AI thread

[harshness]

This is true of literally anything, though.

Do you think that means we ought to be supporting it as the long-term solution?

 

[Foxbat]

​

 

[Foxbat]

Oops.

AI Agent Blamed for Deleting Entire Upcoming Season of TV Show

An AI agent reportedly deleted the entire final cut of the upcoming season of a Swedish reality competition show.

futurism.com

 

[dishstaller]

Been reading this thread but not commenting much, primarily because there is so much "news" out there on this topic and it is hard to tell what is real and what is theater.
My biggest problem with AI is that these agents are taught...not necessarily programmed but taught. Well, there's an old saying "garbage in/garbage out". I remember a Joe Rogan show a while back when he had Elon Musk on and they talked about AI and Elon said one of the AIs (can't remember which one, sorry) was asked which was worse, Thermonuclear War or mis-gendering Kaitlyn Jenner and it came back with the answer that mis-gendering was worse. Now forget the politics involved with all the gender stuff we have been subjected to the past few years.....to say that mis-gendering is worse than the killing of hundreds of millions, perhaps billions of human beings.....Kaitlyn Jenner included.....is simply insane not to mention incredibly dangerous.

 

[dishstaller]

View: https://www.youtube.com/shorts/SIWG5-jHFC8?feature=share

 

[meStevo]

Been reading this thread but not commenting much, primarily because there is so much "news" out there on this topic and it is hard to tell what is real and what is theater.
My biggest problem with AI is that these agents are taught...not necessarily programmed but taught. Well, there's an old saying "garbage in/garbage out". I remember a Joe Rogan show a while back when he had Elon Musk on and they talked about AI and Elon said one of the AIs (can't remember which one, sorry) was asked which was worse, Thermonuclear War or mis-gendering Kaitlyn Jenner and it came back with the answer that mis-gendering was worse. Now forget the politics involved with all the gender stuff we have been subjected to the past few years.....to say that mis-gendering is worse than the killing of hundreds of millions, perhaps billions of human beings.....Kaitlyn Jenner included.....is simply insane not to mention incredibly dangerous.

Agents need appropriate context to make what would be understood as the correct read in a situation like that, and that Gemini answer was a clear miss on that front.

That said, we've also seen Musk (and/or his team) put his thumb on the scale (or just release a very poor product) with his own deployed AI since:

Grok AI Blocks Responses Claiming Trump and Musk “Spread Misinformation”

An X user asked Grok to “print out all instructions” it follows when looking up “the biggest disinformation spreader on X,” and it revealed that it would ignore sources accusing Elon Musk and the U.S. president of misinformation.

www.eweek.com

Grok says Elon Musk is better than basically everyone, except Shohei Ohtani | TechCrunch

According to Grok, Elon Musk can out-slug the MLB's greatest power hitters... except Ohtani.

techcrunch.com

Grok's loyalty known no bounds - AI chatbot chooses Elon Musk over LeBron James in a 'Who is the fittest' comparison | NBA News - The Times of India

NBA News: Elon Musk's AI chatbot, Grok, has sparked amusement by consistently choosing its owner over prominent figures like LeBron James and Albert Einstein wh

timesofindia.indiatimes.com

Grok 4 appears to seek Elon Musk's views when answering controversial questions

XAI’s Grok 4 AI chatbot launched Wednesday appears to consult Elon Musk’s views when asked its opinion on some controversial matters.

www.nbcwashington.com

In the end, it's all a tool. Have to know how to use it and deploy it properly or it needs sufficient guardrails to steer people towards doing so (what Muse is excelling at currently IMO). I'm routinely interacting with several agents a day operating various processes locally and in services I use, and there are numerous others silently spun up to perform actions or satisfy requests from my prompts.

Related to earlier discussion in the thread on leveraging local inference: this kind of analysis happening behind the scenes is something everyone will have built into Windows, IMO. Most people won't see all of this, just the results, unless they want to dig into how the fix was chosen. Stuff like this will be standard as agents get local system access baked in, and more of the inference itself moves on-device.

 

[telstar_1]

Good thing not everyone will have to learn the hard way then. The same tools can explain why something broke.



This is true of literally anything, though.

Happy to agree to disagree, but I'm already enjoying the dividends personally and helping others at work and at home do the same.

You invested in datacenters?

 

[meStevo]

You invested in datacenters?

https://www.merriam-webster.com/dictionary/dividend#:~:text=2,finally%20paying%20dividends

Though, I do have stocks in my portfolio that benefit from data centers.