Modern AI Tech

Bart_

Carpal tunnel level member
Location
GTA
Remember the good old days when telephone annoyance was - [ick up a call, say hello, wait and hear silence and just when you're about to hang up a machine which dialled to you switches you to a scripted (scam/telemarketer/survey) operator who wasted your time? After enough calls I developed the response of of saying hello repeatedly with similar length pauses a few times pretending I couldn't hear the telescripter and then hanging up.

This kicks it up to a whole new level ;)



edit - the url auto jumped to where I paused the video. Hope it now starts at the beginning
 
Last edited:
If any y'all have CS training you'll recognise the recursion being at least one part of the AI's demise. comma comma period period period :)

Has anyone come to recognise AI slop on youtube, common voices techniques etc?
 
If any y'all have CS training you'll recognise the recursion being at least one part of the AI's demise. comma comma period period period :)

Has anyone come to recognise AI slop on youtube, common voices techniques etc?
I can’t quite put words around it, but I have a keen sense of when it’s ai. Sometimes it’s the way a single word or phrase is said, and sometimes it's just a general vibe, but I don't think it's good enough to fool anyone who is paying enough attention and thinking critically.
 
More often than IMO it should occur, I'll see tech or history vids where mismatched media are used. They'll be talking about one type of engine eg small single and then put in a photo of a v8 head and then you recognise the inflection and tone, the exact same as you heard in another vid (and word pacing). Media can also seem too diverse, like it's probably been scraped from the web. Sounds very human except for those clues. Grammar techniques and writing construction patterns that harken back to high school english class are another give away. No one writes like that.

comma comma period period

You could strap your old high school English teacher in a chair like Clockwork Orange and torture via the mangled AI meltdown just like Sir Paul's famous "in which we live in" line. (from Live and Let Die)
 
More often than IMO it should occur, I'll see tech or history vids where mismatched media are used. They'll be talking about one type of engine eg small single and then put in a photo of a v8 head and then you recognise the inflection and tone, the exact same as you heard in another vid (and word pacing). Media can also seem too diverse, like it's probably been scraped from the web. Sounds very human except for those clues. Grammar techniques and writing construction patterns that harken back to high school english class are another give away. No one writes like that.

comma comma period period

You could strap your old high school English teacher in a chair like Clockwork Orange and torture via the mangled AI meltdown just like Sir Paul's famous "in which we live in" line. (from Live and Let Die)

I think Sir Paul's line is "ever-changing world in which we're living", but I could be wrong....
 
I saw an interview where Paul said it was a small jab at his school teacher who gave him a hard time. Could be fabbed up by Paul later on for PR/interview but it seems in character with his mantra "Are you going to tell me I'm not allowed to do that? I'm allowed." One time he was asked who's your favourite bass player? and he said "me. I'm allowed to say that, aren't I" gently challenging the interviewer to correct him. That was late Beatles or Wings era. Even now at 82 the guy's still got presence.

derail sorry
 
AI strikes again with robotics - non-nefarious. If you're interested look up Raibert of Boston Dynamics who was a hero of mine back in the day.

Now for Benny Hill :)


There's hours and hours of raw footage, some impressive.
 
Crash and burn for the win, 100 m sprint under 9 seconds!


Here's an un-censored, not state-run video showing the sacrificial end:

 
Last edited:
Crash and burn for the win, 100 m sprint under 9 seconds!


Here's an un-censored, not state-run video showing the sacrificial end:

I heard about that, that's crazy! Just wait until robo-soldiers can do that kind of time through the woods, hopping logs and fences. I'm sure that's coming and not as far off as we might like.
 
If you watch the opening ceremonies it looks like the marching in the movie I Robot. The kick boxing shows what I consider to be lethal blows. They also run a largely boot camp style obstacle course event.

One technology step that could bleed into battery saws is FOC field oriented control which is a step up on hall effect fixed brushless commutation. Ebikes have been onto it for a while already.
 
I see it getting into my work which I never thought would happen. People’s job requests include AI wording, or their conclusions about their trees are informed by AI inquiries. I’ve had to correct people’s information about trees because they already asked an AI chatbot and were getting bad information. Unexpected… didn’t think the work of a consulting arborist would be so easily ‘replaced’ but there you go.

People certainly are, on the whole, exceedingly dumber by the day.
 
What gets me is AI writing software and just recently I discovered designing a circuit and PCB board. At first I marvelled at it but then when pondering how I came up with a principle. "It's been done before" or close enough to copy/tweak. Still quite an achievement. Professionals consider such code/design as a first draft to examine, tweak, test etc before standing behind it and putting it out there. The cct example had cpu, peripherals, sensors, signal chips, many pins involved, the whole shooting match. In fairness anything I design is "copied" via either someone else's example or chip application notes which give example cct snippets based on testing they've done on successful configurations and to which they've optimised the chip internals to. On data sheets they sometimes give the schematic of the chip internals and you've got to be hard core to glean insight from that. Eg an op amp 3 pins you use and perhaps a hundred internal transistors/diodes/etc.

Don't they call the bad info "hallucinations" when it propagates to generated output?
 
Was digging into a network connectivity problem and youtube big brother served this up to me. Skynet here we come!

Wow, very sobering. It seems like only a matter of time before our infrastructure of keeping track of money transactions in the globally connected banking systems will be compromised.
 
That's kind of the gist of what those guard-rails folks are all about. Some time back an anecdote passed my way saying the team had to use AI to sleuth out what happened and they had to run a Chinese AI without safety guards to do it because the safetied AI they had refused on the grounds that it felt it was being given a mal-acting task.

Watched some more related vids and the skynet moment is when the AI making better AI loop goes out of control, despite robocop prime directives etc. Moral choices can devolve to rational balances it seems in AI, demonstrated a bit in the breakout. In fairness some of he lingo must have infused from hacker based training data perhaps message boards. But the evolving communication and "greater good" stuff was spooky. IMO A connection from computers to real life hardware like in scifi movies is getting simpler every day too. doh.

A huge risk is mal-actors letting loose with an AI attack on purpose. Exponential timeline on AI capability progression!

Friend of mine was a phD in AI back in the days of poor % success rate trying to recognise images, simple back prop nets, before the eureka structural change/training which is all currently black magic to me. Actually another semi-classmate of mine also phD'ed in AI and did musko-skeletal mobility, for a bug model, a lot like those crash and burn for the team's glory running robots. Wonder if he's retiring soon. Probably wants to keep his fingers in the game at this dramatic nexus.


If you want further spooking look up (LG) smart tv's spying on you "We own the glass"! Gamer's Nexus guy.
 
That's kind of the gist of what those guard-rails folks are all about. Some time back an anecdote passed my way saying the team had to use AI to sleuth out what happened and they had to run a Chinese AI without safety guards to do it because the safetied AI they had refused on the grounds that it felt it was being given a mal-acting task.

Watched some more related vids and the skynet moment is when the AI making better AI loop goes out of control, despite robocop prime directives etc. Moral choices can devolve to rational balances it seems in AI, demonstrated a bit in the breakout. In fairness some of he lingo must have infused from hacker based training data perhaps message boards. But the evolving communication and "greater good" stuff was spooky. IMO A connection from computers to real life hardware like in scifi movies is getting simpler every day too. doh.

A huge risk is mal-actors letting loose with an AI attack on purpose. Exponential timeline on AI capability progression!

Friend of mine was a phD in AI back in the days of poor % success rate trying to recognise images, simple back prop nets, before the eureka structural change/training which is all currently black magic to me. Actually another semi-classmate of mine also phD'ed in AI and did musko-skeletal mobility, for a bug model, a lot like those crash and burn for the team's glory running robots. Wonder if he's retiring soon. Probably wants to keep his fingers in the game at this dramatic nexus.


If you want further spooking look up (LG) smart tv's spying on you "We own the glass"! Gamer's Nexus guy.
The technology is crazy today. I have an uncle who is a conspiracy theorist and author (and a government cybersecurity programmer...) who gave me the book Nano (I believe by John Marlow) about 20 years ago and it was a very thought provoking read. Highly recommend it, it will definitely make you think about what AI technology gone awry could potentially do.
 

New threads New posts

Back
Top Bottom