AI - The artificial intelligence thread

[Satans Little Helper]

40 minutes ago, o00o said:

yeah I had the same experience. I  always restart with a better promt when asking coding questions and almost never start a conversation about it. the worst results are when you tell it to correct itself instead of starting a new session 

caude.ai also got an update recently which made it perform better than GPT. I gave it another spin today and the results are really great so far: https://claude.ai/chat/

Yes! As soon as you need to correct chatGPT it's over. And also, as soon as it starts to write a solution, but in the end starts to correct itself and completely starts over. Another red flag. That's when chatGPT starts to get weird. Often, in hindsight it's because the question was too broad and/or complex. Sometimes it can take me a day to get to the right set of questions.

 

37 minutes ago, zkom said:

Well, yes, this is kind of what I meant that when you have well defined problems like you would get in a coding challenge or CS class it gives decent answers. But the real world problems when you're developing software and systems are hardly ever so well defined and everything is kind of murky and undocumented and ill-defined, possibly with inaccurate documentation that no one bothered to update. The kind of algorithmic part is usually fairly straightforward, it's all the technical details and buggy hardware and software that are the pain in the ass.

Really can't comment on your specific situation. Feeding it error messages might help. But I'm sure you've already tried that. Not telling you anything new, I'm afraid. It's just that my current intuition/experience about ChatGPT giving helpful code is about optimising the set of questions.

And yes, you do need to define a problem. That's the art of it. Within a google challenge thats different to real life issues, obvsly. You need to "tell" ChatGPT what it needs to do. Not necessarily how, btw. It can produce the required algorithm all by itself.

In the end, it's the proverbial "stochastic parrot" which returns the most probable sequence of words/code to an input. The input defines the output.

[Nebraska]

 

 

[iococoi]

Quote

Today we're excited to introduce Devin, the first AI software engineer.

[Nebraska]

Digital Marilyn interacts with fans in real-time using advanced natural language processing, deep learning and GPT 3.5. This hyper-real Digital Celebrity answers questions in Marilyn's signature voice and style, shares anecdotes, and even delivers personalized greetings, creating an unforgettable experience for Marilyn fans.

"This collaboration exemplifies the transformative power of AI in connecting brands and consumers. Digital Marilyn showcases our Biological AI, bringing an iconic personality to life through engaging dialogues and emotional intelligence. It's more than nostalgia; it's a glimpse into the future of immersive interactions." Greg Cross, CEO and Co-founder of Soul Machines.

https://www.prnewswire.com/news-releases/soul-machines-to-unveil-groundbreaking-digital-marilyn-monroe-at-sxsw-2024-302083533.html

[iococoi]

[o00o]

On 3/14/2024 at 1:41 AM, Nebraska said:

Digital Marilyn interacts with fans in real-time using advanced natural language processing, deep learning and GPT 3.5. This hyper-real Digital Celebrity answers questions in Marilyn's signature voice and style, shares anecdotes, and even delivers personalized greetings, creating an unforgettable experience for Marilyn fans.

"This collaboration exemplifies the transformative power of AI in connecting brands and consumers. Digital Marilyn showcases our Biological AI, bringing an iconic personality to life through engaging dialogues and emotional intelligence. It's more than nostalgia; it's a glimpse into the future of immersive interactions." Greg Cross, CEO and Co-founder of Soul Machines.

https://www.prnewswire.com/news-releases/soul-machines-to-unveil-groundbreaking-digital-marilyn-monroe-at-sxsw-2024-302083533.html

GPT 3.5 is pretty stupid in comparison to the other models 

[chronical]

This was on my Facebook feed today. It was ringing all my AI picture alarm bells but the hands are correct, the shadows align for the most part. The diner logo photoshopped on is the only thing that stood out at first, but I'm preeetty sure the whole thing is just one ChatGPT/Dall*E process. I'm honestly so done with social media at the moment, the amount of generated content is getting extremely uncomfortable to see when you realize just how much of it is there. I'm so glad I was born before this era of phones being glued to our faces. Fuck this shit

[o00o]

10 minutes ago, chronical said:

This was on my Facebook feed today. It was ringing all my AI picture alarm bells but the hands are correct, the shadows align for the most part. The diner logo photoshopped on is the only thing that stood out at first, but I'm preeetty sure the whole thing is just one ChatGPT/Dall*E process. I'm honestly so done with social media at the moment, the amount of generated content is getting extremely uncomfortable to see when you realize just how much of it is there. I'm so glad I was born before this era of phones being glued to our faces. Fuck this shit

don't use facebook 

[chronical]

22 minutes ago, o00o said:

don't use facebook 

Well yeah. Valid point. My profile would be long gone if it wasn't the only platform my grandparents know how to use. But yeah. It'll be gone soon

[zero]

2 hours ago, chronical said:

I'm so glad I was born before this era of phones being glued to our faces. Fuck this shit

+1

we're the last of the Mohicans here bro...a dying breed.

[ignatius]

2 hours ago, chronical said:

This was on my Facebook feed today. It was ringing all my AI picture alarm bells but the hands are correct, the shadows align for the most part. The diner logo photoshopped on is the only thing that stood out at first, but I'm preeetty sure the whole thing is just one ChatGPT/Dall*E process. I'm honestly so done with social media at the moment, the amount of generated content is getting extremely uncomfortable to see when you realize just how much of it is there. I'm so glad I was born before this era of phones being glued to our faces. Fuck this shit

it's "Coors Light" though  and wtf the hookha. does anyone hold a burger like that? lols.. so good. so bad. 

[o00o]

9 minutes ago, ignatius said:

does anyone hold a burger like that?

yeah it looks like he is holding it mostly at the toppings 

[logakght]

3 hours ago, o00o said:

don't use facebook 

Agree with this. Using Facebook is not IDM at all—in fact most social media is absolute trash, except are.na.

[luke viia]

there are two hookah hoses (and a very large hookah itself situated impossibly close to a fat man's crotch) and two Coors "LITE" cans in that image, suggesting the restaurant is giving both alcohol and shisha to young children as well as burgers that appear to be a half pound of beef and yeah no one in reality has ever grasped a burger like it's a bird that needs its wings held down to feel safe. how is this a convincing AI image lol

[iococoi]

Quote

Mushrooming Risk: Unreliable A.I. Tools Generate Mushroom Misinformation

https://www.citizen.org/article/mushroom-risk-ai-app-misinformation/

 

[logakght]

Filthy over. Honestly this hole AI shit is making me a lot more doomer than normal: https://www.rollingstone.com/music/music-features/suno-ai-chatgpt-for-music-1234982307/

[Freak of the week]

who cares, at this point music is just regurgitation of regurgitation ... (n times), where n->infinity

more ai-generated junk certainly won't change that

[ignatius]

climate change is the final boss of everything anyways... 

edit to add this fckn 'song'

 

Edited March 20 by ignatius

[logakght]

My friends, I do believe this time it is completely over: https://soundry.ai/

[ignatius]

1 hour ago, logakght said:

My friends, I do believe this time it is completely over: https://soundry.ai/

there's going to be a lot of that kind of thing and people who just lean real hard on it and generate tracks alllllll daaayyy long.

[prdctvsm]

 

[ignatius]

41 minutes ago, prdctvsm said:

 

pretty interesting/informative video. seems to have a handle on 'cutting edge' things happening. hmm. 

edit: so, still watching. long video. and fuuuuuck. this is gonna be pretty crazy. 

Edited March 23 by ignatius

[prdctvsm]

 

[o00o]

this shit looks strait out of black mirror 

 

 

https://gizmodo.com/nvidia-wants-replace-nurses-with-ai-1851347917

who is even responsible if the AI is suggesting you something incorrect and you get seriously ill from over / under dosing the medication. 

This is 10 times worth than outsourcing it to a nearshore telemarketing agency. It basically says your health is so worthless to us that we do not even care to have a human talk to you over the phone. this soulless avatar trained on 2 PDFs is all you get. 

Edited March 25 by o00o

[Satans Little Helper]

2 hours ago, o00o said:

this shit looks strait out of black mirror 

 

 

https://gizmodo.com/nvidia-wants-replace-nurses-with-ai-1851347917

who is even responsible if the AI is suggesting you something incorrect and you get seriously ill from over / under dosing the medication. 

This is 10 times worth than outsourcing it to a nearshore telemarketing agency. It basically says your health is so worthless to us that we do not even care to have a human talk to you over the phone. this soulless avatar trained on 2 PDFs is all you get. 

I understand your concerns. And to a degree they are valid. One thing to take into account though is the growing lack of nurses. So the scenario AI where AI is seriously looked at for being a viable option, is the scenario where people otherwise wouldn't receive care.

When it comes to responsibility, I'd normally expect these technologies to be part of the care people receive when being treated by some care organisation. As is currently the case. And the responsibility will likewise be similar to what is currently the case. Or in other words, these technologies should be brought to clients/patients similar to other technologies which are currently being brought to clients/patients within the healthcare system. Like a prescription by a certified doctor.

What is so frustration about this article in gizmodo is that it starts from the assumption that it's about cost reduction. It seems completely blind for the bigger problem which is the growing shortage of nurses. (which is a world wide issue btw)