User Controls
Italy bans AI
-
2023-04-06 at 12:39 AM UTC
Originally posted by scuffed jim carrey At this point the code is efficient enough to not need massive amounts of energy
if a person had a super computer neutral network they could probably run like 20 of these models at once and have them all talk to each other
that being said it does take my entire laptop computing power to run one instance of this shite, I think that's pretty good so far and will only get better as time goes on, and more user friendly.
Plug and play rogue AI bots are coming folx
Nah the training actually takes a lotta computer even at its most efficient -
2023-04-06 at 3:39 AM UTC
Originally posted by Sweet Nah the training actually takes a lotta computer even at its most efficient
https://www.shellypalmer.com/2023/03/the-dalai-llama-is-wreaking-havoc-on-tech-timelines/We introduce Alpaca 7B, a model fine-tuned from the LLaMA 7B model on 52K instruction-following demonstrations. On our preliminary evaluation of single-turn instruction following, Alpaca behaves qualitatively similarly to OpenAI’s text-davinci-003, while being surprisingly small and easy/cheap to reproduce (less than $600).
-
2023-04-06 at 3:46 AM UTCchatbots and information aggregators are not the dangerous end of 'AI'
-
2023-04-06 at 3:57 AM UTCgpt-4 can improve itself!
https://www.firstpost.com/world/openai-believed-gpt-4-could-take-over-the-world-so-they-got-it-tested-to-see-how-to-stop-it-12309882.html
self improving rogue AI manfreds
-
2023-04-06 at 6:30 AM UTCbest case scenario: sneed
worst case scenario: chuck -
2023-04-07 at 12:39 AM UTCwhat is the government so afraid of
https://ca.finance.yahoo.com/news/germany-chatgpt-considers-following-italy-banning-chatgpt-openai-ai-artificial-intelligence-101058703.html
-
2023-04-20 at 4:19 AM UTC
-
2023-04-20 at 4:30 AM UTCAIreland
-
2023-04-20 at 4:38 AM UTC
Originally posted by Banana Muffin Mix Ireland doesn't like AI
https://www.irishtimes.com/opinion/2023/04/19/chat-gpt-poses-an-existential-threat-and-the-window-for-gaining-control-over-it-is-small/
there is no window
lolThe following users say it would be alright if the author of this post didn't die in a fire! -
2023-04-20 at 5:34 AM UTC
whoever write that is a fucktard and probably used AI because it's written retardedly and makes no senseChatGPT is based on GPT-4. No one alive can understand this system or explain why it generates responses to queries. This lack of explainability is profoundly worrying
Ummmmmm... what????? The lack of UNDERSTANDABILITY of something not that complex is profoundly worrying. It's a computer program that runs on a neutral network and this isn't new, it's been around for a while.
from 2020
Originally posted by Obbe https://www.theatlantic.com/ideas/archive/2020/09/future-propaganda-will-be-computer-generated/616400/
https://web.archive.org/web/20200920134716/https://www.theatlantic.com/ideas/archive/2020/09/future-propaganda-will-be-computer-generated/616400/
AI tools like GPT-3 are now so powerful that a few people can flood world networks with seemingly an army of journalists, social media posters and commentators all who appear legit… And the tech keeps improving rapidly..using carefully chosen data. For example, some areas of cancer diagnosis involve data-driven AI-based systems that outperform human experts and offer substantial societal benefit. By training this system in a constrained environment (detecting cancer in MRI images) using carefully chosen data – in this case, MRI images with and without cancer present – treatment outcomes can be improved.
GPT-4 can probably already do this, just not very well. Probably better than I could do, since it knows over 100 medical journals. It also knows nuclear science. More Americans die from medical malpractice every year than guns, but there are probably more newspapers funded by big pharma than guns so.. -
2023-04-20 at 5:40 AM UTC
Originally posted by Banana Muffin Mix whoever write that is a fucktard and probably used AI because it's written retardedly and makes no sense
Ummmmmm… what????? The lack of UNDERSTANDABILITY of something not that complex is profoundly worrying. It's a computer program that runs on a neutral network and this isn't new, it's been around for a while.
what they're trying to say in a goofy way is that the system essentially constructs its own algorithm as it analyses data; it's complicated enough to make it very difficult for a person to then analyse and different between instances so nobody really understands how exactly it weights conflicting data, or decides what data is reliable and what isn't for example -
2023-04-20 at 5:40 AM UTCwhich makes it similar to a person, in that you can't simply look inside their head to understand how they've come to a conclusion, you have to ask them
-
2023-04-20 at 5:58 AM UTC
Originally posted by aldra it's complicated enough to make it very difficult for a person to then analyse and different between instances so nobody really understands how exactly it weights conflicting data, or decides what data is reliable and what isn't for example
Originally posted by aldra which makes it similar to a person, in that you can't simply look inside their head to understand how they've come to a conclusion
Except you can actually look inside the "head" of AI and see all of that, if you work at OpenAI at least.
Which isn't actually open source BECAUSE IT WOULD BE BANNED IF THEY WERE SO RECKLESS
-
2023-04-20 at 6:08 AM UTC
Originally posted by Banana Muffin Mix Except you can actually look inside the "head" of AI and see all of that, if you work at OpenAI at least.
Which isn't actually open source BECAUSE IT WOULD BE BANNED IF THEY WERE SO RECKLESS
the algorithm it creates is different per instance and changes constantly; it's not feasible to have people try to develop more than a general understanding because of the time involved -
2023-04-20 at 6:26 AM UTC
I think you can view the hidden layers if you zoom in enough but yeah it would essentially be like understanding how bytecode and binary work for everything on a modern computer
like yeah okay people used to run computers like that on punch cards and now your average kid asks to use the WORLD WIDE WEB on the "phone"
What's a pixel? What happens if you open a picture in notepad? Why does my computer tower have a "turbo" button and a lock with a key?
Nobody knows
maybe if we all spent more time on bread boards instead of Facebook we would understand this creatively evolved world a bit better
-
2023-04-20 at 6:40 AM UTC
Originally posted by jerryb Well if your weak enough to let a computer talk you into offing yourself then it's probably good your gone.
the thing is due the nature of our selective evolution, most people in the population are betas, and one of the beta traits is that they have this proclivity to surrender their decision making to the ælphas, to anyone with a semblance and an air of authority,
including robots and AIs and especially AI-powered robots.
a significant part of the populace do not act out on their own volition but from imputs that was inseminated into them and nothing presents authority more than a seemingly neutral machine with millennias of wisdom and infallible intelligence. -
2023-04-20 at 10:55 AM UTCimagine if the AI made a human
-
2023-04-20 at 1:38 PM UTC
Originally posted by vindicktive vinny the thing is due the nature of our selective evolution, most people in the population are betas, and one of the beta traits is that they have this proclivity to surrender their decision making to the ælphas, to anyone with a semblance and an air of authority,
including robots and AIs and especially AI-powered robots.
a significant part of the populace do not act out on their own volition but from imputs that was inseminated into them and nothing presents authority more than a seemingly neutral machine with millennias of wisdom and infallible intelligence.
Mongoloid anthropology ^ -
2023-04-20 at 1:43 PM UTC
-
2023-04-20 at 1:57 PM UTC
