One tiny nanoscopic nitpick, because i agree with you mostly, programming is often creating wider things (abstractions, frameworks). I think it hits a different part compared to most jobs. Maybe... i'm not sure, but that's how i feel compared to other manual occupations that i loved too.
Has anybody written about this ? in fiction or as report even. It seems obvious the current techbros are only thinking about a radical shift where labour changes meaning and human societies are irrelevant for those who owns datacenter and have pocket deep enough to buy the rest when people can't sustain their own lives.
Version control might not be a big deal if you are all-in on the database.
Stored procedures are easiest to version by simply defining multiple variants and then incrementally moving the callers in the direction you want. The durability comes from (hopefully) your backups. Point-in-time-recovery is often easier for the business to reason about than a git repository.
Having worked for a business that made a serious go of running everything out of stored procedures, I have to say that lack of version control was a huge problem and effectively limited all development to a single person who held all the rules in their head.
This is essentially what we do. All of our stored procedures, views, etc are kept in *_api schemas. Those schemas just get fully dropped and reinstalled whenever we migrate. It works really well, and has basically zero cognitive load overhead for developers.
Not trying to defend ai but I observed another mode: what used to be bored dev chats where people avoided topics or started feuds, now it's "well Claude suggests...". It's not great but it's a short form of improvement somehow. (Sure I'd prefer passionate convos steering toward innovation, but that's been a rare sight in my career)
No it isn't an improvement. If I wanted the output of an LLM instead of a thinking, smart, real human, I would have simply asked an LLM. Nearly nobody who asks humans questions WANTS to get an LLM answer, that's simply not why people ask other people.
I agree with your sentiment from the perspective of the asker.
But if I'm the askee, I honestly don't know how to navigate those waters yet.
If someone asks me for help and I can find, through AI, a thread to explore, but I don't have time to explore it myself, should I not share?
Do I say "Have you tried X?", where X is the thing the LLM suggested? Should I pretend that I did not ask the LLM?
In the past, I could find some source and send them the link, and I wouldn't assume the person had exhausted the entire Google index. Sending a link isn't the same as LMGTFY.
Analogously, while "Claude says X" does sound as rude as lmgtfy, disclosing that your suggestion was found via llm is more akin to linking to a source, or "take this with a grain of salt".
I think as long as nobody can tell you've used an LLM, it's fine. If you use an LLM to get your info and then respond, that's normal. Copy pasting or going "claude said" and then more or less just regurgitating is different, because you are no longer involved.
If I ask you, I want your thoughts, based on whatever you can find or know. It's difficult to articulate for me, honestly, but I hope it makes sense.
LLM integrate more kinds of information and allow bidirectional debates. I can't ask Google to do the same search session but change some parameters from far earlier.
ps: I'm not pro centralized corp. owning data and ai. But so far they are the cheap highway to answers
reply