andrej.fidel

Fullstack developer, software enthusiast, founder @codecannondev

[andrej.fidel @AndrejFidel]

Stop ignoring those dependency PRs.

Dependabot can handle them while you ship real features. 🚀

👉 https://t.co/RdjsI3AFuy

How do you handle updates?

[andrej.fidel @AndrejFidel]

@LaraconUS @taylorotwell Define your model columns and relations and CodeCannon generates a v0.1 codebase for you. Get CRUD UI and API, Auth, Linter/Formatter/Test suite configuration, basic Docker/Kubernetes configs, and a codebase made for developers who appreciate code https://t.co/2DHhHK65Te

[Тsфdiиg @tsoding]

What I discovered by letting my normie relatives use my laptop is that if you use Linux with i3 window manager you don't need a screen locker at all.

[andrej.fidel @AndrejFidel]

@aarondfrancis Discrete checkouts all the way

[andrej.fidel @AndrejFidel]

If you think about it, graveyards are real life memory leaks 🤔

[andrej.fidel @AndrejFidel]

@MatijaSosic Sync engine and all, or just the general functionality? Would be interesting to see how long it takes..

[andrej.fidel @AndrejFidel]

Woud love to see a gpt 5.2 banchmark, I've been having better luck with vanilla 5.2 than 5.3 codex on frameworks like Laravel, probably because the framework abstracts a lot of concepts into something resembling human speech rather than algorithms or competitive coding style. Haven't done deep dives with Nuxt yet though.

[andrej.fidel @AndrejFidel]

My god. Had a semi nightmare - going through the familiar scene of hanging out with friends before a test I know I'm going to fail because I didn't study. Must be all the AI I've been using.

Learning AI feels a bit like cheating. Not really doing the thing but finding creative ways to circumvent the requirements and get a passing result.

It doesn't feel like craftsmanship to me. Maybe that's why I'm getting nightmares from a time when my life constantly felt out of my hands.

Maybe it's time to take an AI break for a couple days.

[andrej.fidel @AndrejFidel]

AI is a tool. The messy part is the humans and the expectations around it.

And of course the dishonest marketing implanting those ideas.

[andrej.fidel @AndrejFidel]

A lot of workplace “stability” was vibes. When leadership is changing direction every month, the vibes are gone.

When the plan changes constantly, it stops feeling agile and starts feeling like whiplash.

I think we’re gonna find the limit of how much chaos teams can take. We’re just not there yet.

Feels like we’re all re-learning how to work in real time, and it’s kinda exhausting.

Lowkey loving this dynamic. I've got the right brain type for this industry, but not everyone loves chaos.

I wonder who blinks first, devs trying to meet unrealistic ai outcome expectations or leadership sweating as the AI expenses keep climbing with adoption and more and more expensive models and harnesses with often questionable productivity gains.

[andrej.fidel @AndrejFidel]

Even when AI writes a large chunk of code, you can still make stuff you’re proud of, but you have to protect that on purpose. It doesn’t happen by accident.

[andrej.fidel @AndrejFidel]

Using AI properly often feels like experimenting until something works. Which is awkward when your sprint plan wants certainty.

A lot of AI tooling is still janky. Half the time you’re not “building,” you’re just poking it to see what it does today.

The only way to learn what works is to try stuff and fail a bunch. Super fun when you’re also expected to be “highly predictable.”

Not so different from adopting any new tech tho. But because of the relentless marketing people expect phd results yesterday.

Hard to explain to non tech people that tech is still inherently hard...

[andrej.fidel @AndrejFidel]

Companies are kinda in culture shock. Nobody knows what the “right” way to use AI at work is yet, but everyone wants results yesterday.

[andrej.fidel @AndrejFidel]

@glcst Something like this is the best solution I've seen so far I think https://t.co/IPT84R8ckn

[andrej.fidel @AndrejFidel]

Started using ChatKit from @OpenAI and have been running into plenty of bugs and poor documentation issues. Is there a human one can discuss this with? We want to lean heavily into agents and chatkit but lack of documentation and clear guidance on troubleshooting is making this really hard. Would love to work alongside the team to help surface some of the issues we've been facing.

[andrej.fidel @AndrejFidel]

Daily update 34

[andrej.fidel @AndrejFidel]

Daily update 42

[andrej.fidel @AndrejFidel]

Daily update 41

[andrej.fidel @AndrejFidel]

@zeeg We are https://t.co/2DHhHK6DIM

[andrej.fidel @AndrejFidel]

Daily update 39