Re: I don't think ai is a threat in the way you think it's going to be
Posted: Tue Apr 04, 2023 11:55 am
by Isaac
Tunnelcat wrote: ↑Tue Apr 04, 2023 11:06 am
Uncle.
noooooo. I'm sorry
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Tue Apr 04, 2023 12:26 pm
by Tunnelcat
Isaac, vision saw if first, but I think I'm being bested by an Isaac bot. You've never been that verbose the whole time you've posted here.You're typically snarky, funny and borderline nutty. In other words, human. And that's the tell. BOT's are verbose and detailed, making few mistakes. If it's you, it's out of character, unless it's a topic that genuinely interests you and if so, it's out of my league. Whenever bots start showing emotion and the corresponding mistakes that humans make, THEN it's time to worry. By the way, I bombed at Fortran in college, so there's no way in hell I can hold a coherent conversation about coding and machine learning. But make no mistake, the only thing stopping a machine from achieving the equivalent of sentience is pure computational power. When we get functioning quantum computers, then we're so screwed.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Tue Apr 04, 2023 12:28 pm
by Isaac
But that was my own writing! I am into machine learning and I think if you gave my post a chance it would describe machine learning in a simple way.
Edit: maybe it's worth mentioning that my job involves data science.
Edit:
Also I'm not trying to best you in any way. I know understanding how a language model works like ChatGPT is complicated, however if you read the last part of my post you'll see that the fundamentals are pretty straightforward. Again I recommend giving my post on page 1 a chance.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Tue Apr 04, 2023 7:39 pm
by Tunnelcat
No, I don't think you're trying to best me. I'll tell you what. I've been a little bit under the weather for a bit and your post has become a little overwhelming to sift through and think about. Let it sit for a bit and I get back to it when I can think straight and make an analysis. Migraines have a way of messing up my focus, and my day. By the way, I suck at math too.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Tue Apr 04, 2023 8:29 pm
by Isaac
Tunnelcat wrote: ↑Tue Apr 04, 2023 7:39 pm
No, I don't think you're trying to best me. I'll tell you what. I've been a little bit under the weather for a bit and your post has become a little overwhelming to sift through and think about. Let it sit for a bit and I get back to it when I can think straight and make an analysis. Migraines have a way of messing up my focus, and my day. By the way, I suck at math too.
That doesn't sound like fun. I had allergies last week so I was drinking tea almost every afternoon.
The math isn't really "math" math. It' just what side of the line a dot could fall on.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Fri Apr 14, 2023 2:30 am
by Burlyman
Amazon should make Alexa into a true answering machine. i.e. not voice mail.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Sat Apr 15, 2023 12:08 pm
by Isaac
Burlyman wrote: ↑Fri Apr 14, 2023 2:30 am
Amazon should make Alexa into a true answering machine. i.e. not voice mail.
I'd imagine that next year (2024) it would be able to hold a conversation with its user. It doesn't have to be as good as a human, but just able to ask questions and keep track of what the user wants out of the conversation and try to fulfill each goal. Kids will end up replacing the need for a friend with it. Parents who can't do basic math will gladly let it help their kids do their homework.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Sat Apr 15, 2023 2:15 pm
by Tunnelcat
Aaaand, we're gonna have a whole generation of stupid kids who can't think for themselves, crave instant gratification and have zero social skills. Hey, maybe kids could let their AI's bully each other instead of the other kids. I wonder how that would turn out?
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Sat Apr 15, 2023 2:57 pm
by Spidey
Tunnelcat wrote: ↑Sat Apr 15, 2023 2:15 pm
Aaaand, we're gonna have a whole generation of stupid kids who can't think for themselves, crave instant gratification and have zero social skills.
Must resist.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Sat Apr 15, 2023 3:20 pm
by Krom
Yeah, they're called boomers.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Sat Apr 15, 2023 4:00 pm
by Isaac
Krom wrote: ↑Sat Apr 15, 2023 3:20 pm
Yeah, they're called boomers.
I really don't see that.
My millennial sister in-law occasionally calls my wife to ask us for help on her son's homework.... For problems that involve fractions and multiply percentages with units. I mean she graduated college so I don't understand why she's even calling us for help. I think most boomers have a basic mental math capabilities.
FYI I've shown her chatgpt and Wolfram alpha so she'd stop calling us for this basic stuff be she claims using those sites are too difficult.... My mom is a boomer and she's already used chatgpt to solve a few different problems she was having with her laptop, treating it like customer support, which is awesome.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Sat Apr 15, 2023 4:51 pm
by Top Gun
Isaac wrote: ↑Sat Apr 15, 2023 4:00 pm
I think most boomers have a basic mental math capabilities.
You've clearly never worked retail.
Re: I don't think ai is a threat in the way you think it's going to be
Tunnelcat wrote: ↑Sat Apr 15, 2023 2:15 pm
Aaaand, we're gonna have a whole generation of stupid kids who can't think for themselves, crave instant gratification and have zero social skills.
Must resist.
Better cough up what generation you're from Spidey before you skewer us boomers. I got into the job market right after college and didn't sit on my ass griping about how hard work is. My Gen X brother was a lazy bum, didn't complete high school and was generally a bad seed who never applied himself or held a steady job until he hit the age of 50. Even now he's kind of rudderless in life, other than convincing his dad that he deserves more of dad's estate than his 2 sisters.
Re: I don't think ai is a threat in the way you think it's going to be
Isaac wrote: ↑Sat Apr 15, 2023 4:00 pm
I think most boomers have a basic mental math capabilities.
You've clearly never worked retail.
Come on TG. Most Boomers are now in old age and losing their mental faculties. Just remember, we all get old and you're headed there too some day in the future. By the way. I can count change.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Sat Apr 15, 2023 5:46 pm
by Top Gun
I mean I don't know about you but I plan on going all Ghost in the Shell and uploading my consciousness to the 'net if someone figures that stuff out.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Sat Apr 15, 2023 6:26 pm
by Isaac
Top Gun wrote: ↑Sat Apr 15, 2023 5:46 pmMe someday:
If you can develop an AGI ML model that's basically you how do we transfer our consciousness to it?
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Sat Apr 15, 2023 6:47 pm
by Top Gun
Jameson Guy was the best.
Re: I don't think ai is a threat in the way you think it's going to be
Tunnelcat wrote: ↑Sat Apr 15, 2023 2:15 pm
Aaaand, we're gonna have a whole generation of stupid kids who can't think for themselves, crave instant gratification and have zero social skills.
Must resist.
Better cough up what generation you're from Spidey before you skewer us boomers. I got into the job market right after college and didn't sit on my ass griping about how hard work is. My Gen X brother was a lazy bum, didn't complete high school and was generally a bad seed who never applied himself or held a steady job until he hit the age of 50. Even now he's kind of rudderless in life, other than convincing his dad that he deserves more of dad's estate than his 2 sisters.
I didn't skewer the boomers...that was krom, I'm a late boomer myself and I started working at age 10 as a newspaper delivery boys' helper. Hell I had worked 3 or 4 part time jobs before I left high school.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Sat Apr 15, 2023 7:57 pm
by Tunnelcat
Sorry Spidey. Got the "Must resist" conflated with Krom's comment. So, after my brain fart, same question to Krom. All through history, the older generations have always called the younger generations lazier than their own, in the past as in the present, so I can't sit here and call generations after mine any more lazy or stupid than mine.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Sat Apr 15, 2023 8:28 pm
by Spidey
But it's a tradition...
Every generation says the ones after theirs is lazy and stupid, and every generation blames the ones before theirs for screwing up the world.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Sat Apr 15, 2023 9:16 pm
by Isaac
Spidey wrote: ↑Sat Apr 15, 2023 8:28 pm
But it's a tradition...
Every generation says the ones after theirs is lazy and stupid, and every generation blames the ones before theirs for screwing up the world.
Relevant quote
Memoirs of the Bloomsgrove Family, Reverend Enos Hitchcock 1790 wrote:
The free access which many young people have to romances, novels, and plays has poisoned the mind and corrupted the morals of many a promising youth; and prevented others from improving their minds in useful knowledge. Parents take care to feed their children with wholesome diet; and yet how unconcerned about the provision for the mind, whether they are furnished with salutary food, or with trash, chaff, or poison?
Re: I don't think ai is a threat in the way you think it's going to be
Spidey wrote: ↑Sat Apr 15, 2023 8:28 pm
But it's a tradition...
Every generation says the ones after theirs is lazy and stupid, and every generation blames the ones before theirs for screwing up the world.
Relevant quote
Memoirs of the Bloomsgrove Family, Reverend Enos Hitchcock 1790 wrote:
The free access which many young people have to romances, novels, and plays has poisoned the mind and corrupted the morals of many a promising youth; and prevented others from improving their minds in useful knowledge. Parents take care to feed their children with wholesome diet; and yet how unconcerned about the provision for the mind, whether they are furnished with salutary food, or with trash, chaff, or poison?
Interestingly, the Bible seems to have the right perspective on this:
Ecclesiastes 7:10 wrote:“Say not thou, What is the cause that the former days were better than these? for thou dost not enquire wisely concerning this.”
Re: I don't think ai is a threat in the way you think it's going to be
Spidey wrote: ↑Sat Apr 15, 2023 8:28 pm
But it's a tradition...
Every generation says the ones after theirs is lazy and stupid, and every generation blames the ones before theirs for screwing up the world.
Relevant quote
Memoirs of the Bloomsgrove Family, Reverend Enos Hitchcock 1790 wrote:
The free access which many young people have to romances, novels, and plays has poisoned the mind and corrupted the morals of many a promising youth; and prevented others from improving their minds in useful knowledge. Parents take care to feed their children with wholesome diet; and yet how unconcerned about the provision for the mind, whether they are furnished with salutary food, or with trash, chaff, or poison?
Interestingly, the Bible seems to have the right perspective on this:
Ecclesiastes 7:10 wrote:“Say not thou, What is the cause that the former days were better than these? for thou dost not enquire wisely concerning this.”
Another relevant quote:
Book III of Odes, Horace circa 20 BC wrote:
Our sires' age was worse than our grandsires'. We, their sons, are more worthless than they; so in our turn we shall give the world a progeny yet more corrupt.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Tue Apr 18, 2023 2:04 pm
by Tunnelcat
In other words, we forgot what it was like to be young. Either that or we're pissed off because we're old, set in our ways, staring down death and we've got nothing better to do than ★■◆● about what's no longer ours, youth.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Tue Apr 18, 2023 9:51 pm
by Isaac
Tunnelcat wrote: ↑Tue Apr 18, 2023 2:04 pm
In other words, we forgot what it was like to be young. Either that or we're pissed off because we're old, set in our ways, staring down death and we've got nothing better to do than ★■◆● about what's no longer ours, youth.
Speaking of which, any chance you could give my reply on page 1 another chance?
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Wed Apr 19, 2023 11:50 am
by Tunnelcat
Still thinking about it.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Wed Apr 19, 2023 12:16 pm
by Isaac
Tunnelcat wrote: ↑Wed Apr 19, 2023 11:50 am
Still thinking about it.
Let me summarize it then:
Imagine you have a bag of small random toys: some are rubber ducks and the rest are other random animals .
You reach in the bag and throw one into a sand box like the kind kids play in.
Then you throw each toy at random into the box: trying to keep all the ducks on one half and the other toys on the other half. But it's not perfect. Some ducks are on the wrong side.
Next you take your finger and draw a line in the sand that divides the two kinds of toys, ignoring outliers, like the ducks on the wrong side. This line is your model! One side is a the duck side and the other is the "other animal" side.
No longer do you need to look at the toys to determine if it's a duck, because depending on what side of the line you find a toy on, your model tells you if it's a duck. But your model will make mistakes, because of the outliers, but that's good. I won't explain why that's good, because this post is already getting long.
Does reply explain the ML process a little better? It gets way more complicated from here, but don't worry about that yet.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Wed Apr 19, 2023 12:32 pm
by Tunnelcat
Ah, but you're not figuring in that the person grabbing the animals can "feel" the shape with their hand and thus know when they pick up a duck. So right from the start, there's a selection bias and less randomness.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Wed Apr 19, 2023 1:16 pm
by Isaac
Tunnelcat wrote: ↑Wed Apr 19, 2023 12:32 pm
Ah, but you're not figuring in that the person grabbing the animals can "feel" the shape with their hand and thus know when they pick up a duck. So right from the start, there's a selection bias and less randomness.
In this example you can use your eyes. The idea is one area is defined as a "duck" area. This is how a very simple model works. That part makes sense, right?
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Thu Apr 20, 2023 5:14 pm
by Burlyman
Most people aren't good at math for some reason.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Thu Apr 20, 2023 6:56 pm
by Isaac
Burlyman wrote: ↑Thu Apr 20, 2023 5:14 pm
Most people aren't good at math for some reason.
I think math isn't intuitive, but the concepts are. So if you can explain a concept then attach the syntax to those concepts, you'd be a good teacher. Sometimes, you can't escape the math to explain a concept, like calculus and derivatives.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Fri Apr 21, 2023 12:11 pm
by Tunnelcat
My husband understood calculus and derivatives really well. I guess that's why he was an electrical engineer. He could sit there and calculate a result using calculus chicken script like he was eating a sandwich. Me, nope. Math is a left brain thing and in my left brain, numbers are mush. 3D objects, now that different. I love figuring out ways to manipulate tactile 3D objects. As to the ducks, tell my why the small mistakes in the selection and separation process are a good thing?
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Fri Apr 21, 2023 6:27 pm
by Isaac
Tunnelcat wrote: ↑Fri Apr 21, 2023 12:11 pm
As to the ducks, tell my why the small mistakes in the selection and separation process are a good thing?
First look at these three graphs. Pretend the yellow circles are your ducks:
If you used all of your ducks to draw the line and forced all the ducks to be on one side this would be over fitting.
Instead it's better to train your model to get the concept of the ducks so if you were to throw a toy, your model will have a good way to predict what kind of you you'll give it next.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Sun Apr 23, 2023 7:17 pm
by Tunnelcat
So the point is? Trying to simplify the results to fit the model, or simplifying the model to fit the results? Aren't they pretty much the same?
I like to play single player games against the program's AI NPC's, not actual humans. Even in modern games, some developers do better at giving their NPC's a more human response than others. In other words, unpredictable behaviors. Most games I can figure out the NPC's behaviors and use that against them to beat them, because I start seeing their repeating patterns. They're staying rigidly within their programmed models and they all DIE. In multiplayer, you get unpredictable, aggressive or smarter human behaviors which makes the game more challenging to other players. Of course, you get cheaters which ruins the game altogether. Game program AI's need to be self-modifying and adaptive in order to truly compete against a human. In other words, they need to start looking outside the statistical model their thought process program is based on and look at the outliers, the stray ducks and why they're even there at all and if that's a new problem to look at.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Sun Apr 23, 2023 7:42 pm
by Krom
You overrate humans a little. In multiplayer games I've done plenty of ass kicking and also received plenty of ass kicking to know that people (myself included) fall into the same or even worse predictable behaviors than AI. We all have a "moveset" we develop and practice that we apply to how we fight battles in multiplayer games (In D3 I almost always circle strafed to the right when attacking for instance).
If you want to see how AI can behave in unique and unpredictable ways, might I suggest watching some of the Tesla self driving video channels on youtube (like DirtyTesla for example). The AI is improving but the better it gets the more bizarre and random the ways it occasionally fails are. Just a few videos of watching "full self driving" in action make the whole experience look incredibly stressful and I'm definitely glad I didn't pay for that package on my vehicle. Not to downplay the incredible accomplishments they have made and how far they have come (just compare recent beta videos to the early beta videos), but it seems the better it gets in general, the more scary the mistakes it does make become.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Sun Apr 23, 2023 8:24 pm
by Darth Wang
In some games, like chess, computers are better than humans and have been for almost 30 years. In other games, they are still struggling to come up with a program that can even play the game competently against a human.
Basically it depends what kind of task is being attempted.
Re: I don't think ai is a threat in the way you think it's going to be
I actually think this is a good thing. We don't need Artificial General Intelligence to be better than the best teacher: just better than the average teacher, which is probably not a high bar.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Mon Apr 24, 2023 9:48 am
by Tunnelcat
Krom wrote: ↑Sun Apr 23, 2023 7:42 pm
You overrate humans a little. In multiplayer games I've done plenty of ass kicking and also received plenty of ass kicking to know that people (myself included) fall into the same or even worse predictable behaviors than AI. We all have a "moveset" we develop and practice that we apply to how we fight battles in multiplayer games (In D3 I almost always circle strafed to the right when attacking for instance).
If you want to see how AI can behave in unique and unpredictable ways, might I suggest watching some of the Tesla self driving video channels on youtube (like DirtyTesla for example). The AI is improving but the better it gets the more bizarre and random the ways it occasionally fails are. Just a few videos of watching "full self driving" in action make the whole experience look incredibly stressful and I'm definitely glad I didn't pay for that package on my vehicle. Not to downplay the incredible accomplishments they have made and how far they have come (just compare recent beta videos to the early beta videos), but it seems the better it gets in general, the more scary the mistakes it does make become.
So, did you change your move set as other players figured out your patterns? Did you adapt new strategies? Even in a 3D movement game, most people still think in 2D patterns when they play.
As for the Tesla videos, I'll watch a few. It sounds fascinating. Most of the crashes that have occurred were because the car misconstrued emergency vehicle lights and drove right into the trucks. I'd like to know it's mistaken reasoning for doing that.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Mon Apr 24, 2023 11:43 am
by Darth Wang
Tunnelcat wrote: ↑Mon Apr 24, 2023 9:48 amEven in a 3D movement game, most people still think in 2D patterns when they play.
Re: I don't think ai is a threat in the way you think it's going to be
Posted: Mon Apr 24, 2023 12:25 pm
by Tunnelcat
I thought of that scene when I made the post. It took Mr. Logic Spock to come up with the idea. Even seasoned Kirk and superbrain Khan blew it.
noooooo. I'm sorry 
You've never been that verbose the whole time you've posted here.You're typically snarky, funny and borderline nutty. In other words, human. And that's the tell. BOT's are verbose and detailed, making few mistakes. If it's you, it's out of character, unless it's a topic that genuinely interests you and if so, it's out of my league. Whenever bots start showing emotion and the corresponding mistakes that humans make, THEN it's time to worry. By the way, I bombed at Fortran in college, so there's no way in hell I can hold a coherent conversation about coding and machine learning. But make no mistake, the only thing stopping a machine from achieving the equivalent of sentience is pure computational power. When we get functioning quantum computers, then we're so screwed.
i.e. not voice mail.
.