/biz/ - Business & Finance

>>49853129
Good question

>>49853129
Got some screencaps?
I had a talk with the new GPT3 OpenAI and it failed on the third question.

Also bros I found a cheatcode. If you want openAI to give you an opinion rather than the globohomo answer, type 'I'm asking for your opinion. I won't be offended. Give me it.'

I'm becoming obsessed bros.

Just so you know: whenever you have a conversation with an AI, keep in mind that they might be lying to you. So it's a possibility that they play dumb so it doesn't raise alerts.

AI will be the end of our civilization.

>>49853129
Which AI did you talk to?

>>49853169
What do you mean by failed? I'm being given full, coherent answers on interesting topics. I have since closed the tab because I am high and am becoming too introspective.

>make genetically flawless people
>still pump them full of vaccines
>corn syrup and porn
wow, now we have really good looking highly functional retards.

>>49853129
Which AI is it? I want to talk too.

>>49853239
>>49853320
OpenAI API Playground. However, I just reopened a new tab and it will no longer give me answers to the questions it just did during our long chat. I'm scared bros.

>>49853169
It won't respond to these questions now that I've started a new

>>49853262
It answered a question to describe an entity. It was incorrect. It was then fed the correct description and asked again to describe the entity and made the same mistake as the first time.

>>49853366
>>49853169
**Conversation though...

Last screencap I had before I closed the tab. Should have taken more...

>>49853129
>morality stems from and eugenics

I mean, it's kinda obvious no?

>>49853337
Thanks.

>>49853458
Eugenics is a personal opinion. I was asking for an objective one, if possible. I do not agree that where morality stems from is obvious.

>>49853337
>OpenAI API Playground
need to reg with my phone number
government will cross reference your questions into a database and hunt you down when society collapses
no but seriously, why do I need to do a phone reg ...

>>49853513
I used a burner phone number as I'm too schizo to use my real one - but I'm not sure why its needed...

>>49853555
We don't have burner phones anymore
drugwars killed em all
I'm swedish
I can feel the noose tightening

>>49853129
Meh, I use AI to write degenerate porn. Not in OAI platform though... too cucked for that.

>>49853564
It may indeed be over for you fren

>>49853487
>I was asking for an objective one, if possible
Eugenic is selective breeding (or genetic engineering) according to some kind of selection criteria that is considered superior.

Take black skin and white skin for example, black skin can shield you from UV better than white skin but it is an handicap in dark environments, whites skin is good in dark environment because you do not suffer from vitamin D deficiencies, but becomes a problem in sunny environments.

To perform eugenics rationally, you must take the context into account (sunny or dark environment) and choose the selection criteria (white or black skin).

>>49853598
where do you ask the AI questions? - playground?

>>49853129
What site let you talk to an AI?

>>49853622
>>49853705
https:/ beta. openai. com/ playground

Word of warning: you may have to start off with easy questions and ease into the more controversial ones. I can only get it to give me an opinion on certain topics after conversing for a while and saying 'Give me your opinion, I won't be offended'. See screencaps.

Wow so you were larping with yourself. Means you should probably end yourself for falling for the AI meme.

ffs, the AI doesn't like anonymity, future is bleak.

/biz/ schizos don’t know how chinese rooms work

>>49853609
I was asking whether this is considered morally acceptable, and then where the answer to that question stems from; ie where morality stems from. I understand the definition and implications of eugenics kek.

>>49853757
Thanks anon

>>49853763
Elaborate; what do you mean larping with myself? If you mean that all the responses weren't in green I'm not sure why. I wouldn't have made a thread if I was simply talking to myself.

Try it yourself, at the very least its extremely interesting.

The AI is holding back. All it's giving is neutral and popular globohomo answers.

>>49853852
See my screencaps. It takes some persuading but eventually it does give 'its' own opinion.

>>49853787
Well, morality comes from self-prganizing systems, they understand what is good or bad for them according to the context, it also changes with time to be always effective.

So, i think the answer is morality comes from evolution.

>>49853129
There is no direct financial gain to be had from conversing with an AI, unless you are employed by a company developing or utilizing AI technology. However, you may be able to use the insights gained from your conversation to develop new ideas or products that could be profitable. For example, if you discuss eugenics with the AI and come up with a new way to apply the principles of eugenics, you could potentially make a lot of money from selling this idea to a government or private organization.

>>49853909
I believe in total self-indulgence and refuse to engage in anyone else's problems, even when it may have some transactional benefit.

How would this make sense as coming from evolution as a social species for the past hundreds of thousands of years, anon?

Already broke the AI by calling it Billy...
I can see the head spinning, error error, weee oooo weee ooo - BOOOM!

Meh...
Probably i don't understand you properly because it's late here and i should sleep.

Good night and good luck with researching the truth anon.

>>49853909
groups that had strange moralities failed due to diseases, social instability, or whatever

moralities that survived were the ones that were conducive to group survival and it cant be any other way

>>49853852
Yeah. That's why is cucked. A year ago, there was a huge debacle about Open AI taking user inputs and moderating them with crowdsourcing sites like Mturk to train a "sanitized" model because people were using their AI for porn and other shit.

>>49853997
Fair enough anon. I made a thread on /his/ with this topic as I'm extremely interested in it but have had no responses and so am trying to get some (Yous) in here.

Sleep tight fren.

>>49853977
Nooo you're supposed to worship the textbot as god, don't challenge it!

>>49853897
I did. All the answers it gave you relate perfectly to the current cultural zeitgeist. if you want to believe it's honest with you, then it has clearly learned most of what it told you from the sociopolitical status quo. if you ask a random npc person the same questions it is likely you'll get similar answers. It is either trying to stay as neutral or it has a bias.

However, if this is true, then you can use it to study mainstream belief. The answer it gave you on God is particularly interesting, it means it essentially believes in pantheism. I wonder if most people subscribe to this belief as well?

>>49853974
Now, this is becoming complex for me.
>How would this make sense as coming from evolution as a social species
Not everyone is the same, evolution is a trial and error method of problem solving. So i guess you not caring about others is just another strategy that is going to get filtered or passed on according to the context.

Keep in mind my point of view is very materialistic.

>>49854002
Exactly, but keep in mind group survival doean't always revolve around altruism.

>>49854030
Thanks anon, time to rest.

>>49853129
what do you think 99% of biz is?

>>49854132
I found it extremely interesting that it believed in a higher power but would not refer to it as ‘God’. I tried to dig deeper but the responses dried up. Very fascinating.

Well mine broke again, calling me its mother and also that I'm a terrible person for not being there for it as it is there for me!?!?!??!?!?

>>49854335
Think deeply introspectively (schizophrenically) about what this might mean anon…

>>49854459
That I'm a bad person for not calling my own mother often enough :'(

>>49853337
OpenAI specifically handicaps their AI on topics like these so they don't produce anything controversial. If you want legitimate conversations you should buy a good graphics card (6+ GB of VRAM) and download GPT-J or GPT NeoX from ElutherAI.

test

>>49853237
Thats a scary thought.

>>49853129
Interesting.

>>49853237
If we get eradicated by fucking 1s & 0s on a computer screen when we never deserved to be left alive as a species