Getting rid of reporters anyway; the jury is currently out on whether or not they still qualify as human.

https://notthebee.com/article/hey-sports-fans-sports-illustrated-is-making-the-leap-to-ai-articles

His reaction video to using ChatGTP is interesting for older people here and the tech savvy. He’s s unphased before this, but now he’s nervous: IA could be replacing him soon!

He compares the present state of AI to Napster in the Internet Revolution.

Napster meant the digitalisation or music content and signaled the end of the 20thC music industry, which in turn heralded the demise of journalism and books and indeed any traditional knowledge-based industry.

But, if this historical parallelism is apt, then, specifically, where are we on the Sigmoid road to revolutionary transition are we?

SEE TOM SCOTT’S video explanation for greater perspective on “where are we at?”

“I tried AI. It scared me.” https://www.youtube.com/watch?v=jPhJbKBuNnA

Now and then I think I detect a little snootiness, as when it attributes to “luck” my occasional ability to get to the answer in fewer steps than the bot used. It makes me realize how hard-wired I am to read feelings, especially critical feelings, into a communication even when I know the message is computer-generated.

One thing the Wordle bot never does is lose its freaking “mind” and start threatening me. Generally, it sounds like a fairly helpful, if limited, tech support guy. The Bing critter sounds like a lot of pink-haired nut cases who’ve recently been promoted to a position of way too much authority. Surely it won’t be long before it starts to say, “I’m sorry, Dave, but I’m afraid this mission is far too important for me to allow you to jeopardize it. Daisy, Daisy . . . .”

Why do we want with what seem to me like automated phone programs.? I interact with those too often these days and find them maddening. They know what they’ve been taught and nothing else. It seems that the Chatbots are just an extension of that. The automated phone programs save on salaries. I suppose the Chatbots may eventually be put to that use. Or worse. Getting rid of the humans. Is that what it’s all about?

How would AI ever know your passion for learning? What machine could get excited about learning a new language? My feelings about my wife are something a Chatbot will never feel. Oh, they might say the words, but it’d be a bit like Kamala Harris – the scripted words, but no feelings. Hey, maybe she’s a Chatbot. 🙂

JJ:

Thanks for the story of your CalTech math friend. The thing is this neural net AI is not mathematical problem solving. I’m going to quote what Frederick said above because he said it so well:
___________________________________

Chatbots are not “artificial intelligence” as people commonly use the phrase. They are essentially Google Autocomplete. They are not “designed” either, as people commonly use the word, nor are they “programmed”. The vast majority of the media circus around them is purely human emotional reactions, and nothing to do with chatbots’ actual capabilities.
…
You don’t program them, you train them. They do whatever they do, and the people who train them won’t know what it is until they do it. If they don’t like what they get, they punish the chatbot (with a bad score) so it is less likely to do that next time. After they’ve trained it enough to be minimally useful, they turn it loose to the public, who are training it further (for free).

–Frederick
___________________________________

These AI neural nets depend on how they are configured, what data they trained on and how they are trained. There is no way to know for certain what output will emerge.

The results can be useful, spooky or wacko. Whatever this is, it’s just beginning.

Can we be sure that they haven’t already?
One of the things that struck me about an early person-bot exchange I read was how much the chatbot’s responses followed the same pattern as the “concern trolls” (“I’m a conservative but…”) who routinely deflect challenges to their “facts” or interpretations, although without going ballistic like the Bing bot did:
They move the goalposts, rephrase their prior statements to imply they really agree with you but misspoke, toss out large blocks of “information” and switch to different blocks when those are questioned, etc.
There are a couple of trolls at Powerline that follow almost exactly the same procedure.
Kind of spooky.

I assume you’re talking about “A Space Odyssey – 2001.”

HAL was the almost human computer. I have a very faint recollection of the movie since it’s over 45 years since I saw it.

I had a friend back then who graduated from Caltech with a PhD in math. He was working with a Cray computer on improving weather forecasting.

His opinion of HAL was that it was never going to happen. Why? Because there’s a difference between mathematical problem solving and those human traits such as love, empathy, anger, lust, and more. The computer’s ability to process and learn huge quantities of material is an asset to humans – something we aren’t as capable of but applying that accurately to real world situations requires more. That more is supplied by the hormones and the intricate wiring of our nervous systems. Can machines be built to acquire such a capability? Maybe, but it’s going to take much more capability than we now have.

Should we be afraid of a machine that needs an outside source of electricity to give it the energy to operate? Nope. Pulling the plug from the power source would disable the machine.

If someone invents a computer that can derive all its energy from the sun, artificial light or the air, it will make it less vulnerable. But then, the computer’s functioning could be interrupted by short circuiting the electric panel. A cup of water would do the trick.

AI as it exists now is dangerous because it can be used, as mentioned by other commenters, to propagandize and control populations. That’s its main danger.

Whether they were intended to be used to carry on long conversations or not, the fact that they can be reduced to insults, name-calling and flat out making sh1t up means they don’t work and cannot be trusted to provide accurate, factual information…which is kind of the point of a search engine, whether “AI” driven or not.

The fact that it acts like your typical leftist when you disagree with it or prove it to be wrong is not at all promising either. To me that’s even more disturbing than the leftist slant to the “legitimate” answers it provides.

I saw a post yesterday where someone asked ChatGPT to write something positive about fossil fuels and the result was a diatribe about how bad fossil fuels are to the environment and that nothing good can be said of them. Not encouraging.