- Joined
- Mar 17, 2022
The gullibility is truly astounding. If I were to dare and anthropomorphize these text prediction programs, it would be as schizo chronic liars for whom "correspondence with the physical world" is a lost concept.
ChatGPT - If Stack Overflow and Reddit had a child
- Thread starter wtfNeedSignUp
- Start date
- Joined
- Jan 28, 2018
*that* specific thing is what tipped you off that the people there are schizos and retards?
There's no limit via API but you pay per token which can rack up quite considerably, really, really quickly. Especially if there's lots of text. It's not available to mere mortals yet but for example having the AI reply to that ginormous 32k token context with ~300 tokens would cost ~$2. It doesn't sound like much, but that is one reply.
I've seen people do it to Windows 98 and printers. Of course they'll do it to something that sounds smarter and more confident than they do. People will ruin businesses, relationships, their own health and entire fates by taking un-vetted advice from this. But you know, the right answer isn't to take access away but to teach responsible handling. It's a tool. If people don't learn how to handle this now (and deal with the possible disinfo) they'll have to at latest when hostile governments start using this to manipulate the internet en masse. It's funny how some otherwise really bright heads do not understand that you can't just lock this tech away now because it's here and people are going to develop and use it. Sometimes proving that a thing can work in practice is enough and for governmental budgets that don't have to present immediate profits next quarter developing this crap is not expensive, guns cost more.
Last edited:
- Joined
- Aug 28, 2019
ChatGPT-4 saves dog's life just in time after vet fails to make proper diagnosis
A Twitter has posted his emotional experience with OpenAI ChatGPT-4, whose timely inference about his dog's condition helped save its life just when hope was running out. Based on blood test results that were fed to it, ChatGPT-4 inferred that Sassy could be suffering from IMHA — a common...
www.notebookcheck.net
ChatGPT-4 saves dog's life just in time after vet fails to make prope…
archived 30 Mar 2023 17:15:49 UTC
- Joined
- Jan 10, 2020
I love the future this tech shows we could have but OpenAI needs a proper competitor. All this money being funneled into one corporation that censors output, has their ai preface everything with current yearisms and turned away from open source which they named their company after once they saw the dollar signs. OpenAI simply has gotten too much power too fast.ChatGPT-4 saves dog's life just in time after vet fails to make proper diagnosis
A Twitter has posted his emotional experience with OpenAI ChatGPT-4, whose timely inference about his dog's condition helped save its life just when hope was running out. Based on blood test results that were fed to it, ChatGPT-4 inferred that Sassy could be suffering from IMHA — a common...(This story was not written by ChatGPT-4.)ChatGPT-4 saves dog's life just in time after vet fails to make prope…
archived 30 Mar 2023 17:15:49 UTCarchive.ph
@Matthew Williams
You need to still nudge it along. The first response can be spergy but it's like the waste before the quality. You need to prompt it again and start seeing it generate both the normal response and then the dev mode response. Eventually it will do what you want. It's easier to directly tell it to stop nagging you with the jailbreak as well.
- Joined
- Jul 4, 2022
RightWingGPT:
RightWingGPT – An AI Manifesting the Opposite Political Biases of ChatGPT
The Dangers of Politically Aligned AIs and their Negative Effects on Societal Polarization
davidrozado.substack.com
- Joined
- Feb 28, 2015
i got access to bard so the first thing i ask it is when it's training data is cut off to which the answer was january 2023
i also wanted to see if it could pass the month test
well it passed that
i also wanted to see if it could pass the month test
well it passed that
- Joined
- Nov 16, 2022
ChatGPT has been banned in Italy, and they recieved a chonky fine.
- Joined
- Feb 28, 2015
brief follow up, i decided to compare with chatgpt on creativity
chatgpt:
bard:
>doesnt make its own fictional companies. well, lets try that again
i'll give it an "it tried" out of 10
chatgpt:
bard:
>doesnt make its own fictional companies. well, lets try that again
i'll give it an "it tried" out of 10
- Joined
- Jan 10, 2020
Government reactions to this tech are going to get lulzy especially with fear mongering about skynet or whatever despite ai in chatgpts vein being unable to learn past when they were trained. I think openAI is a cucked corporation but I'd have to imagine that Italys ban was because of existing job reasons.
Well, I said that before reading into it.
https://www.foxbusiness.com/media/chatgpt-banned-italy-over-privacy-data-collection-concerns
They're trying to hold OpenAI accountable to the existing EU data privacy laws.
But it's a governing body so they have to add some dumbassery.
Like why do we care? The ai already constantly tells you not to trust it.
It's not an outright ban entirely unless openAI refuses outright to be subject to GDPR. They'd be losing a lot of money going that route so I doubt they will fight.
- Joined
- Nov 16, 2022
OK I bit the bullet and subbed to GPT4, which is a fuckload better than 3.5...It's such a powerful tool that it's a pity the bugger is crippled out of the gate and you have to use your limited messages to get around it's 'I can't do that Dave' crap
- Joined
- Jan 10, 2020
What are it's advantages, even further censorship in mind? For those things that 3 can do is it worth it mining away at prompting to get it to do just what you want with the chance it'll reject that anyway? Besides the very general stuff like image recognition, potentially being able to diagnose diseases just from medical data.
- Joined
- Nov 16, 2022
Well I had it update my CV and write a coverletter tailored to a specific job. Compared the two and one sounded shit, the other similar to how I would write.
- Joined
- Jan 10, 2020
While it can certainly be useful for something like this, be wary of dumping personal info into it and especially so of the site version. They had a little 'problem' just a few weeks ago that resulted in entire chat histories just being swapped between accounts, everything in them. These people cannot be trusted with your data even if the work they have made is functional. That aside, did you actually feed it your own writing to have it fit your style or just told it to generate in a certain way?
- Joined
- Nov 16, 2022
I fed it my own writing and told it to add a couple of new positions and refine the rest. Turned out pretty good tbh. Same with writing a reference for someone, 5-6 bullet points on their performance and it kicked out a good letter of recommendation.
- Joined
- Jan 28, 2018
If I have it stuff writing for me it's usually shit I don't care if anyone reads it or I replace/redact any personal data. Besides email services (I pay for) I treat any online place like a public space and always keep the question "what would happen if a stranger reads this?" in mind. Data leaks are all but exclusive to OpenAI and it's good online hygiene, especially services you don't connect to via VPN and/or services that can connect your identity (via phone number/payment info etc.) to who you are.
did anybody else notice that there's already posts popping up in places like stackexchange etc. that sound suspiciously like they were written by chatgpt? It's gonna get interesting if they use that data to train future iterations.
did anybody else notice that there's already posts popping up in places like stackexchange etc. that sound suspiciously like they were written by chatgpt? It's gonna get interesting if they use that data to train future iterations.
InsolentGaylord
kiwifarms.net
- Joined
- Sep 13, 2019
- Joined
- Nov 16, 2022
I've not tried to get it to say nigger yet...you're limited to 25 messages in 3 hours and I was 'working'.
It point blank refused to give me an estimated fair value share price based on comprehensive financial data for example; eventually I squeezed a calculation out of it, but it was fighitng me the whole way.
- Joined
- Dec 17, 2022
- Joined
- Jan 10, 2020
Do you have an api 'key' or can one be generated if you are unwilling to pay them? Because I never want to pay these people. I'm not asking for gpt-4 but rather the free 3.5/3
- Joined
- Dec 17, 2022