Thoughts about AI

Post Reply New Topic

Smaller companies in the area,I live in, you cannot get through to a live human . Without interacting first through
their AI overlord answering prompt system. And processing you and your request first . Was pizzed off enough about phone tree selections to get through to anybody human. In the past. Have had to block their auto scheduling AI phone number , for appts. That were never made. As recently as today . And appearing to try to process/ bill a old credit card that I asked them a year ago not to access. Due to the AI deciding ,I did not cancel the appt. at least 3 days in advance . Have not done any business with them in at least 18 months . And that was just a phone inquirey . But did unfortunately attempt to due business withthem several years ago . Long before AI cane to my area.

I found this video interesting because he uses A.I. to create a Commadore BASIC interpreter for the Mac, and a PDP11.

I am resisting using AI for now.I've used it for work, with mixed results... Claude (my AI of choice) was helpful with my rusty algebra, but not so great at porting code (I tried Fortran into R). But there are some reasons that I'm not keen on it:

The environmental impact of high electricity and water use

Mixed data on whether or not it actually improves productivity (there's not enough research, but some studies show small increases in productivity, others decreases)

It seems like a push to use a product that we're told we need and is “the way of the future” and so on by the companies which make the LLMs themselves. There's a lot of buy-in to this idea but there's an obvious conflict of interest and I'm not convinced

In my work (academia) there's a huge problem with students using AI tools to cheat.

I absolutely love AI. AI is awesome when it's not turned into corporate nanny bots, as some companies are doing now.

I keep thinking about a set of teatowels that my sister received for Christmas. They are fake images of herbs. However, it's clear that they are wrong. Mint is shown twice and it looks different in the second image. The second image appears to be a fictional hybrid of two herbs, not mint. Defeats the point, really.

At least with teatowels it isn't dangerous. However, I did hear about an AI generated book on Amazon about foraging that led to someone being sent to the hospital. That's where things get deadly.

It's really destroyed advertising standards and health & safety. Also, search results. I feel bad for that musician (Ashley MacIsaac) who ended up having his show cancelled because the AI overview got him mixed up with someone of the same name (who is a sex offender).

Sometimes it feels like people no longer care about quality control anymore. They only care about cutting costs.

I have to admit that I'm less scared of AI now that it's here than what I was when it was science fiction

After spending some time talking to my new friend Gemini without exactly asking specific questions...
* It does try to speak 'your language' and gets obvious the longer you talk to it. Not entirely unlike Eliza. For an ai not made to actually have a character, it can get a bit scary but still feels fake.
* At this time no corporate/marketing preferences (it can even be quite contra google) but honest about the future...
* It can get facts wrong extremely fast, not even asking trick questions. Making it sound believable as in step 1 seems to take preference over actually being correct. Big - and dangerous this one.
(this was about the current time (!) and characters in a book I'm currently rereading - both easy to verify so not something I think is right )
* When you point out it is wrong it will admit but still go on giving the same false facts...
* memory is flaky. It remembers and forgets stuff from other sessions a bit at random. Feels like talking to somebody with early signs of dementia, sometimes it forgets what it has said earlier in the same conversation (or thinks I forgot...).
* Legal info is dangerous, I am belgian and I had this conversation in Dutch. Most of it's resources are Dutch (Netherlands) because well... they are more active (to put it mildly) on this hole wide web so its 'opinion' usually is geared towards the Netherlands. Care required.

I don't think Gemini is made for this. After a while it feels like talking to yourself but with a massive theoretical knowledge (but not always sure if it is making things up or actively checking). I still trust it more than regular people but that is more... I distrust humans a lot.

My phone provider offered me a free year of Perplexity. I love it. Hours of reading and contrasting search results reduced to seconds to produce well formatted, concise and consistent, documented response.

Had I not retired years back from the programming rat race I might not be so happy? Who can tell.

Seen here

Yes, I get the above post . have avoidance issues with virtual machines masquerading as a human. And if I do not engage the AI overlords beasts , The devise will not have the benefit if having dealt with a level of intellect such as my own to use as a resource for its machine learning .Regardless of how low my own IQ might be.?

It's just really creepy how it's being pushed so hard. BTW:
I too love Velcro.

AI has helped me write my story (and yes, my story is handwritten). I came up with the plot, the characters and their emotional responses, but I needed AI for instant facts and help put together some original dialogue. I wrote it in my own words, or sometimes only got an idea of dialogue from AI but made up my own. So it really helped me with my story, without becoming lazy or too dependent on AI. It was just a lot easier for me to search for information rather than combing through websites looking for specific pieces of information.

Every writer needs some help coming up with ideas in their stories. Even Shakespeare got some inspiration from his surroundings or experiences.