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

Using ChatGPT to generate a GPT project end-to-end

github.com

Post reply on HN
Print view View original

101–110 of 225 posts

Re: Using ChatGPT to generate a GPT project end-to-end

#101

Post by shmoogy »

davidktr wrote: post #50
jvanderbot wrote: post #18

Earlier quoted context omitted.

Copilot without -x does this relatively well. But it's a hands-on process. I can't just give it some source files and say "go". But it can easily make you 10x. I often spend more time tab-completing than writing.

I'm still trying to figure out how to do this. Copilot is extremely helpful. But more often than not, it suggests completions which would steer the project in another direction. It is extremely good once the code has been organised and an overall structure exists. But until then it can be a distraction.

It helps me to create thoughtful docstrings and comments which directs the copilot to much stronger completions

Re: Using ChatGPT to generate a GPT project end-to-end

#102

Post by avidphantasm »

choeger wrote: post #65
c7b wrote: post #56

The technological singularity is approaching fast. It's pretty clear that, given terminal access and an appropriate outer loop, GPT models can iteratively create new GPT models (either by writing and executing Pyhton code, or later versions trained on LLM weights may even be able to output new weights directly). If the inner workings of the loop are sufficiently obfuscated (in the code-based version), it wouldn't nec…

Calm down. It is a language model . People have figured out how to predict the next word for a given prefix. That's very cool and it will definitely have a significant impact on software. But it is not artificial intelligence.

So you are 100% certain that no emergent properties can develop from a LLM that transcends the limitations of LLMs. I haven’t read any LLM literature, so I am honestly asking, do you know of anything close to a proof that such emergent properties cannot develop?

Re: Using ChatGPT to generate a GPT project end-to-end

#103

Post by shmoogy »

jiggywiggy wrote: post #34
amelius wrote: post #24

Is anyone letting an LLM code and run its code by itself, then iteratively fix any bugs in it without human intervention until it e.g. passes some black box tests? Would it be possible to significantly improve an LLM using such unsupervised sessions?

Yeah at some point gpt4 loses track and just consistently is wrong. Lately I can't feed it too much info, the longer the context the more issues. With your suggestion, it doesn't know which part of the iteration is correct at the moment. For us iteration is logically but for chatgpt I think it's just more variables that make the chance of being wrong larger. So you need to build that in somehow that it can iterativel…

Isn't the issue that things fall out of context and then it starts hallucinating a lot more ? Sometimes it helps to just start a new prompt

Re: Using ChatGPT to generate a GPT project end-to-end

#104

Post by KaoruAoiShiho »

ZephyrBlu wrote: post #100
KaoruAoiShiho wrote: post #97

Earlier quoted context omitted.

Yes AutoGPT would be the singularity if it had more brain-like subsystems like long and short term memory and is hooked up to real world instruments.

"Yes AutoGPT would be the singularity if it was way more advanced and could manipulate physical reality at will" Way to handwave all the difficulties with this in reality.

Lmao neither of those things I mentioned really cost more than 50 million dollars.

Re: Using ChatGPT to generate a GPT project end-to-end

#105

Post by jiggywiggy »

shmoogy wrote: post #103
jiggywiggy wrote: post #34

Earlier quoted context omitted.

Yeah at some point gpt4 loses track and just consistently is wrong. Lately I can't feed it too much info, the longer the context the more issues. With your suggestion, it doesn't know which part of the iteration is correct at the moment. For us iteration is logically but for chatgpt I think it's just more variables that make the chance of being wrong larger. So you need to build that in somehow that it can iterativel…

Isn't the issue that things fall out of context and then it starts hallucinating a lot more ? Sometimes it helps to just start a new prompt

I don't know what the issue is. Been happening more lately. Before would ask a lot and would manage, now often 4-5 prompts in seems to just answer without previous context

Re: Using ChatGPT to generate a GPT project end-to-end

#106

Post by ZephyrBlu »

KaoruAoiShiho wrote: post #104
ZephyrBlu wrote: post #100

Earlier quoted context omitted.

"Yes AutoGPT would be the singularity if it was way more advanced and could manipulate physical reality at will" Way to handwave all the difficulties with this in reality.

Lmao neither of those things I mentioned really cost more than 50 million dollars.

Money is definitely not the bottleneck here.

Re: Using ChatGPT to generate a GPT project end-to-end

#107

Post by KaoruAoiShiho »

ZephyrBlu wrote: post #106
KaoruAoiShiho wrote: post #104

Earlier quoted context omitted.

Lmao neither of those things I mentioned really cost more than 50 million dollars.

Money is definitely not the bottleneck here.

Yes it is, if not I would've already done it.

Re: Using ChatGPT to generate a GPT project end-to-end

#108

Post by gwoolhurme »

KaoruAoiShiho wrote: post #104
ZephyrBlu wrote: post #100

Earlier quoted context omitted.

"Yes AutoGPT would be the singularity if it was way more advanced and could manipulate physical reality at will" Way to handwave all the difficulties with this in reality.

Lmao neither of those things I mentioned really cost more than 50 million dollars.

I personally think the same as the other comment. Do you really think money is the bottleneck? Let’s assume Sam Altman felt that way. He couldn’t drum up that cash tomorrow if he wanted??? For the pinnacle of human invention. Twitter was purchased for way way more than 50 mil.

Re: Using ChatGPT to generate a GPT project end-to-end

#109

Post by srcreigh »

hnlmorg wrote: post #66
ineedasername wrote: post #54

This is not really the same, but may be interesting to some: I subscribe to ChatGPT plus for a month to check out GPT-4. The rate limits were cumbersome though and it can be easy to waste a prompt, so I started to bootstrap: I would explain my problem to 3.5 and ask it to suggest comprehensive prompts to use with 4 to maximize my limited quota. It worked very well. In the long years to come the most advance AIs may b…

I see this as an example of the reverse: AI is still stupid enough that it takes humans a degree of skill to craft a request which generates the desired output.

Kolmogorov complexity would like a word! It seems intractable for AI to read minds, there should always be some degree of skill involved in prompt writing.

Re: Using ChatGPT to generate a GPT project end-to-end

#110

Post by unshavedyak »

hnlmorg wrote: post #66
ineedasername wrote: post #54

This is not really the same, but may be interesting to some: I subscribe to ChatGPT plus for a month to check out GPT-4. The rate limits were cumbersome though and it can be easy to waste a prompt, so I started to bootstrap: I would explain my problem to 3.5 and ask it to suggest comprehensive prompts to use with 4 to maximize my limited quota. It worked very well. In the long years to come the most advance AIs may b…

I see this as an example of the reverse: AI is still stupid enough that it takes humans a degree of skill to craft a request which generates the desired output.

I don't know how true it is vs how much PR it is, but Khan Academy's use of LLMs was interesting in that they apparently craft a prompt from the AI itself. Ie a two step process where the AI generates the steps, and then the AI reasons about the result of the steps to attempt and judge the accuracy of the data. This was just a blurb from the Ted talk[1], but i'd be interested in seeing a slightly more in depth explanation about this strategy.

Made me wonder if you could have an recursive cycle of thinking. Where the "AI" prompts itself, reasons on the output, and does that repeatedly with guards such that it will stop if it doesn't judge anymore advancement.

[1]: https://www.youtube.com/watch?v=hJP5GqnTrNo&t=715s

Post reply on HN