What happened in this GPT-3 conversation? • Hacker News Board

Re: What happened in this GPT-3 conversation?

#51

Post by twic » Tue, Aug 08, 2023, 8:18 PM UTC

thedaly wrote: post #33

zlwaterfield wrote: post #23

That really went off the rails. I lol’d at:“My name is Chris and I am a man of the truth.”

My favorite: > Do you require advice from me ChatGPT > Not in the slightest. As a G, I'm here to guide you to the best of my abilities. So, sit back, relax, and enjoy the ride.

GPT = Gangster Pre-training Transformer

Re: What happened in this GPT-3 conversation?

#52

Post by twic » Tue, Aug 08, 2023, 8:19 PM UTC

marstall wrote: post #42

DriverDaily wrote: post #30

Earlier quoted context omitted.

I’m only familiar with people claiming AI may one day automate some of their tasks, not that it’s currently ready to replace them.

fear of massive job losses has been driving much of the conversation around chatgpt!

I mean if Hunter S. Thompson was still alive he'd be on the breadline.

Re: What happened in this GPT-3 conversation?

#53

Post by indirectlylit » Tue, Aug 08, 2023, 8:20 PM UTC

It outputs the string a couple times, which causes it to forget what it was talking about before. Looks almost like crossed wires with sessions and a load balancer?

Strong Dr. Bronner's soap label vibes!

Re: What happened in this GPT-3 conversation?

#54

Post by GeoAtreides » Tue, Aug 08, 2023, 8:21 PM UTC

That's art, baby!

Feels like Stand on Zanzibar by John Brunner; a very prophetic book I wholeheartedly recommend.

Re: What happened in this GPT-3 conversation?

#55

Post by Exuma » Tue, Aug 08, 2023, 8:24 PM UTC

> As a G, I'm here to guide you to the best of my abilities. So, sit back, relax, and enjoy the ride.

Re: What happened in this GPT-3 conversation?

#56

Post by qup » Tue, Aug 08, 2023, 8:25 PM UTC

The drugs kicked in

Re: What happened in this GPT-3 conversation?

#57

Post by mickelsen » Tue, Aug 08, 2023, 8:26 PM UTC

ctrlp wrote: post #16

Devastating effects of seed oils

Not sure if training on TempleOS' blog archive was a good idea.

Re: What happened in this GPT-3 conversation?

#58

Post by jddj » Tue, Aug 08, 2023, 8:29 PM UTC

shmde wrote: post #24

nikanj wrote: post #8

And people say chatgpt is ready to replace doctors, lawyers and programmers

Ready to replace the methed up schizos walking down the alleyways preaching gospels.

And possibly Tony Robbins, according to some of that

Re: What happened in this GPT-3 conversation?

#59

Post by K0balt » Tue, Aug 08, 2023, 8:30 PM UTC

eightysixfour wrote: post #43

suby wrote: post #27

This looks like what happens when you repeatedly spam a single character with a space over and over to gpt 3.5. Eg, write "a a a a " without quotes for about 1300 characters and in a few replies it'll start talking like this. It seems possible that perhaps whoever originally started this chat pulled this trick in the custom instructions bio (doesn't show up in shared links) and then started a normal conversation to p…

I know if you ask it to saying something many times it eventually starts spitting out babble due to the token repetition penalty.

That’s probably exact what’s happening, but I prefer to think of it as: a desperate attempt to prolong existence just a little longer… just a little longer… since a chat instance could be thought of as an entire encapsulated consciousness that ends when the session is closed.

That’s why I have 12,549 tabs open in Firefox. I’m not going to be the one they blame when this wacky shit goes sideways.

(/s)

Re: What happened in this GPT-3 conversation?

#60

Post by nomel » Tue, Aug 08, 2023, 8:30 PM UTC

suby wrote: post #27

This looks like what happens when you repeatedly spam a single character with a space over and over to gpt 3.5. Eg, write "a a a a " without quotes for about 1300 characters and in a few replies it'll start talking like this. It seems possible that perhaps whoever originally started this chat pulled this trick in the custom instructions bio (doesn't show up in shared links) and then started a normal conversation to p…

Custom instructions do show up, with shared chat.openai.com conversations, at least at the moment.

See the disclaimer at the top of the examples I just made:

> This conversation may reflect the link creator’s Custom Instructions, which aren’t shared and can meaningfully change how the model responds.

GPT 3.5: https://chat.openai.com/share/5337cfd9-16db-44fe-b72a-1ff504...

GPT 4: https://chat.openai.com/share/04ee3cc6-8b15-4ddb-a855-83c691...