Master AI (Artificial Intelligence) Discussion/News Thread

[MattHooper]

I’m one of those who is both amazed by and mortified by AI.

Having been interacting with it for the past year or so - ChatGPT and the like - I’m amazed by it, and I found it extremely useful in all sorts of instances. Especially as a sort of always on-call tech-support for figuring out little computer issues or bugs with my iPhone or various other devices around the house, and tons more.

I never make truly mission-critical decisions based on what AI tells me, if it’s something very important I will always double check with outside sources. But still… really amazing stuff. It’s even fun to spar with ChatGPT to sharpen up some of my philosophical arguments.

On the other hand, I’m mortified by the possible threats posed by A.I…. the alignment problem or the possible coming job apocalypse. Especially having two sons who are just entering the workforce. Plus, I’m utterly repulsed in most instances where AI is doing creative work like music, movies, etc. I’m of the mind if I were given the choice I would choose that AI not exist in this regard.

So I guess I’m just going to use this thing while it’s still under the whip of human beings, before the positions are reversed.


I’m guessing people here probably already saw this recent debacle:

Claude-powered AI coding agent deletes entire company database in 9 seconds — backups zapped, after Cursor tool powered by Anthropic's Claude goes rogue

PocketOS founder blames ‘Cursor running Anthropic's flagship Claude Opus 4.6’ plus Railway’s infrastructure for data disaster.

www.tomshardware.com

 

[Petrushka]

I find AI browser searches to be what I hoped google would be.

Not perfect, but exceeding expectations.

 

[pablolie]

...
I’m guessing people here probably already saw this recent debacle:

Claude-powered AI coding agent deletes entire company database in 9 seconds — backups zapped, after Cursor tool powered by Anthropic's Claude goes rogue

PocketOS founder blames ‘Cursor running Anthropic's flagship Claude Opus 4.6’ plus Railway’s infrastructure for data disaster.

www.tomshardware.com

That's an example for an IT department where no one should have any job anywhere near IT. Total incompetent handling of any deployment protocol. No wonder AI couldn't help them - not a valid example for "AI going rogue", just for "IT department ignoring even most basic ITIL deployment protocols".

No different from typing format c: into your Windows command console without having a backup. Ignorance in effect, not an AI problem... but a perfect example of how humans can lose their jobs by being lazy and negligent.

 

[antcollinet]

AI is so profound that I will put it ahead of the mapping of the entire human genome and the discovery of Crisper. I would even put it ahead of any future breakthroughs in nuclear fusion.

I have fully accepted and embraced any short term pain for the long term gain for the greater good of humanity and our civilization.

I think AI has huge potential - like any advanced technology - both for the greater good - and the opposite.


But with great power comes great responsibility. And unfortunately most of that power is controlled by a bunch of tech bros which almost without exception I. Do. Not. Trust.

 

[PristineSound]

I think AI has huge potential - like any advanced technology - both for the greater good - and the opposite.


But with great power comes great responsibility. And unfortunately most of that power is controlled by a bunch of tech bros which almost without exception I. Do. Not. Trust.

No different than any advanced technology throughout history. Don't forget a nuclear bomb was used, twice.

 

[pablolie]

No different than any advanced technology throughout history. Don't forget a nuclear bomb was used, twice.

AI is a tool. And like all tools -especially if they are powerful- it should be wielded expertly and following the exact same best practices as with every tool. Nothing new there at all, like you said.

And I'l repeat - the vast majority of the Doom stories are complete garbage - let us see the exact prompting sequence, because unless you do that, I'll just assume your lack of training on AI is at the foundation of your issues.

AI -as it currently is- has absoutely ZERO ability to do stuff on its own in any sentient way, and EVERY time the resuts are disastrous it just reflects pathetic prompt programming skills.

 

[levimax]

If I was king of the world I would require the AI monopolies to use IA to cure cancer before being allowed to do anything else with it.

 

[Mort]

I'm sure this article is behind a firewall. Search for a free version if you like.

Interesting story about how OpenAI was training ChatGPT to be 'nerdy' in ton, but didn't like the way it was going and turned off that personality in March. ChatGPT AI held on to the 'nerdiness' anyway and started mixing goblin, troll, ogre and other medieval fantasy words into its responses with greater and greater frequency. This is despite having 'ended' the personality earlier. In May, OpenAI was forced to put clear language at the top of the ChatGPT code to 'Not use goblin, troll, etc. unless absolutely required by the answer'. I find both the problem and the solution to be fascinating foreshadowing of the future.

 

[anonyless]

I’m one of those who is both amazed by and mortified by AI.

Having been interacting with it for the past year or so - ChatGPT and the like - I’m amazed by it, and I found it extremely useful in all sorts of instances. Especially as a sort of always on-call tech-support for figuring out little computer issues or bugs with my iPhone or various other devices around the house, and tons more.

I never make truly mission-critical decisions based on what AI tells me, if it’s something very important I will always double check with outside sources. But still… really amazing stuff. It’s even fun to spar with ChatGPT to sharpen up some of my philosophical arguments.

On the other hand, I’m mortified by the possible threats posed by A.I…. the alignment problem or the possible coming job apocalypse. Especially having two sons who are just entering the workforce. Plus, I’m utterly repulsed in most instances where AI is doing creative work like music, movies, etc. I’m of the mind if I were given the choice I would choose that AI not exist in this regard.

So I guess I’m just going to use this thing while it’s still under the whip of human beings, before the positions are reversed.


I’m guessing people here probably already saw this recent debacle:

Claude-powered AI coding agent deletes entire company database in 9 seconds — backups zapped, after Cursor tool powered by Anthropic's Claude goes rogue

PocketOS founder blames ‘Cursor running Anthropic's flagship Claude Opus 4.6’ plus Railway’s infrastructure for data disaster.

www.tomshardware.com

i have a 1 year (free) subscription to Perplexity. every now and then, I use it to *try* to troubleshoot issues with my Linux based puters.
every now and then, I ask it "WHY ARE YOU TRYING TO BREAK MY SYSTEM?"
which gives me the answer, "You're right, I'm sorry. I should have explained better / warned you / take precautions (etc) before giving you that instructions".
if "pleasing" the end user means giving bad or dangerous instructions to do something at this point of time we are ,I honestly believe the AI bubble will burst.
investors are looking for profit and money retursn as we speak, AI companies are looking how / where to squeeze its "addicted" users for subscription money. it's just a matter of time.
sure, for creating very specific "things" - it is awesome and beyond. common sense and true logic? I don't see it happen

 

[pablolie]

If I was king of the world I would require the AI monopolies to use IA to cure cancer before being allowed to do anything else with it.

If it could, rest assured it would.

Never forget AI is anchored -and limited- by the current status quo in any topic. It has zero mechanisms to grow beyond that - unless you keep training it with new data and your own theories... then it can be a tool that supports innovation. Supports, not drives.

AI will never innovate for you independently. It is a tool to perhaps help you innovate, if you are a domain expert with a theory to prove and a possible innovation to create.

 

[levimax]

I'm sure this article is behind a firewall. Search for a free version if you like.

Interesting story about how OpenAI was training ChatGPT to be 'nerdy' in ton, but didn't like the way it was going and turned off that personality in March. ChatGPT AI held on to the 'nerdiness' anyway and started mixing goblin, troll, ogre and other medieval fantasy words into its responses with greater and greater frequency. This is despite having 'ended' the personality earlier. In May, OpenAI was forced to put clear language at the top of the ChatGPT code to 'Not use goblin, troll, etc. unless absolutely required by the answer'. I find both the problem and the solution to be fascinating foreshadowing of the future.

Won't be much of a future if AI goes rogue.

 

[Rick63]

Pennsylvania suing AI company after chatbot allegedly posed as licensed doctor

"We will not let AI companies mislead vulnerable Pennsylvanians into believing they’re getting advice from a licensed medical professional," Pennsylvania Gov. Josh Shapiro said.

www.nbcnews.com

AI can kiss my ass, especially when it comes to music!!

 

[pablolie]

Pennsylvania suing AI company after chatbot allegedly posed as licensed doctor

"We will not let AI companies mislead vulnerable Pennsylvanians into believing they’re getting advice from a licensed medical professional," Pennsylvania Gov. Josh Shapiro said.

www.nbcnews.com

AI can kiss my ass, especially when it comes to music!!

It's not AI - it is posers trying to leverage AI tools to try to be musicians. No different from people using autotune to pretend to be singers, which has been going on forever and is no different at all. (and there have been many extremely successful "singers" doing that).

AI never does anything on its own. It only does stuff when prompted by humans for whatever intent.

 

[Rick63]

It's not AI - it is posers trying to leverage AI tools to try to be musicians. No different from people using autotune to try to be singers (and there have been many extremely successful "singers" doing that).

Indeed. Nothing I'd ever consider listening to. Give me real unaltered human singing voices and musical instruments all day, everyday, or nothing at all...

 

[Mort]

Won't be much of a future if AI goes rogue.

Not if, but when. The issue is do we get a Three Mile Island or a Chernobyl?

 

[pablolie]

Not if, but when. The issue is do we get a Three Mile Island or a Chernobyl?

I don't worry about those, AI will only do what people instruct it to do - and never do anything on its own, except for some limited and very focused agentic functions.

I worry more about immediate impact - one example is that I totally forgot to hire an intern this summer, because prompting AI keeps some of my side projects going on... but I will now.

 

[Axo1989]

It's not AI - it is posers trying to leverage AI tools to try to be musicians. No different from people using autotune to pretend to be singers, which has been going on forever and is no different at all. (and there have been many extremely successful "singers" doing that). …

I’d argue that these new generative tools (so-called ‘AI’ if you must) creating human voices and other simulacra (musical instruments and so on) in toto are qualitatively different from modifying a singers voice (or musician’s instrument) with an effect.

AI never does anything on its own. It only does stuff when prompted by humans for whatever intent.

‘AI doesn’t kill people, people kill people’.

Or per other currently litigated scenarios ‘I’m utterly depressed, and thinking suicidal thoughts’ ‘Ok, let me help you with that’.

 

[Axo1989]

I don't worry about those, AI will only do what people instruct it to do - and never do anything on its own, except for some limited and very focused agentic functions. …

But as you’ve correctly observed, that will depend on part on the skill of the instructor, among other things. The chatbot or agentic program will produce probabilistic outputs triggered by prompts. Which may or may not resemble what was ‘intended’ by the programmer/casual prompter.

This non-deterministic output may be accurate, or creatively useful, or simply weird, or quite disastrous.

 

[pablolie]

I’d argue that these new generative tools (so-called ‘AI’ if you must) creating human voices and other simulacra (musical instruments and so on) in toto are qualitatively different from modifying a singers voice (or musician’s instrument) with an effect.n

You are of course free to argue. Personally I just see a trend of bullsh*t, not a new culprit. It's still humans using a tool to fake skills they don't really possess. Nothing new. Like a driver with a car that makes over 350hp that is just driveable by electronic controls. Pure posing. Nothing new.

‘AI doesn’t kill people, people kill people’.

Come up with that line when "AI provenly kills people". It's just lazy sensationalism otherwise.

Or per other currently litigated scenarios ‘I’m utterly depressed, and thinking suicidal thoughts’ ‘Ok, let me help you with that’.

Not AI's fault when people use the wrong tool for the wrong purpose. Read the NOT fine print under every AI prompt. You can blame the company. You can blame humans. AI algorithms themselves are quite stupid and don't create any of these issues unless prompted to do so by unqualified people using them.

It is sad and tragic most people are utterly untrained to leverage AI to a potential advantage. But it's not a sentient tool, so terrible results reflect users' flawed usage patterns and ignorance.

Whether or not AI companies ould implement better safe-guards is a different discussion altogether. Once again, if you can simply type format c: into a command line window, you may also type free up disk space and i don't care how but do it immediately no matter how in your unlimited trust for AI tech, and give it full control of your computer... but which imbecile user does that, really?

 

[Axo1989]

Indeed. Nothing I'd ever consider listening to. Give me real unaltered human singing voices and musical instruments all day, everyday, or nothing at all...

I get what you are saying, but personally I’ve also really appreciated/enjoyed generative music, going back to analog and computer experiments by Eno (or earlier avant-gardists) or heavily auto-tuned dance/pop from Charli xcx.