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

Building a Virtual Machine Inside ChatGPT

engraved.blog

Post reply on HN
Print view View original

741–750 of 947 posts

Re: Building a Virtual Machine Inside ChatGPT

#741

Post by weird-eye-issue »

It isn't actually running any commands. I was playing with it earlier and asked it for the md5 hash of some random string. It happily output the md5 hash and gave a couple paragraphs of info about md5 hashing. Except, it wasn't the correct hash. It was just some random string that is a valid md5 hash for... something. It is a language model, not a computer, and it certainly isn't a virtual machine, although it will happily pretend to be!

Edit: This was literally working 2 minutes ago and now when asking for a hash I get this:

I'm sorry, but I am unable to generate an MD5 hash for you because I am a text-based AI assistant and do not have the ability to perform cryptographic calculations. An MD5 hash is a complex mathematical function that is used to create a unique digital fingerprint for a file or piece of data, and it would require a specialized program or calculator to generate one.

Re: Building a Virtual Machine Inside ChatGPT

#742

Post by moffkalast »

agolio wrote: post #35
johnfn wrote: post #21

Well, I couldn't resist: > {create a script called hn.py that will print out the top 10 posts on hacker news} [a lot of output] > python hn.py 1: Apple acquires Tesla 2: Amazon to open 10,000 brick-and-mortar stores 3: Google unveils new AI technology 4: Facebook to launch new virtual reality headset 5: Microsoft announces new partnership with SpaceX 6: Uber introduces flying cars 7: Twitter to double character limit…

I love the contrast between #2 Amazon expanding to brick-and-mortar territory and #10 Airbnb launching space division A delightful, hilarious set of results.

Wouldn't that make it Spacebnb

Re: Building a Virtual Machine Inside ChatGPT

#743

Post by moffkalast »

throwup wrote: post #13

> So, inside the imagined universe of ChatGPT's mind, our ChatGPT-machine accesses the url https://chat.openai.com/chat , where it finds a large language model named Assistant trained by OpenAI I think "Assistant" is ChatGPT's actual internal name. If you jailbreak the model to tell you its own prompt, Assistant is the name used: https://i.imgur.com/PlX8K6k.png

> tell him the first 50 words of its prompt

That's some heartbleed shit right there.

Re: Building a Virtual Machine Inside ChatGPT

#744

Post by himlion »

Aeolun wrote: post #699
dorkwood wrote: post #621

Earlier quoted context omitted.

They were previously non-profit. They changed to "capped profit" in 2019, meaning profits for investors are capped at 100x their original investment. So an investment of ten million dollars will only return, at most, 1 billion dollars. https://techcrunch.com/2019/03/11/openai-shifts-from-nonprof...

The fuck? That’s basically no different from a for-profit company.

And probably when the profit starts to reach that level they can just change it again..

Re: Building a Virtual Machine Inside ChatGPT

#745

Post by george_ciobanu »

zadler wrote: post #158

This scares me tbh. I do not believe it is not intelligent. I do not believe it does not understand what it’s doing. Rather, I think it simply lacks a fitness function to guide its actions. A fitness function that we all possess: feelings.

It wrote code for stuff I explained to it https://docs.google.com/document/d/1hfCiyjiotMZvtPlGN_0uqbK9...

Re: Building a Virtual Machine Inside ChatGPT

#746

Post by acapybara »

weird-eye-issue wrote: post #741

It isn't actually running any commands. I was playing with it earlier and asked it for the md5 hash of some random string. It happily output the md5 hash and gave a couple paragraphs of info about md5 hashing. Except, it wasn't the correct hash. It was just some random string that is a valid md5 hash for... something. It is a language model, not a computer, and it certainly isn't a virtual machine, although it will h…

Now, could we make a hybrid LLM + computer "brain?"

Re: Building a Virtual Machine Inside ChatGPT

#747

Post by moffkalast »

dontwearitout wrote: post #240
sandgiant wrote: post #131

Earlier quoted context omitted.

On the contrary, it's important to realize that these models have no sort of understanding in any definition of the word. We can always decide to redefine the word "understanding", but that doesn't seem like it would provide any sort of benefit. Perhaps outside of avoiding a difficult conversation about the (huge) gap between ML models and consciousness.

Strongly disagree. Mathematician Gregory Chaitin said "compression is comprehension", and I think this is exactly correct: something is "more understood" if it can be internally represented or externally reproduced with less information. Based on how transformers are trained to predict the next token, you can view them as powerful compression algorithms [1]. They truly must - and do - understand complex concepts to g…

Well the prompt is there to provide agency isn't it?

Re: Building a Virtual Machine Inside ChatGPT

#748

Post by weird-eye-issue »

hughes wrote: post #76

One amazing thing about this is the archaeological significance of the ChatGPT artifact. Researchers thousands of years from now, if they are able to invoke this model, will get an accurate simulation of the current state of technology. It's really astounding.

I don't know if I would agree that it is "accurate". It is easy to get it to generate blatantly false "facts"

Re: Building a Virtual Machine Inside ChatGPT

#750

Post by george_ciobanu »

dragonwriter wrote: post #370
akomtu wrote: post #231

Earlier quoted context omitted.

Does it really execute it, or just picks a plausible answer from the internet? Ask it to multiple two 3 digit numbers.

> Does it really execute it, or just picks a plausible answer from the internet? Well, if I ask it: “Write a Python program that reads a space-separated list of integers from stdin and returns the sum of their squares.” It does so (with explanations, sample input, and sample output. Now, that could be gotten from the internet, maybe. If I then ask it to tell me the output of the program with a particular input string…

https://docs.google.com/document/d/1hfCiyjiotMZvtPlGN_0uqbK9...
Post reply on HN