egeozcan

You cannot call all the bullshit. You need to call what's important for you. That defines your values.

It's also important to base your actions on what's at hand, not teaching a lesson to "those people".

I thought assuming good faith does not mean you have to sympathize. English is not my native language and probably that's not the right concept.

I mean, do not put the others into any stereotype. Assume nothing? Maybe that sounds better. Just look at the hand you are dealt and objectively think what to do.

If there is an arsonist, you deal with that a-hole yourself, call the police, or first try to take your loved ones to safety first?

Getting mad at the arsonist doesn't help.

IMHO, you should deal with actual events, when not ideas, instead of people. No two people share the exact same values.

For example, you assume that guy trying to cut the line is a horrible person and a megalomaniac because you've seen this like a thousand times. He really may be that, or maybe he's having an extraordinarily stressful day, or maybe he's just not integrated with the values of your society ("cutting the line is bad, no matter what") or anything else BUT none of all that really helps you think clearly. You just get angry and maybe raise your voice when you're warning him, because "you know" he won't understand otherwise. So you left your values now too because you are busy fighting a stereotype.

IMHO, correct course of action is assuming good faith even with bad actions, and even with persistent bad actions, and thinking about the productive things you can do to change the outcome, or decide that you cannot do anything.

You can perhaps warn the guy, and then if he ignores you, you can even go to security or pick another hill to die on.

I'm not saying that I can do this myself. I fail a lot, especially when driving. It doesn't mean I'm not working on it.

I'll give you an insider info: There's no guy! Your response would be filtered away by the profanity filter and nobody working in Telekom will ever read any of it.

Hell, I can even say, likely, nobody will ever read it, regardless of how you answer.

Those companies only respond to lawyers.

In life, you have to pick your battles.

I like using it at work. I probably would use something lighter if I ever move off github though, as I've seen the IT throw more and more resources at it for simple usage with just 30-40 MRs per day.

Also the free version doesn't have PR requirements or multiple reviewers etc.

I was part of a big software development team once and that necessity I felt there, namely, being able to let go of the small details and focusing on the big picture is even more important when using llms.

I think it won't be bigger than the giant set of rules people are supposed to read through (they never do) when onboarding.

At least with AGENTS/CLAUDE.md file, you know the agent will re-read those rules on every new session.

You can see if a test is changed. If you're paranoid like I am, you can simply make the tests read-only for the user you're running the agent from.

I added a bunch of lines telling it to never do that in CLAUDE.md and it worked flawlessly.

So I have a different experience with Claude Code, but I'm not trying to say you're holding it wrong, just adding a data point, and then, maybe I got lucky.

I think you can keep the vibe coding totally under control with good tests. I'm baffled how such a huge regression would not be caught.

Then again, the google home page was broken on FF on Android for how long?

That's something a horse pretending to be AI would say.

We don't know it, up to the point we observe it.

Which was right, today gopher has more users than ever! :)

I've been developing an rts using coding agents https://egeozcan.github.io/unnamed_rts/ - lots of borrowed ideas from red alert.

it's super buggy and yet doesn't work on mobile but it's been my childhood dream to make an rts and I have something I can have fun adding different dynamics.

it's much more fun to watch the agent struggle developing an AI instead of play the game itself (tried that at the beginning too).

Some things you build not because they are necessary, but because you can.

the site really needs something like

img { max-height: 100vh !important; width: auto !important; object-fit: contain !important; }

jQuery but gets compiled out like svelte... Not a bad idea at all.

It's easier to write with LLMs. One-off projects (the way I treat userscripts) is where they really shine.

Oh the horrible things I do with Instagram...

I learned about this thanks to claude always using it.

My driving skills are probably below average. I really like that my car warns me of zebra crossings and can follow the car in front of me with a safe distance.

Many of the modern car features are just useless marketing fluff, but there is some really good progress too.

> I have serious doubts as to the ability of AI to ever have original thought, which is a core requirement of being a programmer

If you find a job as an enterprise software developer, you'd see that your core requirement doesn't hold :)

claude opus 4.5 is much more impressive when used in claude code. I also tried it through antigravity but from a users perspective, claude code is magic.

For-of loops are fast enough, most of the time.

You probably still don't want to risk it in a tight loop. Relying on those optimizations may not be the most reliable choice when you think about all the different JS engines, platforms, system load, operating systems...

Yes, foreign intervention worked wonderfully in Iran, Iraq, Afghanistan and other parts of the world. We should also thank Russia for trying hard to extend its thriving democracy to Ukraine.

Just like how Denmark and Greenland stole American land that happens to be where Greenland is. Or Canada.

Seriously though, even the imperial ambitions from the guy feels racist :)

I guess Turkey can stop worrying on thanksgiving days.

I have a lot of conflicting views with both the "left" and the "right" these days, but it seems the so-called "conservatives" are not that conservative in their ambitions, no?

I can't prove it (because they show no stats) but I feel like you get more Gemini Pro with a normal subscription than Opus with a Max plan.

Maybe the new more efficient models made it better for Claude users but that was my experience a couple months ago.

For professional usage though, Calude Code is so much ahead of Antigravity that it didn't even make sense to make a formal comparison. That, even when using the same model (Opus).

I thought it was a specific prompt that breaks it, and that it's just something they never tested against, but when I saw it happen in antigravity, which supposedly must have been tested with a very specific use case, then I was very surprised.

Gemini is very prone to go into an infinite loop. Sometimes, it even happens with Google's own vibe coding IDE (Antigravity): https://bsky.app/profile/egeozcan.bsky.social/post/3maxzi4gs...

I needed to recreate it recently because summary element does not allow headings inside, because it has a button role.