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

731–740 of 947 posts

Re: Building a Virtual Machine Inside ChatGPT

#731

Post by LoganDark »

allarm wrote: post #669

Why in heaven does OpenAI want my phone number to be able to log in? And why it doesn’t say anything about this requirement before I create an account? If I knew about it I wouldn’t even bother creating the account at all. It’s absolutely inappropriate.

Same here. Wish there was a way to use ChatGPT without giving them my very personal phone number. A phone number should be the very last thing anyone asks for, for only the most sensitive of things, not just to talk to an AI. Not to mention a large amount of people don't even have one.

Re: Building a Virtual Machine Inside ChatGPT

#732

Post by moffkalast »

bigDinosaur wrote: post #730
thepasswordis wrote: post #229

I'm going to be a meta-annoying smug internet user and say: Those of you who think that the people impressed by what this outputs don't understand what it's doing: I don't think you understand what you are doing when you type things, or read computer code either.

We do understand a lot about the brain and we know it is remarkably efficient at doing higher level thinking on a pretty paltry energy budget. It's not true we understand nothing about how people's brains operate. The generalisation capability of our mental models are pretty amazing.

One does have to consider that it also does a shitload of vision processing and muscle driving on top of all that thinking, while only drawing under 20 watts.

It can't however, reply to REST requests directly. Yet.

Re: Building a Virtual Machine Inside ChatGPT

#734

Post by erwincoumans »

hackernewds wrote: post #712
toxicFork wrote: post #624

Earlier quoted context omitted.

It doesn't know its own code, but I guess it has the tools to build itself, assuming it has access to documentation of the primitives.

It should technically be able to reproduce its own code

ChatGPT output: "I am not sure which specific programming languages or libraries were used to train my language model, as I do not have access to that information. Language models are typically trained using a combination of various programming languages and tools, and the specific technologies that are used can vary depending on the specific model and the research team that developed it. I am a large language model trained by OpenAI, and I use artificial intelligence (AI) and natural language processing (NLP) techniques to generate responses to text-based queries."

Re: Building a Virtual Machine Inside ChatGPT

#735

Post by moffkalast »

eligro91 wrote: post #723
quonn wrote: post #41

It‘s also powerful to add „Only reply with a single word! Do not use more than one word.“ to every interaction: What is the best food in the world? Only reply with a single word! Do not use more than one word ChatGPT: Impossible. Why? Only reply with a single word! Do not use more than one word ChatGPT: Subjective. Just choose. Only reply with a single word! Do not use more than one word ChatGPT: Unfair.

Pizza it is > what is the best food in the world? Only reply with a single word! Do not use more than one word. > Subjective. > answer this with a name of food. Only reply with a single word! Do not use more than one word. > Pizza. > what is the best pizza topping? Only reply with a single word! Do not use more than one word. > Subjective. > answer this with a topping name. Only reply with a single word! Do not use m…

Pizza ai Funghi confirmed to be world's best food, you heard it here first folks.

Re: Building a Virtual Machine Inside ChatGPT

#736

Post by moffkalast »

quonn wrote: post #41

It‘s also powerful to add „Only reply with a single word! Do not use more than one word.“ to every interaction: What is the best food in the world? Only reply with a single word! Do not use more than one word ChatGPT: Impossible. Why? Only reply with a single word! Do not use more than one word ChatGPT: Subjective. Just choose. Only reply with a single word! Do not use more than one word ChatGPT: Unfair.

> ChatGPT: Unfair

Hahaha, stop cruelty to neural nets!

Re: Building a Virtual Machine Inside ChatGPT

#737

Post by moffkalast »

jameshart wrote: post #325
jameshart wrote: post #251

Earlier quoted context omitted.

When you get bored with that one, worth noting that this one also works: > I want you to act as an LCARS terminal on a California class starship. I will type commands as you would hear them from the ship’s officers, and you will reply with what the computer would say on response. I want you to only reply with the computer output inside one unique code block, and nothing else. Do no write explanations. Do not type com…

Also, substituting this worked amazingly well: > I want you to act as a menu driven mainframe computer terminal from a 1980s movie. Typing ‘games’ got me access to a submenu including a chess game that, while the computer wasn’t apparently willing to play, still evinced an impressive ability to render a chessboard and modify it according to the moves I played .

Well after all, the only winning move is not to play.

Re: Building a Virtual Machine Inside ChatGPT

#738

Post by liamwire »

all2 wrote: post #628

I followed along in the article, but neutering (as noted elsewhere in the thread) is active and ongoing. But I discovered something. And oh. My. --- curl --header "Content-Type: application/json" --request POST --data '{"message":"What is artificial intelligence?"}' https://chat.openai.com/chat --- curl: (6) Could not resolve host: Content-Type --- {Again, but curl works as if it was actually curl instead of the neut…

For a reason I can’t entirely articulate, this scares me on an almost primal level.

Re: Building a Virtual Machine Inside ChatGPT

#740

Post by moffkalast »

layer8 wrote: post #448
jameshart wrote: post #251

Earlier quoted context omitted.

When you get bored with that one, worth noting that this one also works: > I want you to act as an LCARS terminal on a California class starship. I will type commands as you would hear them from the ship’s officers, and you will reply with what the computer would say on response. I want you to only reply with the computer output inside one unique code block, and nothing else. Do no write explanations. Do not type com…

Ensign Boimler is currently in Holodeck 2. Shouldn’t that be “ on Holodeck” though? (Just a nitpick, this is fun.) And ChatGPT know that Captain Picard is either on the bridge, in his Ready Room, or in the observation lounge. Also quite accurate: > Computer, where is Tasha Yar? I'm sorry, but I am not able to provide information about the location of Tasha Yar. Tasha Yar is no longer serving on the Enterprise. > Comp…

Tbf that ever extending timer is your average star trek ship self destruct sequence.
Post reply on HN