Artificial Intelligence and Public Comment

These are all ChatGPT, which is basically 3 years old now. GPT 4 is significantly better, but behind a waiting list. What they are developing now, that the public cannot yet see, is freaking the experts out.



Something really counterintuitive to me, is that those who "grew" this, actually do not know how it works under the hood.

 
The latest headlines have been that experts warn that AI could lead to human extinction. From the letter published yesterday...'Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.” What I haven't read is what these experts think may happen. How would AI kill us? All of this is beyond my comprehension. Does anyone know what the dangers are?
Pretty sure we were are on a good pace to do the extinction thing without help from AI. We just have to keep in ming that people get more attention if they say more shocking things. A headline that says "The future with AI...meh" doesn't get clicks. And of course politicians need people to be scared about something, so they eagerly jump aboard the train.
 
Pretty sure we were are on a good pace to do the extinction thing without help from AI. We just have to keep in ming that people get more attention if they say more shocking things. A headline that says "The future with AI...meh" doesn't get clicks. And of course politicians need people to be scared about something, so they eagerly jump aboard the train.
I would argue that just as often, politicians benefit from fostering an attitude of "meh" as much as they do, " holy shit we should probably do something."

To be clear though, I'm not really arguing for any future I think is likely, other than that of deep-fakes being indiscernible from reality and position statements written by nonhumans causing us real problems in terms of how we engage in sense-making.
 
I would argue that just as often, politicians benefit from fostering an attitude of "meh" as much as they do, " holy shit we should probably do something."

To be clear though, I'm not really arguing for any future I think is likely, other than that of deep-fakes being indiscernible from reality and position statements written by nonhumans causing us real problems in terms of how we engage in sense-making.
All to often have we have gotten the mis/disinfo tag that didn't age well. That has done much to foster the meh malaise. We either don't believe according to predisposed tribalism or we wait til another shoe falls....or did it fall? Add AI to the growing list of information shading to disambiguate.
 
All to often have we have gotten the mis/disinfo tag that didn't age well. That has done much to foster the meh malaise. We either don't believe according to predisposed tribalism or we wait til another shoe falls....or did it fall? Add AI to the growing list of information shading to disambiguate.
 
All to often have we have gotten the mis/disinfo tag that didn't age well. That has done much to foster the meh malaise. We either don't believe according to predisposed tribalism or we wait til another shoe falls....or did it fall? Add AI to the growing list of information shading to disambiguate.

I agree. It's tough for anyone to sift through the takes. Over decades, but in particular the last few years, "authorities and experts" burned through a hell of a lot of Social Trust Capital, succumbing to some weird form of audience/funding capture.

The tragedy will be that a day will come where we will really need the authorities/experts, and the boy who cried wolf won't be the only one who gets bitten.
 
I would argue that just as often, politicians benefit from fostering an attitude of "meh" as much as they do, " holy shit we should probably do something."
Have you ever listened to members of The Freedom Caucus? Nothing is 'meh' and everything is the end of civilization. Not a whole lot different than the Progressive Caucus. This is what happens when we elect politicians that have zero experience in any relevant field making laws and trying to figure out the effect of those laws. It devolves into figuring out what name gets put on the Pay-To-the-Order-Of line for US government checks. AI will be no different. Politicians have zero knowledge of the space and an equal ability to effectively regulate. AI is so new, people can claim to be an expert if they have the proper narrative for the specific cable channel.

...deep-fakes being indiscernible from reality and position statements written by nonhumans causing us real problems in terms of how we engage in sense-making.
We are already there. To be optimistic, humans are very adaptable. Maybe we will start thinking for ourselves again?

I clearly have spent too much time down this rabbit hole. One possible solution I have heard to the problem you pointed out is everything that is recorded (music, film, speeches, whatever) becomes an NFT that is verified at production as being authentic. That said, can you fake an NFT?
 
We are already there. To be optimistic, humans are very adaptable. Maybe we will start thinking for ourselves again?

I wonder if people will start to distance themselves from technology and social media? I’ve been thinking a lot lately about trying to get away from Facebook. Maybe get rid of my smartphone. They are both such a waste of time. I need to read more books, spend more time outside, exercise more, and countless other more productive things. I say all of this as I’m typing from my iPhone.
 
I wonder if people will start to distance themselves from technology and social media? I’ve been thinking a lot lately about trying to get away from Facebook. Maybe get rid of my smartphone. They are both such a waste of time. I need to read more books, spend more time outside, exercise more, and countless other more productive things. I say all of this as I’m typing from my iPhone.
Maybe. But social media is popular because it gives everyone a voice and audience.

Is maybe the film War Games a more accurate conclusion? A autonomous AI will certainly discover that people make selfish decisions often hurting both other humans and the planet, but extinction is mutually assured destruction.

1685573742899.jpeg
 
I agree. It's tough for anyone to sift through the takes. Over decades, but in particular the last few years, "authorities and experts" burned through a hell of a lot of Social Trust Capital, succumbing to some weird form of audience/funding capture.
But it has to be this way in a capitalistic market where you need to out compete your competition in garnishing or maintaining more viewers.
Maybe. But social media is popular because it gives everyone a voice and audience.
But how different is that than typical talk radio? Not everyone listens to it because it's mostly garbage and most people recognize that. There's some hope.
 
But how different is that than typical talk radio? Not everyone listens to it because it's mostly garbage and most people recognize that.
Didn’t mean to ignore this, just wanted to think about it. This is what I would guess. Most radio shows have producers that screen calls. The bat-shit crazy never gets on air. This isn’t true in the internet. On the internet/social media, the levels and nuances of crazy self-stratify into “like” groups. HT is no different. We all may disagree on topics, but when something truly crazy gets posted it gets called out. No one is immune. But HT is tame. You can go to “dark web” and see some truly crazy stuff. We just hope they are and remain the minority.

You say most people don’t listen to talk radio, but I’m not sure that is true (ok, Most is true, but plenty still do). With hundreds of stations on satellite radio, it doesn’t take someone long to self select the station that fits their narrative. In that instance, recognizing the BS is a lost skill.

I still remain hopeful there is a “middle” or “independent”, politically, socially, or fiscally. But it is hard to defend that space in the current environment.
 
Something really counterintuitive to me, is that those who "grew" this, actually do not know how it works under the hood.

I am reading a book called “The Alignment Problem” by Brian Christian that explores this issue in depth. It is very well written. The frequency with which AI programs go off the rails is startling and because they are not “ interpretable” it can be very difficult to understand why the programs failed to perform the way they were intended. The author discusses dozens of real world examples. I’d seen the book on several “best AI book” lists and would recommend it if you are interested in technical info.
 

Fascinating and insightful podcast. It’s long but breaks down a lot of issues. Overall they do a good job not diving into the tech details too much but there is some.
 
Back
Top