AI starts out by being taught by a human brain.
-
Enjoy XS650.com? Consider making a donation to help support the site.
XS650.com receives a small share of sales from some links on this page, but direct donations have a much greater impact on keeping this site going.
Should we be afraid of Artificial Intelligence?
- Thread starter Mailman
- Start date
I'd have to point out that we've been sidetracked by chatbots.
ChatGPT, Grok, and all the others are nothing more than next-word predictors, like when you start to type into a search engine. Well, ok, the companies promoting them have 'trained' them by feeding terabytes of internet text so the machine does more than find the next word, it suggests some taradiddle boiler-plated from text matching the context. I called it boiler-plated, they call it generated; semantics.
But just as there's more to human intelligence than typing rubbish on a computer keyboard, there is much more to AI than chatbots. They've used AI for ages in the stock-markets to anticipate price movements. With some pretty worrying consequences such as accelerating & exaggerating those movements.
More usefully, AI has been assisting in medical fields including scanning tissue samples to 'find' cancerous cells. Or developing highly complex molecules for use in medicines.
Also, developing microchips for specific processing tasks - that one is a bit worrying because the humans using the AI can't see how these AI developed chips actually work, though testing shows they do work.
Today on breakfast radio, Sir Brian Leveson was talking about his inquiry into the UK criminal justice system which has become sclerotic with not enough judges and magistrates and growing backlogs of court cases. One very large factor he identified in cases taking a long time to come to court and longer to process through the courts is the vast amount of data from phones and computers which has to be sifted and analysed to see if there is anything pertinent to the case. I immediately thought, prime area to employ an expert system which could process thousands or even millions of emails and text messages to sort the wheat from the chaff. Would potentially save a huge amount of time for both defence and prosecution lawyers.
ChatGPT, Grok, and all the others are nothing more than next-word predictors, like when you start to type into a search engine. Well, ok, the companies promoting them have 'trained' them by feeding terabytes of internet text so the machine does more than find the next word, it suggests some taradiddle boiler-plated from text matching the context. I called it boiler-plated, they call it generated; semantics.
But just as there's more to human intelligence than typing rubbish on a computer keyboard, there is much more to AI than chatbots. They've used AI for ages in the stock-markets to anticipate price movements. With some pretty worrying consequences such as accelerating & exaggerating those movements.
More usefully, AI has been assisting in medical fields including scanning tissue samples to 'find' cancerous cells. Or developing highly complex molecules for use in medicines.
Also, developing microchips for specific processing tasks - that one is a bit worrying because the humans using the AI can't see how these AI developed chips actually work, though testing shows they do work.
Today on breakfast radio, Sir Brian Leveson was talking about his inquiry into the UK criminal justice system which has become sclerotic with not enough judges and magistrates and growing backlogs of court cases. One very large factor he identified in cases taking a long time to come to court and longer to process through the courts is the vast amount of data from phones and computers which has to be sifted and analysed to see if there is anything pertinent to the case. I immediately thought, prime area to employ an expert system which could process thousands or even millions of emails and text messages to sort the wheat from the chaff. Would potentially save a huge amount of time for both defence and prosecution lawyers.
Seems to me the 1990's was that last distinct decade. It's all been pretty homogeneous since.
Perhaps the simulation is already in place and AI is already running the show.
How will/would we know if/when it happens?
Perhaps the simulation is already in place and AI is already running the show.
How will/would we know if/when it happens?
There's a statistical argument which goes something like, given a very large universe, somewhere, sometime, it is inevitable some civilisation has evolved with the technology to create a Matrix-like simulation. And as soon as that happens, the number of simulated realities and nth-level simulated realities within those soon explodes to the point we are OVERWHELMINGLY MORE LIKELY to be leading our sorry lives as simulations inside somebody else's AI than in a real gen-u-ine reality.
To the extent that our unknown-to-us creators take any interest at all in us, the joke is When will they wake up and discover they don't exist?
Perhaps it's best we never wake and go on playing with old motorbikes.
Gary, I think you're still stuck on chatbots and generating more rubbish but what I'm pointing out is the potential to use computing power as a tool, using it's capacity to review lots of data and find connections at fantastic speeds, in medicine, research, weather forecasting, and I'm suggesting criminal justice. There surely is an upside to AI?
Oh, bugger, for a moment I believed that until I read my own last post.
Oh, bugger, for a moment I believed that until I read my own last post.
So the bias is even more intimately ingrained than I thought
That’s where I’m working, as applied to aircraft maintenance. We’re trying to predict when the airplane will break. First, it has to learn to read the data. Then, some engineers have to learn to analyze it. It’s an evolutionary process.
Some interesting quotes here. Things to think about.
https://www.axios.com/2025/07/09/ai-rapid-change-work-school?utm_source=firefox-newtab-en-us
https://www.axios.com/2025/07/09/ai-rapid-change-work-school?utm_source=firefox-newtab-en-us
Future is scary, but then I think it probably always has been.
Things always change. Doors close and new doors open. Just have to wonder if enough new doors are going to open.
Kind of plugging the hole in the dyke with my finger, but I've taken to blocking AI intrusions onto my web pages, via the block function of uBlock and others. I can manage perfectly well without the drivel, thank you very much.
Good'a place as any for this....
Denmark is fighting back
https://www.theguardian.com/technol...denmark-copyright-law-artificial-intelligence
https://www.theguardian.com/technol...denmark-copyright-law-artificial-intelligence
Watch this podcast, it'll take a while before you see why I think you should watch.
Thought provoking.
I think she's lonely and needs a male voice to accompany her "desk rocket"
As an after thought, I've listened to many motorcycle related You Tube videos, reviews etc. on bikes that have the same semi real voice and makes the same type errors so it's obviously being used more frequently than we generally realize.
Largely why I've deleted my personalized phone message and gone to the company provided generic, obviously computer generated message.
Reason being that in a typical phone message you use your natural speaking voice, you provide your name and the typical message like
"This is (your name here), Sorry I can't take your call right now, please leave your name, number and a message and I'll return your call as soon as possible. Have a nice Day"
From that much information any geek can replicate your voice utilizing AI, and make it believable even to one of your family members
As an after thought, I've listened to many motorcycle related You Tube videos, reviews etc. on bikes that have the same semi real voice and makes the same type errors so it's obviously being used more frequently than we generally realize.
Largely why I've deleted my personalized phone message and gone to the company provided generic, obviously computer generated message.
Reason being that in a typical phone message you use your natural speaking voice, you provide your name and the typical message like
"This is (your name here), Sorry I can't take your call right now, please leave your name, number and a message and I'll return your call as soon as possible. Have a nice Day"
From that much information any geek can replicate your voice utilizing AI, and make it believable even to one of your family members
Last edited:
I've heard numerous AIs stumble over Roman numerals - sometimes quite funny.
That was incredibly realistic, I’m kinda shocked actually. Right at the end of the video she said a couple of things, she was actually open to AI as a useful educational tool. And I myself have been surprised on numerous occasions when I look something up and get an immediate AI generated response that is both accurate and thorough.
The comments she made were ( and I might be paraphrasing here ) that while AI can be a useful tool, she believe we must be careful not to diminish the human experience or lose our critical thinking skills. And I think that observation is pretty on point.
One of my big fears with AI is that people will get so lazy about thinking for themselves that they will just believe and repeat as fact, whatever their phone or computer spits out.
The other big fear I have is the loss of employment opportunities this will all cause. Particularly technical fields will be vulnerable, doctors , teachers, engineers, office workers, anyone who processes and enters data and issues reports, and on and on. The only jobs that I can see that are safe in the near term are manual labor jobs, construction, plumbers, electricians , mechanics and such.
I guess you'll now want to let your babies grow to be Cowboy's, rather than Doctor's and Lawyer's and such