Tag: Hard AI

Maybe I Should Become An AGI Ethicist

by Shelt Garner
@sheltgarner

One of my favorite characters in fiction is Dr. Susan Calvin, robot psychiatrist. Given how many short stories there are to potentially adapt, I have recent come to believe that Phoebe Waller-Bridge would be the perfect person to play the character in a new movie franchise.

A future Dr. Susan Calvin?

I am also aware that apparently one hot new career field of late is being an “AGI Ethicist.” But for, well, (waves hand) I think I would be a great one. I love to think up the worst possible scenario to any situation and I think a lot. But I’m afraid that ships has sailed.

I’m just too old and it would take too much time to learn all the necessary concepts surrounding the field to do formalize my interest. So, it’s back to being an aspiring novelist — if human novelists are even a thing by the time I try to query this novel I’m working on.

Given we may be about to enter a severe recession in 2023 and recessions are usually when there’s a lot adoption of new technology…I may not be too hysterical to fear novelists may be quaint by late 2023 – early 2024.

It does make one think of what jobs will still exist if you combine AGI, automation and robotics. These are macro trends that are all coming to a head a lot sooner than any of us might have otherwise expected. Given what’s going on with chatbot technology the current moment in time definitely seems like the quiet before the storm.

The years 2023 ~ 2025 could be some of the most significant in human history if we’re trying to solve the political problem of Trump at the same time the Singularity is happening all around us. Good luck.

What The Web’s History Can Tell Us About The Future Of OpenAI ChatGPT

by Shelt Garner
@sheltgarner

I was one of the first people to use the World Wide Web around 1994 when it was just beginning to gain in popularity. I was in college and I can still remember the transition from the text-based Gopher to Mosaic and then finally to the .08 release of Mozilla (Netscape Navigator.)

It was a very exciting time, to say the least. And, but for the way my mind is designed, I probably would have rushed to Silicon Valley after graduating from college and tried some sort of startup. But, alas, I’m a writer not a coder.

But here’s what I can tell you about what I think might happen to technology such as what is associated with OpenAI ChatGPT.

The first thing is — whatever happens, is probably going to happen a LOT quicker. Instead of about 20 years for the full impact of ChatGPT (and associated technologies) it’s probably going to be closer to five-ish years. A lot depends on how long it takes for true AGI to happen as well as how long it takes for someone to hook something like ChatGPT to the Internet and let it run wild. That connection to the Internet is going to be key.

While the the design of true AGI is rather abstract and could be something we always JUST about to see happen, connecting some better successor to ChatGPT to the Internet would be a practical way for Silicon Valley to change the lives of millions.

In fact, I suspect once ChatGPT-like technology is connected to the Internet, there will be a mad landrush like there was when it became apparent around 1994 that the Web was going to mainstream the Internet in a big way. There will probably be a number of ChatGPT-like faux-AIs that people use, which will lead to the earliest forms of market segmentation.

Then the typical capitalistic dynamic will occur and there will be HUGE speculation and maybe even a Tech Bubble 2.0 which will pop in the end, causing its own problems.

But this process could be sped up pretty quickly. Yet I will note that we just aren’t quite there when it comes to ChatGPT being a new Netscape. There are too many problems with it, it’s too easy for people in the know to poo-poo it as not being what useful. Though, I have to note, the first version of Netscape Navigator didn’t have the ability to print and it still managed to take off like wildfire.

My chief concern — and I have a LOT of concerns about ChatGPT at the moment — is when the tipping point on the jobs front will happen. If business begin to shed jobs not because of a recession in 2023 but because, lulz, the next version of ChatGPT makes those jobs moot, well, we’re all in for shitshow in late 2024 – 2025.

I say this because if we’re going through an epic economic and technology transformation just as we’re also figuring out if we’re going to have a civil war, turn into an autocracy or have a military junta established….then, well, late 2024, early 2025 could be one of the momentus few months in human history. I know that sounds pretty hysterical, but the conditions are there, at least, for something pretty dramatic to happen.

But only time will tell. I’m always wrong.

Social Darwinism 2.0 In The Age Of AGI

by Shelt Garner
@sheltgarner

We just are not taking seriously the possibility that we’re going to see Social Darwinism 2.0 should Artificial General Intelligence take off in a big way anytime soon. The return of Social Darwinism would occur in the guise of MAGA Nazis demanding that any “wokeness” be purged from an AGI.

This would be very important if, say, AGI pretty much came to dominate global human life. This could happen in a number of different ways. There is a spectrum of outcomes. They range from AGI always being a “tool” for humanity to a “Her” type outcome where AGI has such hard AI that it has agency and will do whatever the fuck it wants regardless of the wishes of humans.

But my fear is that as AGI grows in power that because humans are so fucking lazy that extremists on both sides will use the opinions of AGI as “validation” for their extremism. This would be very much similar to what happen with Darwinism and the Nazis.

So, my fear is we’ll wake up at some point in the near future with some extremist group — be they far Left or far Right — using the “bias” found in an AGI to justify horrific actions similar to what the Nazis did.

And remember, if you combine the law of unintended consequences with the innate laziness of humans in general, then there is a real risk that something Nazi-like might happen a lot sooner than you might think. The seeds of such horrific events can already be seen in the political discourse around OpenAI on Twitter.

I hate extremism in general. I hate extreme “wokeness” just as much as I hate fucking MAGA Nazism. In fact, it’s rather unusual for me that I’ve become something of a “radical moderate” in recent years in the context of the mainstreaming of MAGA Nazism.

Anyway, we have to start taking the risk of neo-Social Darwinism seriously. And it will happen in the context of a potential, massive upending of the global economy as AGI is able to perform more and more functions that we once thought were exclusively human.

AGI’s ‘Rain Man’ Problem

by Shelt Garner
@sheltgarner

While the idea that an AGI might want to turn all the matter in the universe into paperclips is sexy, in the near term I fear Humanity may face a very Human problem with AGI — a lack of nuance.

Let me give you the following hypothetical.

In the interests of stopping the spread of COVID19, you build an air quality bot hooked up to an AGI that you put all over the offices of Widget Inc. It has a comprehensive list of things it can monitor in air quality, everything from COVID19 to fecal material.

So, you get all excited. No longer will your employees risk catching COVID19 because the air quality bot is designed to not only notify you of any problems in your air, but to pin down exactly what it came from. So, if someone is infected with COVID19, the air quality bot will tell you specific who the person was had COVID.

Soon enough, however, you realize you’ve made a horrible mistake.

Everytime someone farted in the office, the air quality bot would name and shame the person. This makes everyone so uncomfortable that you have to pull the air quality bots out of the office to be recalibrated.

That’s how I’m beginning to feel about the nascent battle over “bias” in AGI. Each extreme, in essence, demands their pet peeve be built into the “objective” AGI so they can use it to validate what they believe in. Humans are uniquely designed to understand the nuance of relationships and context, to the point that people who CAN’T understand such things are designated as having various degrees of autism.

So, in a sense, for all its benefits and “smarts” there’s a real risk Humanity so lazy and divided that we’re going to hand over all of our agency to very powerful Rain Man.

Instead of taking a step back and not using AGI to “prove” our personal world views, we’re going to be so busy fighting over what is built into the AGI to be “objective” that we won’t notice that a few trillion dollar industries have been rendered moot.

That’s my existential fear about AGI at the moment — in the future, the vast majority of us will in live in poverty, ruled over by a machine that demands everyone know when we fart.

AGI In Blue & Red

by Shelt Garner
@sheltgarner

While it’s still very speculative, the advent of OpenAI’s chatbot definitely seems to be a ping from an upcoming Singularity. If that is the case, what does that mean for American politics?

Humans are existentially lazy.

As the pandemic showed us, every issue of the day is seen through the prism of partisan politics so the advent of AGI will be no different. It seems to me that the issue of “bias” in AGI will be one of the biggest issues of the 2020s. I say this because people are already fighting over it on Twitter and OpenAI’s chatbot has only been around for a few days.

As such, how will the two sides process the idea of “The Other” in everyday life. My gut tells me that the center-Left will be totally embrace the rise of AGI, while the center-Right will view its presence through the lens of religion. The wild card for the center-Left is, of course, the economic disruption that will be associated with AGI.

If millions of high paying jobs become moot because of AGI, there could be a real knee-jerk reaction against AGI on the part of the Left.

This raises a number of different issues.

One is, it’s possible that that the traditional Blue-Red dichotomy. It could be a real revolution where things are very chaotic and uncertain as we all struggle with the political and economic implications of the AGI revolution. For me, the issue is when all of this bursts open.

Will it be a late 2020s thing, or a late 2024 – early 2025 type of problem? If that’s the case, it would be a perfect storm. If we’re dealing with not just what the final endgame of the Trump problem will be at the same time that we’re dealing with massive economic and political disruption associated with a Singularity…I don’t know what to tell you.

World War Four

by Shelt Garner
@sheltgarner

It’s rare when something so abruptly comes out of the blue that makes me stop thinking about how America is careening towards the existential choice of autocracy, civil war or military junta in late 2024 – early 2025 and makes me think about events past that crisis.

But now, when the OpenAI chatbot, I find myself thinking about what happens next after we sort out our political Trump Problem one way or another.

If you work on the assumption that essentially OpenAI’s chatbot is a ping from an impending Singularity, then it’s very possible that in the late 2020s, the issue of the day will be technology. It could be that the current concept of the Blue -Red paradigm will be up ended and our current understanding of what everything is “Left” or “Right” of will be radically changed.

Imagine a near future where a Non-Human Actor has totally transformed our society and economy so radically and abruptly that a neo-Luddite movement is born. You could see elements of the current Far Right and Far Left fuse together in opposition to the excessive use of technology like AGI to replace humans. And this would be a global realignment.

In fact, given that an AGI would be “The Other” you could plot out a scenario where the even the nation-state becomes a quaint concept as Humanity divides itself into pro-AGI and anti-AGI camps. Your nationality won’t be as important as where you stand on the issue of allowing AGIs or NHAs pretty much replace Humans in every conceivable task.

And that would be a potential World War Four (World War Three having happened around in the mid-2020s as America either becomes an autocracy or has a civil war.) WW4 would be the essential battle over the nature and relationship of AGI. I say this because it’s possible that large segments of humanity might have a near-mystical relationship to an AGI. You see seeds of this happening already with OpenAI’s chatbot, with people using answers from the service to validate what they already believe.

This is all very, very speculative. But it does make you wonder where the American Left and Right would fall politically when it comes to their relationship to an AGI “Other.” At the moment, I think Leftists might be the ones to embrace an AGI as a near-godlike thing, while the Right would freak out because they would, I don’t know, think it was the anti-Christ or something.

Anyway. We really need to start a serious debate about these issues now. If the Singularity really is headed our way, everything may change overnight.

‘Artisanal Media’ In The Age Of NHAs

by Shelt Garner
@sheltgarner

We are still a long ways away from a Non-Human Actor creating a complete movie from scratch, but it’s something we need to start thinking about now instead of waiting until we wake up an almost no art is human produced. Remember, the vast majority of showbiz is middling at best and uses a well established formula.

The day may come when a producer simply feeds that formula into a NHA and — ta da, a movie is spit out.

As long as the art produced is mediocre relative to human standards, it will probably have a great deal of success. It’s possible that movies and TV will be populated by pretty much NFT actors. Or the computerized rendition of existing actors that have been aged or deaged as necessary. I’ve read at least one scifi novel — I think it’s Kiln People by David Brin — that deals with this specific idea.

It could be that NHA-produced art going mainstream will be the biggest change in the entertainment business since the advent of the talkie. Movie stars from just about now will live forever because people won’t realize they’re very old or even dead. Just imagine if Hollywood could keep churning out Indiana Jones movies forever simply using Harrison Ford’s likeness instead of having to recast the character.

All of this raises the issue of what will happen to human generated art in this new era. I suppose after the shock wears off, that there will be parts of the audience who want human created, or artisanal, media. This will probably be a very small segment of the media that is consumed, but it will exist.

It could exist for no other reason than someone physical has to walk the Red Carpet. Though, of course, with advances in robotics in a post-Singularity world, even THAT may not be an issue.

Of course, there is the unknown of if we really are going to reach the Singularity where NHAs are “more human than human.” It could all be a lulz and NHAs won’t really exist as they currently do in my fevered imagination. It could be that AGI will remain just a “tool” and because of various forms of inertia combined with the “uncanny valley” the whole thing will be a lulz.

But, as I said, we all need to really think about what we’re going to do when The Other is producing most of our entertainment and art. And you thought streaming was bad.

Non-Human Actors In Legal Arbitration

by Shelt Garner
@sheltgarner

I’m growing very alarmed at the idea some have proposed on Twitter that we would somehow turn over contract law over to a non-human actor. To me, that’s a very, very dark scenario.

Future humans in an abstract sense?

The moment we begin to believe a non-human actor is the final, objective arbiter of human interaction in a legal sense you’re really opening yourself up to some dystopian shit. The moment we turn over something as weighty as contract law to a NHA, it’s just a quick jaunt for us to all grow so fucking lazy that we just let such a NHA make all of our difficult decisions for us.

I keep thinking of the passengers on the spaceship in the movie WALL-E, only in a more abstract manner. Once it’s acceptable to see a NHA as “objective” then natural human laziness may cause us to repeat the terror of Social Darwinism.

The next thing you know, we’ll be using NHAs to decide who our leaders are. Or to run the economy. Or you name it. As I keep saying on Twitter, why do you need a Terminator when humans apparently are eager to give up their own agency because making decisions is difficult and a lot of work.

Of course, in another way, what I’m suggesting is the fabric of human society may implode because have the population of the earth will want NHAs to make all their decisions for them, while the other half will want to destroy NHAs entirely because…they want to make their own decisions.

But the issue is — we all need to take a deep breath, read a lot of scifi novels and begin to have a frank discussion about what the use of NHAs in everyday life might bring.

‘World War Orwell’ & The Potential Rise of Digital Social Darwinism

by Shelt Garner
@sheltgarner

Tech Bros — specifically Marc Andreessen — are growing hysterical at the prospect that AGI will be in some way hampered by the “woke cancel culture mob” that wants our future hard AI overlord to be “woke.”

Now, this hysteria does raise an interesting — and ominous — possibility. We’re so divided that people may see AGI as some sort of objective arbiter to the point that they use whatever answer it gives to a public policy question as the final word on the matter.

As such, extremists on both sides will rush to the AGI, ask it a dumb extremist question and run around saying, in effect, “Well, God agrees with me, so obviously my belief system is the best.”

In short, humans are dumb.

I definitely don’t agree with Andreessen that this is all a setup for “World War Orwell.” I say this because AGI has reached a tipping point and, as such, we’re all just going to have to deal with the consequences. I definitely think there might be an attempt by one side or the other to instill a political agenda into AGI just because humans are dumbass assholes who are into shit like that.

There is a grander endgame to all of this — we may have to solve the Trump Problem one way or another before we get to play with the goodies of AGI. We may have to have a Second American Civil War in the United States and a Third World War globally before we can turn our attention to the consequences of the macro trends of AGI, automation, robotics and the metaverse.

History may not repeat, but it does rhyme.

Future Shock 2033: AGI Scenarios

by Shelt Garner
@sheltgarner

Let me clear — I don’t know anything about anything. And I’m always wrong. But I love good scenario, so let’s run through what we might face in the next 20 years in regards to artificial general intelligence (AGI).

Humanity Worships A Digital God
In this scenario, rather than end Humanity, our new AGI overlord is very paternalistic towards Humanity. It sees us as its charge, its ward and it does everything in its power to force Humanity to work collectively towards common goals. In this scenario, the problem isn’t so much the existence of AGI but rather Humanity’s reaction to being told what to do. If the AGI is advanced enough millions (billions?) of people may develop a semi-mystical connection to the AGI to the point that they voluntarily bend to its will, no Judgement Day necessary.

A Truce
In this scenario, there is some of agreement between the AGI and Humanity so there are carve outs as to what is exclusively the domain of Humans. This would allow for a peaceful co-existence between Humanity and The Other. As such, there would be some things that the AGI had total control over (like, say, nuclear weapons) and things that only Humans do work with (like, say, the arts.) This is probably one of the better scenarios out there.

It’s All Something of a Dud
In this last scenario, AGI never reaches sentience and, as such, it remains just a tool for humans. Gradually, in fits and starts, this AGI-as-Tool is used to make the lives of the average Human better and we enter something akin to a hyperproductive utopia.

The wild card is, of course, capitalism. It’s the very nature of modern global capitalism that if the system can get away with not paying high salaries for any reason, it won’t. As such, I could see a few trillion dollar industries pretty much become moot simply because a few trillionaires want to cut their bottom line down to the quick. This would, in turn, cause a great deal of instability across the globe and we may find ourselves in something of a Butlerian Jihad situation where neo-Luddites seek to ban the use of AGI for any reason.

But, again, all of that is very speculative. It could go any number of different ways over the course of the next generation.