Hacker News Reader

Comment by bmitc | Hacker News Reader

I had this disagreement with people on this site just the other day. People basically were like "you're asking it too complicated questions", but my response was then why does everyone make statements like the commenter you replied to?

View on HN

Because 99.9% of people who say things like this are just using ChatGPT itself and not any of the various awe-inspiring tools with full access to your codebase dynamically inserted into context via RAG. I have yet to talk to anyone who has actually worked for any amount of time against the GPT4 API or through Cursor, say, who underestimates their capabilities. Sincerely hoping this 'coup' doesn't mean the beginning of the end of that experience for most...

> who underestimates their capabilities

Did you happen to mean overestimates? Just trying to make sure I understand.

I just mean people who have actually used the API directly or through task-specific applications like Cursor that are meant to maximize use of AI for their needs know how much of a breakthrough we’ve had this year. People who doubt or downplay the already existing capabilities of this technology tend to have just played with ChatGPT a little bit (or have whatever ideological or other reason to do so).

Context is very important in these kinds of use cases. If you work with something niche, I think these tools are less valuable because the training data becomes sparse.

For example, GPT-4 produces Javascript code far better than it produces Clojure code. Often, when it comes to Clojure, GPT-4 produces broken examples, contradictory explanations, or even circular reasoning.

Have you tried Cursor out of curiosity? No ties to the company and long-time dev (Scala mostly), just genuinely found it to be transformative to my development practice like no tool before.

Reply on news.ycombinator.com