tests generated by ai, especially which generates each time codex are completely useless and needed only for beautiful "44/44 tests passed"

I disagree. Personally, I would work in company which respects more code/infra quality than coding speed. When you vibe-deliver a lot of features per sprint after some time codebase becomes an unpredictable AI Slop.

> I would work in company which respects more code/infra quality than coding speed

Same. But that doesn’t have to be in contrast to using AI tools. There’s a lot of grunt in day to day coding as well.

The problem is not the tools

The problem is the tools. The tools encourage you to hand over all control to the AI, taking the user out of the drivers seats resulting in slop. If we were more honest about where the utility of using LLMs for coding, we wouldn’t be burning nearly as many tokens.

What's your grade?

Senior. So I still get coding tests. I wouldn’t worry if I’m Staff and above.

I would. Just finished up a job search recently for staff+, and still got leetcode interviews for most companies.

Thanks. That's tough. Did you only interview at big tech?

Big, medium, and startups. Big and medium definitely biased more towards leetcode, but I also got it at around a quarter of startups I applied to.

Thanks for sharing. This is really surprising. Sigh, we still have to waste time grinding on these pointless stuffs. Just wish I get my FU money and never look back.

Good luck to you!

"LLM-summarized" lmao

I'm not especially anti-LLM but emergency messages seems like a terrible use case. If you're typing out a detailed emergency message then I would think you would need all of your details to be sent or you could simply type a shorter message.

Do you know another way to fit a 3000 long message into a 160ish SMS?

Don't write long messages in an emergency. If it can't fit in one message, it probably doesn't matter. Instead of trying to predict everything, allow communication. Let them figure it out, don't command them.

Also, you might consider using existing messaging app. Then you can interact if have them install the app. Matrix might work. I think Matrix can do SMS.

Why 3000 ? Set a character limit. If the detail is important they can just send 2.. or 10.

19 SMS messages.

Does anyone still pay per text?

> Does anyone still pay per text?

It makes much more sense outside the US than to pay for a text bundle. My phone plan includes 0 SMS messages, and I don't know what was the last time when I've had to send one was. It's only useful when someone doesn't have internet, which only happens once every few months at most.

You are generalizing quite a bit here, not everywhere "outside the US" is the same. Here almost all phone plans that aren't data only come with unlimited SMS - and we never paid for receiving them anyway.

write a shorter emergency message?

"I was in a car accident, come to general hospital downtown"

Ok I understand the usecase now.

I would have used two textboxes, Title and description, but this works as well.

btw, you can claim "relax" name instead of "relaxai" on pypi

pypi.org/project/relax is abandoned library, which owner registered via email with expired custom domain, so you can claim this domain and reset owner's account by email.

It's probably a better idea to follow the process documented in PEP 541 [1] and contact the PyPI admins to request a transfer of the name. Taking over the domain to impersonate the original owner would look indistinguishable from a supply-chain attack.

[1] https://peps.python.org/pep-0541/#how-to-request-a-name-tran...

Yeah, I noticed this library few years ago when checking pypi.org for supply chain attack vulnerability and scanned all libraries. There are a lot of such libraries which you can take over.

Thank you! We'll check this out.

And create yet another completely non-descriptive package name?

If everybody presses red button - everybody survives. Ez solution

I noticed another limitation: "An image in the conversation exceeds the dimension limit for many-image requests (2000px). Start a new session with fewer images."

So I can't continue my claude code session I started yesterday.

Double tap ESC and revert the conversation.

makes sense, “a picture is worth a thousand tokens” as they say. They probably lowered the limit due to capacity issues.

just use free antigravity subscription with Opus 4.5 and reverse engineered API, and bunch of cheap google accounts

They should measure for different temperatures, where at 0 it will be the same output every time, but it's interesting to see how results will change for different temperatures from 0.01 to 2. But, I'm not sure if temperature is implemented the same way in all llms