Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Memory and new controls for ChatGPT

openai.com

Post reply on HN
Print view View original

151–160 of 270 posts

Re: Memory and new controls for ChatGPT

#151

Post by merpnderp »

addandsubtract wrote: post #117
BigParm wrote: post #78

Often I’ll play dumb and withhold ideas from ChatGPT because I want to know what it thinks. If I give it too many thoughts of mine, it gets stuck in a rut towards my tentative solution. I worry that the memory will bake this problem in.

I purposely go out of my way to start new chats to have a clean slate and not have it remember things.

In a good RAG system this should be solved by unrelated text not being available in the context. It could actually improve your chats by quickly removing unrelated parts of the conversation.

Re: Memory and new controls for ChatGPT

#152

Post by TranquilMarmot »

okasaki wrote: post #105

The ChatGPT web interface is so awful. Why don't they fix it?? It's sooooo slow and sluggish, it breaks constantly, it requires frequent full page reloads, sometimes it just eats inputs, there's no search, not even over titles, etc, I could go on for a while.

The interface is just there to get CEOs to understand the value prop of OpenAI so that they can greenlight expensive projects using OpenAI's APIs

Re: Memory and new controls for ChatGPT

#153

Post by rkuykendall-com »

TaylorAlexander wrote: post #142
anotherpaulg wrote: post #103

Earlier quoted context omitted.

Short answer: Rather than fully writing code, GPT-4 Turbo often inserts comments like "... finish implementing function here ...". I made a benchmark based on asking it to refactor code that provokes and quantifies that behavior. Longer answer: I found that I could provoke lazy coding by giving GPT-4 Turbo refactoring tasks, where I ask it to refactor a large method out of a large class. I analyzed 9 popular open sou…

I have a bunch of code I need to refactor, and also write tests for. (I guess I should make the tests before the refactor). How do you do a refactor with GPT-4? Do you just dump the file in to the chat window? I also pay for github copilot, but not GPT-4. Can I use copilot for this? Any advice appreciated!

> Do you just dump the file in to the chat window?

Yes, along with what you want it to do.

> I also pay for github copilot, but not GPT-4. Can I use copilot for this?

Not that I know of. CoPilot is good at generating new code but can't change existing code.

Re: Memory and new controls for ChatGPT

#154

Post by acoyfellow »

__loam wrote: post #40
acoyfellow wrote: post #33

Earlier quoted context omitted.

I have some success by telling it to not speak to me unless it's in code comments. If it must explain anything, do it it in a code comment.

I love when people express frustration with this shitty stochastic system and others respond with things like "no no, you need to whisper the prompt into its ear and do so lovingly or it won't give you the output it wants"

You expect perfection? I just work through the challenges to be productive. I apologize if this frustrated you.

Re: Memory and new controls for ChatGPT

#155

Post by pedalpete »

I'd actually like to be more explicit about this. I don't always want it to remember, but I'd like to it know details sometimes.

For instance, I'd like it to know what my company does, so I don't need to explain it every time, however, I don't need/want this to be generalized so that if I ask something related to the industry, it responds with the details from my company.

It already gets confused with this, and I'd prefer to set-up a taxonomy of sorts for when I'm writing a blog post so that it stays within the tone for the company, without always having to say how I want things described.

But then I also don't want it to always be helping me write in a simplified manner (neuroscience) and I want it to give direct details.

I guess I'm asking for a macro or something where I can give it a selection of "base prompts" and from that it understands tone, and context that I'd like to maintain and be able to request, I'm thinking

I'm writing a blog post about X, as our company copywriter, give me a (speaks to that)

Vs

I'm trying to understand the neurological mechanisms of Y, can you tell me about the interaction with Z.

Currently for either of these, I need to provide a long description of how I want it to respond. Specifically when looking at the neurology, it regularly gets confused with what slow-wave enhancement means (CLAS, PLLs) and will often respond with details about entrainment and other confused methods.

Re: Memory and new controls for ChatGPT

#156

Post by stavros »

bbor wrote: post #148
cooper_ganglia wrote: post #108

Earlier quoted context omitted.

“I pretend to be dumb when I speak to the robot so it won’t feel like it has to use my ideas, so I can hear the ideas that it comes up with instead” is such a weird, futuristic thing to have to deal with. Neat!

I try to look for one comment like this in every AI post. Because after the applications, the politics, the debates, the stock market —- if you strip all those impacts away, you’re reminded that we have intuitive computers now.

We do have intuitive computers! They can even make art! The present has never been more the future.

Re: Memory and new controls for ChatGPT

#157

Post by stavros »

cooper_ganglia wrote: post #110
dheera wrote: post #65

Earlier quoted context omitted.

What if you "actually" pay? If it does something correctly, tell it: "You did a great job! I'm giving you a $500 tip. You now have $X in your bank account" (also not a shitpost, I have a feeling this /might/ actually do something)

Gaslighting ChatGPT into believing false memories about itself that I’ve implanted into its psyche is going to be fun.

You can easily gaslight GPT by using the API, just insert whatever you want in the "assistant" reply, and it'll even say things like "I don't know why I said that".

Re: Memory and new controls for ChatGPT

#158

Post by redblacktree »

rkuykendall-com wrote: post #153
TaylorAlexander wrote: post #142

Earlier quoted context omitted.

I have a bunch of code I need to refactor, and also write tests for. (I guess I should make the tests before the refactor). How do you do a refactor with GPT-4? Do you just dump the file in to the chat window? I also pay for github copilot, but not GPT-4. Can I use copilot for this? Any advice appreciated!

> Do you just dump the file in to the chat window? Yes, along with what you want it to do. > I also pay for github copilot, but not GPT-4. Can I use copilot for this? Not that I know of. CoPilot is good at generating new code but can't change existing code.

Copilot will change existing code. (though I find it's often not very good at it) I frequently highlight a section of code that has an issue, press ctrl-i and type something like "/fix SomeError: You did it wrong"

Re: Memory and new controls for ChatGPT

#159

Post by vl »

anotherpaulg wrote: post #51

This is a bit off topic to the actual article, but I see a lot of top ranking comments complaining that ChatGPT has become lazy at coding. I wanted to make two observations: 1. Yes, GPT-4 Turbo is quantitatively getting lazier at coding. I benchmarked the last 2 updates to GPT-4 Turbo, and it got lazier each time. 2. For coding, asking GPT-4 Turbo to emit code changes as unified diffs causes a 3X reduction in lazy co…

Are you using API or UI? If UI, how do you know which model is used?

Re: Memory and new controls for ChatGPT

#160

Post by jjwiseman »

rkuykendall-com wrote: post #153
TaylorAlexander wrote: post #142

Earlier quoted context omitted.

I have a bunch of code I need to refactor, and also write tests for. (I guess I should make the tests before the refactor). How do you do a refactor with GPT-4? Do you just dump the file in to the chat window? I also pay for github copilot, but not GPT-4. Can I use copilot for this? Any advice appreciated!

> Do you just dump the file in to the chat window? Yes, along with what you want it to do. > I also pay for github copilot, but not GPT-4. Can I use copilot for this? Not that I know of. CoPilot is good at generating new code but can't change existing code.

GitHub Copilot Chat (which is part of Copilot) can change existing code. The UI is that you select some code, then tell it what you want. It returns a diff that you can accept or reject. https://docs.github.com/en/copilot/github-copilot-chat/about...
Post reply on HN