggling pretty bad for us. Any new update? Status page still says builds should be back to normal for us.
@voidzerodev I did some benchmarking for our monorepo earlier this month with tsrs, ts-rust, bun, and tsgo. For what it's worth tsrs is a big contender, because it gets a lot of benefits from its memory approach and incremental checks (which bun doesn't support).
The purpose of estimating software projects is not business planning its to create a self imposed deadline that creates social pressure for teams to finish the project.
This sounds insane. When I was an engineer I hated estimation and deadlines. My feeling was, it will be done when it’s done and we don’t know when it will be done. I would obnoxiously give a probability distribution whenever asked if we would ship on time.
But in reality, in a complex project with many people or teams, people need a good deadline, one that is aggressive but not insane. This is what lets everyone involved solve for what really matters, cut things that don’t matter, and push to complete something really hard.
It’s a silly social convention, but it really helps. In fact setting aggressive but not insane deadlines, slipping them when needed but holding them when possible, is one of the key capabilities in engineering management.