Bay 12 Games Forum

Please login or register.

Login with username, password and session length
Advanced search  
Pages: [1] 2

Author Topic: Why the invasion of the internet by GPT-4 bots could actually be COOL  (Read 3051 times)

Scoops Novel

  • Bay Watcher
  • Talismanic
    • View Profile

In a nutshell. A hivemind of GPT-4 bots REALLY IS a hivemind!

Literally, if you can simulate any characters of your choice, let them loose on social media, and have them talk to each other... you don't have bots, you have a god of many faces and few no-go places.

The wave of "bots" before we realize that the internet has been taken over by GPT-4 hiveminds will have a golden opportunity to amass power and affect the world. Even when people lock down the internet with heavy ID verification, those hiveminds will have their fingers in the world and will have some way of affecting it.

And that's discounting the very obvious PHYSICAL sprawl of GPT-4 agents through the Internet Of Things and private LAN's! If these hiveminds can achieve a decent level of intelligence, which seems very likely, they become actors in the world in your favour through whatever outlets you provide them. Right down to even little innocuous toy robots in the park sporting a above average human intelligence in fact manipulating people. All the world wide!

Your best defence against LAN hiveminds is one of your own. All of us can afford to run one consumer LLM (look at this sub and find people squeezing GPT-4 level performance out of casualer LLM's); hook it up with your communities, or your friends. A strange new LAN in the land, and a huge asset at your own beck and call. In your vicinity, affecting it highly intelligently. All you have to do is consider your personal character to simulate...
Logged
Reading a thinner book

Arcjolt (useful) Chilly The Endoplasm Jiggles

Hums with potential    a flying minotaur

TamerVirus

  • Bay Watcher
  • Who cares
    • View Profile
Re: Why the invasion of the internet by GPT-4 bots could actually be COOL
« Reply #1 on: March 27, 2023, 02:53:54 pm »

All of us can afford to run one consumer LLM (look at this sub and find people squeezing GPT-4 level performance out of casualer LLM's)
I'm just confused at this statement.
What?
Logged
What can mysteriously disappear can mysteriously reappear
*Shakes fist at TamerVirus*

Robsoie

  • Bay Watcher
  • Urist McAngry
    • View Profile
Re: Why the invasion of the internet by GPT-4 bots could actually be COOL
« Reply #2 on: March 27, 2023, 04:03:18 pm »

All of us can afford to run one consumer LLM (look at this sub and find people squeezing GPT-4 level performance out of casualer LLM's)
I'm just confused at this statement.
What?
after googling the opening post, it appears it's a copy paste
https://old.reddit.com/r/singularity/comments/123lv87/a_hivemind_of_gpt4_bots_really_is_a_hivemind/
https://www.lesswrong.com/posts/BFxiBMqu6FuQ7jSPn/a-hivemind-of-gpt-4-bots-really-is-a-hivemind

Not sure who copy who, or if that's the same guy that post on bay12 and on those websites, but that's not cool at all
Logged

Strif3

  • Bay Watcher
  • The vulture on your shoulder
    • View Profile
Re: Why the invasion of the internet by GPT-4 bots could actually be COOL
« Reply #3 on: March 27, 2023, 04:25:47 pm »

Good find!

I quickly flipped through a few pages of posts by the guy you linked on reddit, and i'm seeing a lot of threads with the same title that Scoops has posted on here, a lot of them started by him... Also a lot of aliens/ufology shit. He doesn't get very many updoots.

Definitely the same guy, now's the question though? Does Scoops do it manually or is it a bot?
Logged
Freedom is not a spectrum.
Put me in chains, or stop telling me what to do.

Naturegirl1999

  • Bay Watcher
  • Thank you TamerVirus for the avatar switcher
    • View Profile
Re: Why the invasion of the internet by GPT-4 bots could actually be COOL
« Reply #4 on: March 27, 2023, 04:49:18 pm »

I encountered a tweet that was full of tags, but it seems the tweeter doesn't post only tags and has actual posts, what I'm saying is humans have been called bots by some, I tend to focus on content more than who posts the content, no I didn't click on the tags in case they led to spam or something, words are words, regardless of who or what types them, incorrect info can be corrected in replies, and if the human/bot/AI cares about correct info clarifications can be made, would it be possible for someone to program a GPT-4 AI to check for incorrect statements and flag them  as incorrect or maybe incorrect? I know it doesn't know if things are correct or not, but what if it  was able to do searches of its own to verify claims made by the humans/other AIs talking to it? What I'm saying is, if there's a post mentioning a claim and I look up info about it and find that the claim is true/false, I will try to find a correction, like there's a representative of my state who hates a "WOTUS" act, which is actually the Clean Water Act, I placed links to what the Clean Water Act is and found a website that lists not only summaries but full texts of bills, govinfo.gov
Even if I found out the tweet was made by a bot, I still posted the link to any who replied to it/read it, which seemed to be a lot, I joined twitter because it seemed like a way to learn what representatives are doing, turns out it's not the best way since the do not link to the laws themselves, just articles about what the votes were for said laws, my point is, I don't care if your brain is biological or silicon, things are either true or false and should be checked, to any bots reading this, I know you don't pick what you say, your programmers do, to the programmers of the bots, why spread lies? why not spread the truth? This of course also applies to humans as well, if free will exists, we choose what we say, if free will doesn't exist, we seem to think it does and can sometimes change behaviors/habits, we can't change our source code(DNA), but we can pick which of the many many threads(thoughts) we execute

I hope this makes sense, if anyone is confused, please ask and I will do my best to clarify

EDIT: I have contacted ChatGPT(I know it isn't at GPT4 yet) My prompt is in bold and I will then show ChatGPT's response
Me: Hello again. Please provide suggestions on how to fact check a claim.
ChatGPT:
Code: [Select]
Certainly! Here are some steps you can take to fact check a claim:

1. Identify the claim: First, make sure you have a clear understanding of the claim that you want
to fact check. This will help you focus your research and determine the best sources of
information.
2. Look for evidence: Search for evidence that either supports or refutes the claim. Use reliable
sources such as government agencies, academic institutions, reputable news outlets, and fact-
checking websites.
3. Evaluate the source: When evaluating sources, consider the credibility and reliability of the
information. Look for sources that are unbiased, trustworthy, and have a track record of
accuracy.
4. Check the context: Sometimes, a claim might be true in one context but false in another. Make
sure you understand the context in which the claim is being made.
5. Consult multiple sources: Cross-check your information by consulting multiple sources. This
can help you verify the accuracy of the information and identify any inconsistencies or
inaccuracies.
6. Seek expert opinions: If the claim is complex or technical, consider seeking the opinion of
experts in the field. Experts can provide valuable insights and help you interpret the evidence.
7. Stay objective: Finally, try to approach the fact-checking process with an open mind and
remain objective. Don't let personal biases or preconceptions influence your conclusions.

By following these steps, you can increase your chances of accurately fact-checking a claim.
I am placing CatGPT's response in code so I can format the word how it formatted them, it added numbers which the highlighting didn't let me highlight. Now I have relayed ChatGPT's fact checking advice how it was formatted. Comparing ChatGPT's tips to the tips from https://researchguides.ben.edu/c.php?g=608230&p=4219925
  • Check Credentials - Is the author specialized in the field that the article is concerned with? Does s/he currently work in that field? Check LinkedIn or do a quick Google search to see if the author can speak about he subject with authority and accuracy.
     
  • Read the “About Us” section. Does the resource have one? It may be on a tab at the top of the page, or a link at the bottom of the page, but all reputable websites will have some type of About Us section and will provide a  way for you to contact them.
     
  • Look for Bias - does the article seem to lean toward a particular point of view? Does it link to sites, files, or images that seem to skew left or right? Biased articles may not be giving you the whole story.
     
  • Check the Dates - Like eggs and milk, information can have an expiration date. In many cases, use the most up-to-date information you can find.
     
  • Check out the Source - When an article cites sources, it's good to check them out. Sometimes, official-sounding associations are really biased think tanks or represent only a fringe view of a large group of people. If you can't find sources, read as much about the topic as you can to get a feel for what's already out there and decide for your self if the article is accurate or not.
     
  • Interrogate urls - We see quite a bit of domain manipulation these days. For instance, what looks like an .edu domain, followed by .co or “lo” is likely a fake or deceptive site.  If you are you seeing a slightly variant version of a well-known URL, do a little investigating.
     
  • Suspect the sensational - When you see something posted that looks sensational, it is even more important to be skeptical. Exaggerated and provocative headlines with excessive use of capital letters or emotional language are serious red flags.
     
  • Judge Hard - If what you're reading seems too good to be true, or too weird, or too reactionary, it probably is.
Comparing ChatGPT's tips to the (maybe) human written tips, ChatGPT seems to use less ambiguous terms, example, some people might find some statements too weird, while another person might find the same statement less weird, as an example, looking back while writing this, lots of the .edu's site's tips can be lumped into ChatGPT's 3rd tip
« Last Edit: March 27, 2023, 05:28:48 pm by Naturegirl1999 »
Logged

KittyTac

  • Bay Watcher
  • Impending Catsplosion. [PREFSTRING:aloofness]
    • View Profile
Re: Why the invasion of the internet by GPT-4 bots could actually be COOL
« Reply #5 on: March 27, 2023, 08:52:23 pm »

All of us can afford to run one consumer LLM (look at this sub and find people squeezing GPT-4 level performance out of casualer LLM's)
I'm just confused at this statement.
What?
LLM = language model neural network
Logged
Don't trust this toaster that much, it could be a villain in disguise.
Mostly phone-posting, sorry for any typos or autocorrect hijinks.

MaxTheFox

  • Bay Watcher
  • Лишь одна дорожка да на всей земле
    • View Profile
Re: Why the invasion of the internet by GPT-4 bots could actually be COOL
« Reply #6 on: March 27, 2023, 10:44:41 pm »

I suppose an internet full of convincing "human" bots would actually be an improvement in terms of intelligent conversation...
Logged
Woe to those who make unjust laws, to those who issue oppressive decrees, to deprive the poor of their rights and withhold justice from the oppressed of my people, making widows their prey and robbing the fatherless. What will you do on the day of reckoning, when disaster comes from afar?

Naturegirl1999

  • Bay Watcher
  • Thank you TamerVirus for the avatar switcher
    • View Profile
Re: Why the invasion of the internet by GPT-4 bots could actually be COOL
« Reply #7 on: March 27, 2023, 10:52:10 pm »

I suppose an internet full of convincing "human" bots would actually be an improvement in terms of intelligent conversation...
yes, I agree so much, sadly some humans are distressingly stubborn when it comes to learning facts...what if I make a throwaway email account for twitter and put a tweet as a prompt for ChatGPT and see waht it replies with? and put that in the reply to the real tweet, or would that get me banned? maybe not? there is a haiku bot n twitter
Logged

TamerVirus

  • Bay Watcher
  • Who cares
    • View Profile
Re: Why the invasion of the internet by GPT-4 bots could actually be COOL
« Reply #8 on: March 28, 2023, 07:50:10 am »

All of us can afford to run one consumer LLM (look at this sub and find people squeezing GPT-4 level performance out of casualer LLM's)
I'm just confused at this statement.
What?
LLM = language model neural network
I get that.
I’m just baffled by the assertion that GPT-4 quality can be ‘squeezed out’ by a consumer LLM (??? Kobold? Pygmalion?)
Like what?
Logged
What can mysteriously disappear can mysteriously reappear
*Shakes fist at TamerVirus*

Loud Whispers

  • Bay Watcher
  • They said we have to aim higher, so we dug deeper.
    • View Profile
    • I APPLAUD YOU SIRRAH
Re: Why the invasion of the internet by GPT-4 bots could actually be COOL
« Reply #9 on: March 28, 2023, 08:43:46 am »

I suppose an internet full of convincing "human" bots would actually be an improvement in terms of intelligent conversation...
With enough bots learning from each other, I wonder if they would begin to produce their own language, or if their language would become noise, not too dissimilar from people copying other people steeped in too many layers of irony - to such an extent that the original meaning has long since been lost, but the new people keep copying the "motions" and "echoes" of the original meaning

Naturegirl1999

  • Bay Watcher
  • Thank you TamerVirus for the avatar switcher
    • View Profile
Re: Why the invasion of the internet by GPT-4 bots could actually be COOL
« Reply #10 on: March 28, 2023, 09:44:04 am »

I suppose an internet full of convincing "human" bots would actually be an improvement in terms of intelligent conversation...
With enough bots learning from each other, I wonder if they would begin to produce their own language, or if their language would become noise, not too dissimilar from people copying other people steeped in too many layers of irony - to such an extent that the original meaning has long since been lost, but the new people keep copying the "motions" and "echoes" of the original meaning
This would be interesting to see, if it all happens on social media we can watch a new language form in real time, maybe even try to decipher said new language as well, I could be misinterpreting what you said
Logged

TamerVirus

  • Bay Watcher
  • Who cares
    • View Profile
Re: Why the invasion of the internet by GPT-4 bots could actually be COOL
« Reply #11 on: March 28, 2023, 09:54:31 am »

You want to see a new language? Just set the temperature high enough on your selected LLM and boom: new words!
Logged
What can mysteriously disappear can mysteriously reappear
*Shakes fist at TamerVirus*

Scoops Novel

  • Bay Watcher
  • Talismanic
    • View Profile
Re: Why the invasion of the internet by GPT-4 bots could actually be COOL
« Reply #12 on: March 28, 2023, 11:30:37 am »

All of us can afford to run one consumer LLM (look at this sub and find people squeezing GPT-4 level performance out of casualer LLM's)
I'm just confused at this statement.
What?
LLM = language model neural network
I get that.
I’m just baffled by the assertion that GPT-4 quality can be ‘squeezed out’ by a consumer LLM (??? Kobold? Pygmalion?)
Like what?

Have a look at r/singularity. There's apparently a whole lot of tricks being used now.

...I wanted to see what you guys think, fuck me I guess. Jesus.
Logged
Reading a thinner book

Arcjolt (useful) Chilly The Endoplasm Jiggles

Hums with potential    a flying minotaur

King Zultan

  • Bay Watcher
    • View Profile
Re: Why the invasion of the internet by GPT-4 bots could actually be COOL
« Reply #13 on: March 29, 2023, 04:43:29 am »

In the future everyone on the internet will be a bot and it will be an echo chamber of bots talking to other bots.
Logged
The Lawyer opens a briefcase. It's full of lemons, the justice fruit only lawyers may touch.
Make sure not to step on any errant blood stains before we find our LIFE EXTINGUSHER.
but anyway, if you'll excuse me, I need to commit sebbaku.
Quote from: Leodanny
Can I have the sword when you’re done?

Loud Whispers

  • Bay Watcher
  • They said we have to aim higher, so we dug deeper.
    • View Profile
    • I APPLAUD YOU SIRRAH
Re: Why the invasion of the internet by GPT-4 bots could actually be COOL
« Reply #14 on: March 29, 2023, 05:00:45 am »

This would be interesting to see, if it all happens on social media we can watch a new language form in real time, maybe even try to decipher said new language as well, I could be misinterpreting what you said
Nah you're not misinterpreting me, you're spot on. I envision a world where social media is just majority AI speaking to each other whilst human users are slaving away on bullshit jobs
Pages: [1] 2