Hacker News Reader

Comment by victor106 | Hacker News Reader

> SQL is an undervalued skill

Wondering if chatgpt would make sql irrelevant sometime in the future?

Partly. Not because SQL isn't working technologically, but because the hardest part is figuring out what to ask and how and gpt can figure out the how while having a discussion with you about the what and giving you examples dynamically. If I just need natural language to try things dynamically across tables and I can have a discussion faster than I can google specific stackoverflow issues, I'll pick the bot. It might be confidently wrong but my colleagues do that too, often enough. The difference is that it takes 3 minutes instead of two hours, and no Teams call where you have to appease egos. If you're just really good at SQL and you craft amazing requests or figure out schemas from business needs, you probably have a bit more time to enjoy being the go-to SQL person. Probably less and less as the simpler needs will be increasingly covered, leaving a niche of expert until major advances would happen to AI (unpredictable).

SQL is just a means of expressing what you're trying to do. For some use cases it will be easier to use cgpt but if you are already fluent then it's easier to write SQL rather than put your thoughts into natural language in the form of a detailed prompt.

You’d still need to be able to specify _exactly_ what you want from the database in order for gpt to provide the corresponding sql. You’ll still need to understand the database structure and be able to think in terms of relations and what’s possible via queries.

Make irrelevant? Probably not. Make more convenient? Probably.

I guess the usefulness would be from client software like plsql or any other, implement a natural language processing powered by chatgpt, since the client already has access to the structure then it can include it in the prompt, the user would only write something vague like i want all the clients who bought with more than 500$ this year.

For everyday stuff like that it's easier to put the NLP in a BI tool, and some of them already did that.

If you're doing more involved stuff then writing a good prompt will probably need as much analytical and domain knowledge as writing SQL.

This is a great thread! As you might imagine I put in a lot of thought about whether Chat GPT made learning this stuff obsolete. Conclusion I came to is that those tools make this all more accessible (in terms of helping people to avoid things like frustrating single character mistakes) but that kind of makes knowing some SQL MORE valuable because it makes it much easier to know what direction to move in, and helps a person more easily debug confidently incorrect suggestions.

A big problem with SQL is that sometimes the answers can LOOK right if a person doesn't know the pitfalls of, say, table joins, but can in fact be quite off the rails. I also think that there's still space for highly skilled experts doing the REALLY advanced stuff but for the average user - just knowing the basics probably helps them get even more out of LLM helpers

When crafting more complex SQL queries it's quite easy to make a subtle mistake, mess up cardinality etc; I'm sure chatGPT would be excellent at mirroring those kinds of error unfortunately.

I tried to get ChatGPT do sparql, to query wikidata and I can never get my head around queries in that language. It did not go well at least beyond basics.

Reply on news.ycombinator.com