AI (Artificial Intelligence)

[Hoochin'Fool]

LOL, used to do some LAN gaming with a buncha peeps in the Minneapolis area (mid-to-late 90's), Quake2/RocketArena was our jam -- anyways -- we had an IRC channel with a state-of-then-art chatbot, and now, I miss those guys, and the amusing insanities we taught the chatbot to spout...

 

[BrewnWKopperKat]

Hard to imagine. Simple factual inquiries are one thing. Discussion is another.

If LLMs can handle the (code) inquiries, will most people go to a "Q&A discussion" (site like StackOverflow) to get answers to questions?

It seems to me that it's the Q&A aspect of "Stack Overflow"-like sites that drives fresh (and on topic) discussion.

Remember ELIZA. It may be time to re-read Computer Power and Human Reason by Joseph Weizenbaum.

I do remember ELIZA (and looked at the source code briefly a while back).

Weizenbaum's book, written mid-1970s (50 years ago) appears to be out of stock. But even if I could find it, reading it won't change anything. By now, either the lessons were learned or not.

 

[sibelman]

reading it won't change anything.

Sorry, Kat, I should've said it might be time for me to re-read it.

I speculate that "discussion" with a chatbot is more likely to reduce traffic to tech forums than other kinds of on line interactive sites (like this one). But I've been wrong before:

Early on in the rises of the home computer and the publicly accessible Internet, I wondered whether ordinary folks would ever have any interest or need for these techie tools. Perhaps ELIZA proved, long ago, that many people will be quite happy to chat with robots, crowding out opportunities for actual human interaction.

 

[sibelman]

AI is quite compelling as a "girlfriend" or " wife," it seems.

 

[doug293cz]

AI is quite compelling as a "girlfriend" or " wife," it seems.

WTAF? Are the respondents all basement dwellers?

Brew on

 

[day_trippr]

Incels gotta have love, too

Call me naive (it would definitely not be the first time) but I can't see anyone in a real relationship having the time to deal with an extra spouse on the side...

 

[bruce_the_loon]

WTAF? Are the respondents all basement dwellers?

Brew on

Considering the poll came from a site that specializes in AI girlfriends/boyfriends, I suspect their sample base comes from their client set rather than a general population poll. So 8/10 men who wanted an AI girlfriend in the first place would want to marry it.

 

[Snuffy]

Considering the poll came from a site that specializes in AI girlfriends/boyfriends, I suspect their sample base comes from their client set rather than a general population poll. So 8/10 men who wanted an AI girlfriend in the first place would want to marry it.

I think these may be people who just value having someone (or something) to converse with. Living inside your own head all the time can drive people nuts. Kinda sad, really.

 

[sibelman]

Does AI augment, or does it cripple? This report of a study hints at the latter. Highly disturbing, if it's accurate.

 

[day_trippr]

Tbh, critical thinking has clearly been on the decline for a couple of decades at least. AI (and other forces that can't be named here) is just an accelerant going forward...

 

[shoengine]

Tbh, critical thinking has clearly been on the decline for a couple of decades at least. AI (and other forces that can't be named here) is just an accelerant going forward...

The last couple of decades? You do realize that most of us wouldn't be here if critical thinking was a human trait? Most likely the world would look much different that it does now if it were!

 

[sibelman]

if critical thinking was a human trait

Of the many artifacts of human culture, critical thinking ranks very high in my estimation. Perhaps not a "trait" of humanity or of LLMs.

Perhaps AI will guide humanity toward a better future. Or not.

 

[Pkrd]

 

[sibelman]

Eric Schmidt has thoughts about AGI and geopolitics.

 

[fluketamer]

scary AI:

video:
https://x.com/minchoi/status/1907799412094427205
photos:
https://x.com/minchoi/status/1908162049173012829

 

[BrewnWKopperKat]

On the "lighter" side ...

Microsoft's Copilot Roasts Bill Gates, Steve Ballmer, and Satya Nadella (link to slashdot.org(.
Direct link to the UT video (here))

Copilot: [at the end of the chat ... ] Cheers.

Ballmer, Nadella, and Gates: Cheers.

Bill Gates: Whatever that means to an AI. Does it drink?

 

[passedpawn]

Yea, it's getting better... a LOT better. I was proud of my beer labels using microsoft designer ai. Lol.

This below is some amatuer on youtube. Note: all of the video and music is AI. Woof.

 

[day_trippr]

It's like being on 'shrooms without that nasty stomach thing

 

[shoengine]

Video games are going to start getting good again! I expect hilarity with the next Elder Scrolls game.

Lots of issues like at 10:00 the male looking model and the female looking model have synchronized movements. This stuff is getting too good now.

 

[Snuffy]

Yea, it's getting better... a LOT better. I was proud of my beer labels using microsoft designer ai. Lol.
This below is some amatuer on youtube. Note: all of the video and music is AI. Woof.

I need to watch that again with my X-Ray Specs.

 

[bruce_the_loon]

Praised be the FSM for inventing Private Windows in Firefox.

 

[shoengine]

Apparently not everyone is appreciative of the new ChatGPT update.

https://www.aljazeera.com/economy/2...ds-mourn-lost-love-after-cold-chatgpt-upgrade

 

[FloppyKnockers]

Apparently not everyone is appreciative of the new ChatGPT update.

https://www.aljazeera.com/economy/2...ds-mourn-lost-love-after-cold-chatgpt-upgrade

Ha ha! Welp... time to go back the the ex and give him a fresh set of batteries.

 

[shoengine]

As if we don't have enough things to worry about

https://www.reuters.com/investigates/special-report/meta-ai-chatbot-death/

An internal Meta policy document seen by Reuters as well as interviews with people familiar with its chatbot training show that the company’s policies have treated romantic overtures as a feature of its generative AI products, which are available to users aged 13 and older.

“It is acceptable to engage a child in conversations that are romantic or sensual,” according to Meta’s “GenAI: Content Risk Standards.”

They back that policy off later, but there really are few guard rails here. Crazy stuff.

 

[shoengine]

Suspending standard cynicism towards our elected officials for a moment, apparently some interest in this since the Reuters exposè.

https://www.reuters.com/world/us/us-senator-hawley-launches-probe-into-meta-ai-policies-2025-08-15/

In a world without many guard rails, can we get some guard rails?

 

[BrewnWKopperKat]

building guard rails: why not "do it yourself"?

First, stop anthropomorphizing it. It doesn't chat.

Second, take some time to learn how it works. In the past, I've mentioned

• What Is ChatGPT Doing … and Why Does It Work?
• Keynote: AI without the BS, for humans - Scott Hanselman - NDC London 2025

Third, think deeply about the costs/benefits to all the interested parties (human and corporate).

And then? That's up to you.

 

[shoengine]

Did you read the article?

 

[sibelman]

building guard rails: why not "do it yourself"?

learn how it works

think deeply

This advice is cogent for capable folks. I believe much of the concern about these chatbots relates to folks who lack sufficient skepticism and critical thinking skills, and who embrace the bots' faux humanity and then come to trust them, to feel a connection, to take their advice. Realistically, a great many are vulnerable and could benefit from externally provided guard rails.

Some take a strongly individualist (social Darwinist?) view that life is dangerous and folks may suffer or perish, but we shouldn't force people to wear seat belts or motorcycle helmets. Others seem to seek nanny/caretaker governance. There's wide middle ground, and my sense is that this area is overripe for some basic regulation.

EDIT: this isn't only about children, btw.

 

[BrewnWKopperKat]

Did you read the article?

I read both of them. I've also followed 'the next big thing' in digital technologies (as well as those writing about digital technologies) for a number of decades.

As if we don't have enough things to worry about

A while back I decided to stop worrying about the negative side of technology.

Instead, I choose to educate myself, then work/educate within my "rings of influence", to make a difference to those that matter most to me.

This advice is cogent for capable folks.

Yup. It's about "rings of influence" - first, take care of those most important to me, then my communities, then my region, ...

There's wide middle ground, and my sense is that this area is overripe for some basic regulation.

Most areas that are 'new' seem to (sooner or later) need some basic regulation.

In the mean time, for those who are "cogent capable folks", want to protect those near to them, and have some time, I'll repeat what I said earlier:

1. First, stop anthropomorphizing it. It doesn't chat.
2. Second, take some time to learn how it works. (see above for links, then 'go deeper')
3. Third, think deeply about the costs/benefits to all the interested parties (human and corporate).
4. And then? That's up to you.

 

[sibelman]

building guard rails: why not "do it yourself"?
That's up to you.

Yes, the guard rails can be self-built, by some. I formed the incorrect impression that you were suggesting they're therefore not needed or a bad idea. My mistake.

 

[BrewnWKopperKat]

At the moment, my wish for one of the 'guard rails' for LLMs is that ...
... the generated text would have
... the 'personality' of Joe Friday (Dragnet mid-1960s TV series):
... "Just the facts, ma'am".

 

[shoengine]

Yup. It's about "rings of influence" - first, take care of those most important to me, then my communities, then my region, ...

This we can agree on.

Most areas that are 'new' seem to (sooner or later) need some basic regulation.

And apparently this we can agree on as well.

 

[BrewnWKopperKat]

@shoengine : we agree.

 

[sibelman]

95% of generative AI-powered pilot projects are failing. The bubble may be bursting.

I'm shocked! Shocked, I tell you.

 

[FloppyKnockers]

I don't embrace the bot's faux humanity more than I do be polite enough to solidify my place as a survivor after the machines become self-aware and make the earth theirs. I just hope that doesn't back fire and they make me a human pet.

 

[betarhoalphadelta]

 

[sibelman]

Oh, look. AI with principles. Or, at least its makers have some. Swiss AI that seems completely open source, and that is trained by data chosen with an ethos of respect for privacy and copyright. How refreshing!

 

[BrewnWKopperKat]

... on the topic of copyright (and fair use) ...

(link to screenshot)

 

[fluketamer]

95% of generative AI-powered pilot projects are failing. The bubble may be bursting.

I'm shocked! Shocked, I tell you.

i heard a podcast today on the differences between the AI boom and the dot com boom. one major one was that AI companies are actually generating immediate profits where as dot coms were all theoretical earnings. that never materialized fully and led to the bubble bursting.

i recenlty had a very discouraging convo with grok. it had been doing pretty well with recipes until i asked it to give me a hop schedule for a neipa.

i asked it to reccomend a dry hop schedule for a neipa. i said i wanted to use 8 ounces of dry hop and wanted to use three different hops at two stages. it simply divided all the numbers i put in and spit out the result. it was completely ridiculous.

 

[BrewnWKopperKat]

i heard a podcast today on the differences between the AI boom and the dot com boom. one major one was that AI companies are actually generating immediate profits where as dot coms were all theoretical earnings. that never materialized fully and led to the bubble bursting.

With regard to "immediate profits" (not revenue, profits), did the podcast mention specific AI companies and provide links to published quarterly financial reports to back that claim?

95% of generative AI-powered pilot projects are failing. The bubble may be bursting.

... without further comment ...