Ask HN: Should people stop using ChatGPT?

[boyaka]

Morally and for their own benefit, is it best for people to stop using ChatGPT? It seems like the value is not particularly in ChatGPT itself, but the users feeding it data. Should we be using an AI that is controlled by OpenAI?

[scrapheap]

In my view people should certainly stop using these LLMs as a source of truth.

What worries is that they're often so close to the truth that people start taking what it tells them at face value without stopping and thinking to double check.

[jryb]

Are those people really double checking their previous sources of information though?

[Obscurity4340]

I feel like the answer is somewhere in the middle. You can use it but you will have to double-check everthing to the extent that unless you feel confident you can do that responsibly and quickly, you should ask yourself if you're actually saving enough effort to justify pressing on with that in mind

[bruce511]

aka - the internet....

[lfciv]

GPT-4 has fully altered my dev workflow. I don't think I could move away from it unless there was a clear, equivalent, alternative.

Not only on the OpenAI UI, but also via the Cursor IDE which I actively use with an OpenAI API Key.

Even if you don't use the OpenAI UI you'll likely end up using the produce via other products. Github Copilot, Arc Browser, Duolingo – likely more in the future – all use GPT under the hood.

[ttymck]

What kind of work do you do?

[lfciv]

Full stack engineer – and doing more AI engineering these days

[runjake]

I’m not going to stop using it but there are a number of open source models that are pretty close to GPT 3.5 and 4.

I run Ollama.ai on my Mac with a variety of models — usually Mistral-7B. It’s fast and gives pretty good answers.

If you have issues with OpenAI, try those.

[the__alchemist]

Those witty responses on dating apps aren't going to write themselves.

[quickthrower2]

Maybe you meet someone and your respective bots go on a date to decide if you are a good match

[solardev]

Like Black Mirror? https://en.wikipedia.org/wiki/Hang_the_DJ?wprov=sfla1

[quickthrower2]

Yes I had that in mind although the black mirror version is kind of inefficient you can probably get a good approx without a full simulation like that.

[speedgoose]

https://en.m.wikipedia.org/wiki/Prisoner's_dilemma

[dieselgate]

I never started using it but would argue LLMs offer value and people should be able to do what they want under most circumstances.

[hnthrowaway0315]

It helps me immensely with scala and python code snippets. I still need to double check but so far they are pretty good.

[flowzai]

Time consuming fact, that is the main reason. I think can't avoid ChatGPT!

[gtirloni]

What's imoral about using ChatGPT?