Andrei

Trying to stay relevant in Big Tech, also hi-end audio and photo

[Andrei @andrei_32510]

@Joi2James https://t.co/mKNqCc3Ko2

[Andrei @andrei_32510]

@lightsilver323 https://t.co/zfgq6Fqmar

[Andrei @andrei_32510]

This is how you deal with an overload as an EM. https://t.co/4bSLpwPmmA

[Andrei @andrei_32510]

@Alecco Yeah :(

[Andrei @andrei_32510]

I built https://t.co/hxcGTae8Pq Now you can share you collections of stuff

[Andrei @andrei_32510]

I built a local-LLM metal discovery tool on my MacBook and, somewhere along the way, learned that the most interesting part wasn’t running huge models locally - it was figuring out exactly where they actually make software better. https://t.co/ziUQsnnwuj

[Andrei @andrei_32510]

AI coding is becoming software’s 3D printer: making deeply personal ideas worth building. I wrote about HavenOS, a native macOS app that turns a Mac mini into a private home library—and what it taught me about reliable systems, backups and iteration. https://t.co/FfNwrZBSI5

[Andrei @andrei_32510]

@AndyjnrUmaru Diablo 2&4

[Andrei @andrei_32510]

@heyblake https://t.co/kj9Re7FnAV HavenOS installs and manages private services for books, music, movies, and files. Keep the collection on your hardware, use it around the house, and skip the server-admin ritual.

[Andrei @andrei_32510]

@catalinmpit 💪🏻💪🏻💪🏻🇳🇱

[Andrei @andrei_32510]

Don’t waste time on prompt engineering- build your own harness.

[Andrei @andrei_32510]

AI-native developer ≠ prompt engineer.

The key idea is that AI does not remove the need for software engineering judgment. It shifts the engineer’s role one level up.

An AI-native developer is less of a pure “code producer” and more of a system designer, orchestrator, and verification owner.

The important skills become:

Vision — clearly define what needs to be built, why it matters, and what constraints exist.

Direction — break work into small, clear tasks that AI tools or agents can execute safely.

Verification — review, test, challenge, and validate the output before it reaches production.

The biggest risk is not that AI writes code. The biggest risk is that humans become passive approvers of code they do not understand.

So the future of software engineering is not “less engineering.” It is engineering with more delegation — and therefore more need for architecture thinking, testing discipline, debugging skills, security awareness, and ownership.

AI can accelerate delivery, but accountability still stays with the engineer.

That is what makes the AI-native developer interesting: not faster typing, but better judgment at higher leverage.
https://t.co/HbnjAWyrVk