AI musings and maximum cringe

I have started to observe developers virtue signaling when something is done without the help of AI (code, technical writing, even blogging).

This reminds me of a story I always tell friends about when I was doing the FB bootcamp right after I started working there. In one of the 3rd week classes, which was about how Android development was done at the company, one dude came into the class (where nobody knew anybody else) and completely unprompted, started babbling about how he was better than everybody else there (which he didn't know BTW) because he could code in vim faster than any of us with an IDE. Sure, buddy. It was the most surreal stupid thing I've ever seen another dev do (at least in person). Maximum levels of cringe. Of course, everyone ignored his ass.

Your work is your work, I don't really care what you used, as long as the result is good and you 100% understand what you did and why.

I'm not the biggest proponent of AI, and IMO generative AI, while it has some uses, is not trustable at all (and likely won't ever be). I also think these companies betting everything on gen-AI coding a whole app or whatever are delusional (although they must know, and are just willfully ignorant trying to game the market or boost valuations).

And, of course, the constant slop populating the internet, making everything dumber and search results useless, is a big problem. AI translators can't substitute humans (and this is why localization companies betting heavily on AI still use humans to do "machine translation post-editing", aka fix the shitty mess the AI has produced). Image generation can create fun stuff, but mostly slop, and can't substitute a good human artist. Suno, while fun to use to make songs about farts and people you know, can't compete with a real musician.

However, AI is still useful when used as a tool. If I have to write a 70-pager as a non native English speaker, I'm reassured if I have some suggestions about how to improve the language (and fix all my misspellings) and suggest me what to improve. Reviews will still be done by humans, I just want to save me the trouble of shipping cringy mistakes. Bear in mind, I sent emails with pretty embarrassing typos like "Kind retards, Nacho" instead of "regards", or "Salidos" instead of "Saludos".

If I use it to help me outline a document, it's still me writing the document, I'll have to fill the gaps and shape it as I want it, but it'll save me time.

With code it's kind of the same. If I have to write repetitive code, if some AI can help me with that, that's alright. I read all output, if it matches what I wanted to do, LFG.

❗

Big IF: I think it's risky to go all in with coding with an AI agent if you are just starting your career, or learning something new. You have to always have in mind that AI might be hallucinating, and suggestions are not to be taken as gospel.

I lately used Supermaven and I'm happy with it, even if I might be disregarding lots of its suggestions. But any keystroke I don't have to do is a net positive.

And by the way, nothing in this post was written using AI. I want my prize now.

I'm just too lazy to copy-paste the contents to ChatGPT.