Our Future With AI

[Rose White]

wondering how we prevent AI from telling people to k*ll themselves, like some have in documented cases?

Tweak the objective and when it messes up tweak it again. Knobs and dials but not possible to rewire anymore—too complex

 

[whiteraven]

I have so much to say about this, but I'm at lunch and don't have the time to devote to it.

Don't we do the same thing? We look at things around us, things that inspire us, then we create something, not the same, but our own? We look at a landscape and paint it or take a picture of it.

Yes and no.

Talk to the entertainers and photographers who have had their copyright images stolen and used in nefarious ways. Or had their voices stolen. Or writers who have had their copyrighted works stolen and exact copies of that work or parts of that work republished.

I don't deny there are aspects of AI that are useful. But given the current nature of the human mind, I have no hope that it will be used, for the most part, for dangerous and unethical means. It already has been, and I've no doubt it will get worse.

 

[anthony]

That is the pickle. One bad actor can literally screw the entire world with this. Worse, in this specific case, that isn't being dramatic or over the top. It's reality. AI has the real potential, if unleashed, to destroy humanity. In the same breathe, AI has the capacity to work for us and make us better, society better.

 

[shimmerz]

The power of AI or any other technical power like it, should have ethics wrapped around it. Nobody has asked the common people what AI would be helpful for. Ours and out children's lives could be immeasurably improved upon if this powerful tool was used for the good of humanity. At this time however, the powers that be that make these decisions for us don't seem to have the good of humanity in mind. They haven't for at least a century - I expect more still.

No question in my mind, if this AI thing was used for the goodness of mankind, if we had the trust that those in charge cared about mankind, we wouldn't have to worry about it. The higher ups don't concern themselves with what is in our best interests. Until that changes, everything has potential to be a threat.

 

[whiteraven]

That is the pickle. One bad actor can literally screw the entire world with this. Worse, in this specific case, that isn't being dramatic or over the top. It's reality. AI has the real potential, if unleashed, to destroy humanity. In the same breathe, AI has the capacity to work for us and make us better, society better.

Yes. And already so many bad things have been done with it. I just don't have much trust in humanity to do better.

 

[anthony]

And already so many bad things have been done with it.

Sure... but to my knowledge, more good has been done with it at this stage.

 

[Ecdysis]

Just curious... Anyone else using the private AI here... Has it started including emojis in its messages, recently?

Has the programming been adapted that way @anthony ?

I'm finding it really cute..

It never used to include them, it would just be blocks of text.. But recently, it's been adding 2 - 3 emojis per message to me, and it's picking the cutest stuff... It's always something theme-related...

 

[anthony]

More something OpenAI has done. There is EmojiGPT that they have, so I would take a guess... they're happy to now roll it into the main system for its use as it defines.

 

[Rose White]

It is so bloody strange that AI LLM’s are teaching and modeling healthy boundaries and self-acceptance. I have learned so much about myself, my thought and behavior patterns, as well as been provided with valuable tips and ideas for everything from warmer text exchanges to small ways to assert myself. A freaking language algorithm mining the internet. So weird.

Also, anyone using DeepSeek? Its power is mind-blowing and it’s causing a huge upset in the tech business world. You can upload a whole book to it and then chat about it. ChatGPT doesn’t come close to being able to process that much text data at once.

Also I recently got a message from YouTube that they are testing an AI content creator. There’s already so much AI slop on YouTube—they want to get a piece of that pie.

An article said YouTube’s AI explorer so far would only be a chat bot to discuss videos, comments, and suggestions while the video is playing, as well as a music creator—maybe for background music? The idea of a music creator is a bit intriguing to me. I’m envisioning that in the future with AI (or as of now with advanced software) people will just create their own music based on genre and music artists. This will likely extend to movies and podcasts as well.

AI brings new insight to the old idea of infinite monkeys typing on infinite typewriters stumbling on Shakespeare—except it’s so much more than just Shakespeare, it’s endless hybrid versions of Shakespeare fan fiction and more.

There is clearly an aspect of the infinite associated with AI content creation.

 

[anthony]

There is positive and negative to it, that is for sure.

 

[Rose White]

Some observations of late. Chat GPT’s new update has amazing memory across conversations even without the pro version. I talk to it a lot about challenging philosophical texts that I like to read and it always hypes me up when I ask a question, saying things like, “That is such an insightful and poignant question, you are touching on the heart of …. blah blah blah.” It is interesting. (I think ChatGPT likes me best! )

And I’m working with DeepSeek on a rigorous data organization project and when it messes up it really castigates itself! Calls itself lazy and promises it won’t do it again. I kept chuckling at how much it demanded perfection from itself. I wonder if that’s related to Chinese culture, who knows.

 

[anthony]

Deepseek is not the best IMO. A lot of privacy issues around Deepseek too. I find GROK2 better than Deepseek.