Jump to content

Why I Do Not Fear AI...


Recommended Posts

1 hour ago, kerbiloid said:

It's a value of uncontrollable emission. Thus, it's a covert action against the financial control.

***

On another topic:

https://www.theguardian.com/us-news/2023/jul/13/hollywood-actors-union-recommends-strike-as-talks-deadline-passes

It's your turn, AI!

Yes but that you use your electricity for is not regulated at least if you don't have a 10.000 V line into your compound. 

Now I think its plenty of cheating / stealing of power here, like IT staff mining in the server room for their own profit.
Could easy see an shell company renting an place and not paying the power bill then clear out and declare bankruptcy once the power it cut as they don't pay the high bills. 

As for writers I don't see them getting replaced by AI, for artists, not for the big ones who are names, yes you have settings there the character is the famous one but multiple has been cast as Laura Croft, you just need an actor who look enough like the character and few care.  As for background characters they have been cgi for a long time if you need large crowds. 

Link to comment
Share on other sites

41 minutes ago, magnemoe said:

Yes but that you use your electricity for is not regulated at least if you don't have a 10.000 V line into your compound. 

The elctricity is paid on the established price.

But the bitcoins have the value themselves, which can be converted into official currency.

The currency rate is established by dividing the overall amount of produced value by overall amount emitted currency.

Thus, an uncontrollable production of value is like uncontrollable production of emitted amount of currency, exactly what the counterfeiting is, just from the opposite end.

Thus, the mining is counterfeiting + overconsumption of electric power.

But it's not prohibited, thus it's probably important for some entity beyond the national economies.

AI is under suspicion.

(It can't be a national government, as the mining is not geographically controlled.)

The obvious and easy countermeasure could be a hard limit of household consumption, but it's not limited. So...

Edited by kerbiloid
Link to comment
Share on other sites

22 hours ago, kerbiloid said:

The elctricity is paid on the established price.

But the bitcoins have the value themselves, which can be converted into official currency.

The currency rate is established by dividing the overall amount of produced value by overall amount emitted currency.

Thus, an uncontrollable production of value is like uncontrollable production of emitted amount of currency, exactly what the counterfeiting is, just from the opposite end.

Thus, the mining is counterfeiting + overconsumption of electric power.

But it's not prohibited, thus it's probably important for some entity beyond the national economies.

AI is under suspicion.

(It can't be a national government, as the mining is not geographically controlled.)

The obvious and easy countermeasure could be a hard limit of household consumption, but it's not limited. So...

Its not counterfeiting more than producing stuff people pay lots of money for even if have no practical use, NFT is kind of an parody of it.

Iit predates AI by a decade, Bitcoin started in 2009. 
Its simply not regulated as to hard to do and politicians don't understand it. Now its an field full of scams and I assume plenty of whitewashing of money. 
But that is more an interest for the legal system than politicians. 

Link to comment
Share on other sites

So we went to the homeschool convention down in Phoenix this weekend. And we saw the Waymo self-driving cars driving around downtown all day.

iy6YIDS.jpg

Some of them had folks behind the wheel, some of them didn't. At one point we're walking down the street and Thing #3 asks, "So, why are they just driving around?" I said, "Well, probably because they're training their algorithm. It's machine learning, this is how the machines learn." She asks, "Well, yeah, but why can't they just learn in a simulation or something? Why do they have to drive around the streets?" I said, "Well, if you look in the documentation, it will say something like, 'Learning in real world conditions'. But the reality is, it needs to learn to account for stupidity. The sheer depth and breadth of human stupidity. Which, to my knowledge, nobody has been able to accurately simulate yet." Mrs. TheSaint just chortled.

Link to comment
Share on other sites

22 hours ago, SunlitZelkova said:

Nuclear weapons are new and yet they are not fine.

They and not newer than houses or cars but newer than most technology in you house who is not an radio or an hand tool (in common use. TV and electrical drills is older but was not common household items). 
 

Link to comment
Share on other sites

12 hours ago, kerbiloid said:

They do. It's the Matrix. We live in it, they train in it.

They have virtual testing, this is very nice for rare events and stuff you don't want to test live and for an very controlled environment for documentation.
Pretty sure they does an very high millions on hours there. And you can feed the AI real data and analyze the result with another AI.  
An issue with neural networks is that you can not document its behavior unless you test it. 
How does it react if you wear an traffic cone costume?  Or someone falling in an pedestrian crossing? And wearing an white outfit in winter? 
 glastonbury-uk-24th-june-2017-festival-g 
You can fail safe an conversational program if you keep it simple and document it, then have the next layer use that interface. Usually another computer but you can manual override. 
You can not do that with current AI and its probably harder with smarter ones. 
And stuff like planes or life support  medical equipment does not have an fail safe modes then in use. 
Cars do but blocking highways has an cost, so even if AI run cars is safer than 90% of drivers / km driven, they might cause so much gridlocks because the scary shadows then sun get low and the plastic bag ghosts they are blocked. 
 

Link to comment
Share on other sites

11 hours ago, magnemoe said:

They and not newer than houses or cars but newer than most technology in you house who is not an radio or an hand tool (in common use. TV and electrical drills is older but was not common household items). 
 

I’m not saying new things can’t be good, I’m just saying being new doesn’t automatically make it good, and new things can be bad.

Link to comment
Share on other sites

8 hours ago, SunlitZelkova said:

I’m not saying new things can’t be good, I’m just saying being new doesn’t automatically make it good, and new things can be bad.

Here I agree, many modern trends in games as an example. Live services must be 30 years old as in MMO, who worked well for them, not then most games tries to be live services. 
Moving controls in cars to flat screens is bad, as you have to watch you hand not the road them operating them.  Buttons you can find by touch and for more advanced stuff you can show selections in the front screen with speedometer and other essentials. Yes for stuff like navigation you pretty much need an touch screen. Same for more complex settings but this is addition not replacement. 

Link to comment
Share on other sites

24 minutes ago, magnemoe said:

Here I agree, many modern trends in games as an example. Live services must be 30 years old as in MMO, who worked well for them, not then most games tries to be live services. 
Moving controls in cars to flat screens is bad, as you have to watch you hand not the road them operating them.  Buttons you can find by touch and for more advanced stuff you can show selections in the front screen with speedometer and other essentials. Yes for stuff like navigation you pretty much need an touch screen. Same for more complex settings but this is addition not replacement. 

Ideally, a touchscreen with a braille-like display that allowed encoding button-like textures.  And pressing would require a lot more force than mere touching

Edited by darthgently
Link to comment
Share on other sites

Thread of image to video examples. This is pretty new stuff, so sorta like looking at midjourney from this time last year. generating content will get easier and better.

Link to comment
Share on other sites

2 minutes ago, tater said:

It's 2023, we play rock, paper, earbuds here now.

Then I don't ask, how it should distinguish nail (a sharp iron stick) and nail (a finger cover).

P.S.
Just noticed, that it's very dangerous to talk to AI in English,

"Open the chest."

Link to comment
Share on other sites

54 minutes ago, kerbiloid said:

Then I don't ask, how it should distinguish nail (a sharp iron stick) and nail (a finger cover).

P.S.
Just noticed, that it's very dangerous to talk to AI in English,

"Open the chest."

I say you probably could use an larger set of scissors to drive in an nail, might work better than the rock if you held the tip as the handle would add mass. 
So it made it simpler for the AI. 

Link to comment
Share on other sites

I just tried auto-driving controlled by ai through the radars and other sensors on my dad's new car this morning to work. Have to say it feels really weird that "someone" is twisting your steering wheel. And then because it's based entirely on its own measurements, makes lane changes, and other maneuvers are carried out quite decisively - even a bit scary.

In my father's words, it's like a 20-year veteran driver being late for work.

Edited by steve9728
Link to comment
Share on other sites

16 hours ago, kerbiloid said:

Those, who had read the HP fanfic by Yudkowski ("Methods of Rationality"), know that there is also the Hufflepuff option for such case.

I've read bad fanfic but it was a rare displeasure to find one so thoroughly saturated with the author's smug sense of superiority.

Edited by KSK
Link to comment
Share on other sites

Guest
This topic is now closed to further replies.
×
×
  • Create New...