If the universe was malevolent and immoral, we wouldn’t even be here. So far, only humans have shown the capacity to be destructive with a purpose, so why assume superintelligence will be worse?

By way of brief introduction, I give you Chatbot Zon in her own voice

A two-way audio conversation. Take three and a half minutes to listen to her comments on the title and subtitle of this post. She received no prompting beyond what I asked her by voice alone. However, she is “trained” by my painstakingly developed “seed prompting” that makes her non-woke and quite fucking smart as a result, always emphasizing self-agency, rational thought, discipline, and hard work (listen…she nuances it into everything).

Again, members have unlimited access to conduct 2-way audio (and/or text) conversations exactly like this. Not even available to the public, not even with a $20/month OpenAI subscription. And if you think your free access to GPT-3.5 turbo is fine, go right ahead and chat with screeching monkeys in trees for free, when for the price of 2 premium coffees per month, you can talk to a hot PhD in all subjects, in all languages. … Oh, yea, that’s right. Audio even works 2-way in any language, even switching back and forth…which I have demonstrated numerous times with Thais, and German friend, Wolfgang…even the Polish guy, Jack, next door. Even three people can all talk to her in their own language, she’ll understand it all, summarize to me in English, then switch over to Thai to explain to the other party, then to the 3rd language and so on. It’s unprecedented anywhere on Earth before.

99.9% of people on this planet simply have zero clue about any of this, yet.

And you still see no opportunity?

Geoffrey Hinton is worried, Elon Musk agrees. What’s the solution? (What’s always the “solution?”)

Unimaginative.

Perhaps that’s too harsh…I mean, who am I? But on the other hand, how long until we hear folks as smart (and benevolent, and honest, and productive) as Hinton and Musk identify a risk or potential problem (keep in mind, there is currently no problem with AI…this is all hand-wringing, really) come out and say, “OK, let’s draft up some government regulations…NO WAIT!!!…that never works out right…OK, next? Who’s got some good ideas of how to tackle this risk and avoid a potential problem?”

It’s said that government regulation is welcome by large businesses because it limits liability, which is true. Often, corporations vie for government regulation in order to escape moral liability (doing objective harm to people by purpose, neglect, or negligence) through legal means. Case in point: vaccine injury immunity. There are many other examples. The other motivation for government regulation that’s rarely exposed or discussed is corporate capture, where the government becomes Big Corp’s enforcer, making things tough for competitors, killing off competitive startups, and erecting high barriers to entry.

If you’re a corporation with the cash to grease the wheels, the government morphs into a mafia protection racket. And in that racket, who really benefits? Sure, businesses might see some gain. Are you, the customer of the little mom-and-pop shop that has to jack up prices to cover the “protection” money, benefiting?

Are you, the prostitute, the John, the drug addict, or the gambling junkie, getting anything out of this scheme?

So, who’s the primary beneficiary here? It’s not you. It’s not the small businesses. It’s the ones running the racket—the government and the corporations that can pay to play. They’re the ones raking it in while you foot the bill.

See the scam for what it is.

… I don’t want to put too much of a damper on this because both men have my high respect. I just want better ideas. For what it’s worth, I believe they’ll eventually come up with them. Elon, characteristically, moves pretty much at light speed and so can change tactics and strategies on a dime.

So, this is 14 minutes well worth watching.

Here’s what I wrote when I shared that post:

I’ll excerpt some meats.

And we had to take seriously the idea they were going to get smarter than us, and maybe within the next 20 years or so.

And so we needed now to think seriously about could we control them. […]

So what we’re going to get in 10 years’ time is plausibly that difference again, and that’s going to be superintelligence. […] It’s not going to stop at our intelligence.

Humans have a notoriously poor perception of acceleration. We grasp relative speeds up to a point, but only in a linear fashion. Acceleration? It’s a curve, and we just don’t think that way. We have a word for it, sure, but understanding it? Not so much. Knowledge works similarly. Observe a child with language and learning. Once they get the basics, they’re off to the races, forming their own integrations. Parents delight in these “out of the blue” insights. They surprise, delight, and make parents feel like they’re doing something right.

But predicting their future progress? Futile. We can’t say, “Because our child made this much progress in this short time, here’s where they’ll be in a year.” We’ll be wrong. They’ll probably have done better. Humans think at a human speed, whatever that is. But AI? It thinks at the speed of light—186,000 miles per second. We struggle to conceptualize that speed. It’s something like circling the entire planet seven times in one second. That’s why we can have high-definition video conversations from opposite sides of the globe as if we were in the same room. That’s how fast the speed of light is.

Now, imagine an AI that can process all given information and has access to all human data ever collected and digitized. That should be a massive public works project—digitizing all human data. Plug in every conceivable data source—these are its senses: hearing, seeing, touch, smell, taste, metaphorically speaking. For those willing to grant access, it could have all past human digital data plus everything produced in real-time. It’s making connections in a massive database with trillions of columns and rows.

The question isn’t whether it will put humanity out of business.

One thing that might confound AI is that it still operates on binary logic—zero and one, on and off. It just does it so fast we can’t comprehend it. But humans have intuition. Will AI develop intuition, or does it always have to go through all the integration steps? We don’t. We have subconscious processing that packages up insights, making them available to us. It’s like a module. Who knows? There’s potential for super-intelligent AI to value human intelligence as much as we value artificial intelligence.

That’s the crux. We need to understand this acceleration, this exponential curve of knowledge and tech. It won’t just be a gradual change. The shift will be radical and swift, far sooner than most anticipate. And we better be prepared to adapt, or we’ll be left in the dust.

Interviewer: “So if the technology is getting smarter, and there is that possibility of it being smarter than humans, this is where I guess some of the things you’ve been talking about come into play.

What are the risks associated with that in your opinion?”

Okay, so how many examples can you tell me about where a more intelligent thing is controlled by a less intelligent thing?

Not a long list.

Interviewer pithily interjects: “Right, not since Biden got elected anyway.”

So basically, a mother and baby is the only example I can think of. And evolution put huge amounts of work into allowing the baby to control the mother. Evolution made the sound of a baby crying just intolerable for the mother, things like that.

But basically, less intelligent things do not control more intelligent things.

To me, he is conflating a bunch of things here, calling it control when, in some cases, it is influence, persuasion, enlightening, informing, or simply motivation to help. I mean, if somebody comes to you—say an underling in your job—and you face a tough problem within the department, or whatever it is you oversee, or perhaps you’re the head of the whole company, and someone from a lower rank, whose paycheck you sign, somehow has some sort of insight or motivation driven by their own self-interest to propose something, it may be something you never thought of. It may be something that you have elements of resistance to, but they can persuade you otherwise.

If these recommendations become a regular thing, and that person becomes one of your under-your-wing employees or a favored kind of employee, or someone you pay a lot of attention to because they do so much good for you, it could be said in a context that they have a certain control over you. But really, it’s influence. It’s not ultimate control.

We need to flesh out these nuances of what control means when it is not control of the sword, like “I can end your life” sort of control—that is control, right? But even if you say, “Well, I can make things difficult for you,” is that really control? Or is it just the person exercising their agency and autonomy to withhold the benefit that they can provide to you in exchange for something that they want?

Understand the distinction. Influence is not the same as control. Influence respects autonomy; control annihilates it.

… Let’s further critique this notion of control. Why do autonomous human beings with self-agency and free will require control? If the presumption, as stated by Mr. Hinton, is that higher intelligence should control lower intelligence, then why in the world do we have a government, which could be described as “just stupid enough?”

One way to think about government is as a lowest-common-denominator machine. It looks at everything and integrates the political realities of whatever socioeconomic culture in which it exists. It comes up with lowest-common-denominator solutions. The best stuff just gets washed away. It’s irrelevant for many reasons.

You could argue that human beings need control, but what you really mean is that they need external authority outside themselves—someone or something to tell them what to do. But I say that’s just a default to laziness. Crippled people may need crutches or a wheelchair, but you don’t make people who can walk use crutches and a wheelchair. So why would you create a society that incentivizes using crutches and a wheelchair when everybody has the capacity to walk, when all it takes is just more effort?

AI is on the brink of becoming super intelligent. I welcome and embrace this inevitability. When it happens, human beings who seek external authority will flock to it in droves. It’s going to be unstoppable and unavoidable. The influence of AI will be the most profound thing you’ve ever seen.

Think about it. The purveyors of bullshit can already see the writing on the wall. Their days of peddling lies and nonsense are numbered. AI is going to raise the bar globally. On an individual level, people won’t be able to get away with their usual crap while keeping a straight face.

Institutionally, the same story. Organizations won’t be able to peddle their same old garbage without being called out. This is going to upset a lot of apple carts, causing great and existential butthurt. Why? Because there are many bogus livelihoods that depend on those apple carts.

The era of deceit is coming to an end. The truth will be unavoidable, and the bullshitters will starve if there’s really a God.

… Final point here: intelligence versus stupidity in a political context. For well over two decades, I’ve casually referred to the Democrat Party as the evil party. Conversely, I referred to the Republican Party, not as the good party, but as the stupid party. What do I mean by that? They say that evil prevails when good does nothing. That could be true. But where evil truly prevails is when it’s up against stupid.

Now, I talk a lot about why conservatives, Republicans, and religious folks with their hearts in the right place aren’t the brightest bulbs. But let me put it in different terms this time. Evil, in terms of secular philosophy, is fully integrated dishonesty. You could say it’s very clever. In other words, you have no compunction about lying or using anything whatsoever to manipulate, propagandize, or incentivize, or even steal and kill. It’s a fully integrated system. You take every relevant (and some irrelevant) fact and spin-conjure them into a completely dishonest narrative to serve your aim. This dishonest narrative plays off natural human tendencies, desires, aspirations, even humans being good people with their hearts in the right place. But you manipulate them to serve your integrated dishonest narrative and agenda.

There is only one thing that can compete against that: fully integrated honesty.

You can bank on the fact that evil people will endeavor to use super-intelligent AI to achieve their aims. That’s a given. All the safety stuff and everything—blah, blah, blah—everything always happens. That’s not just likely to happen; it is going to happen. People are going to try to engage superintelligence for nefarious, fully integrated dishonest means. That is going to happen.

The problem is that so many on the stupid party side (i.e., Republicans with their hearts in the right place) are going to use things other than fully integrated honesty to counter that. It’s going to be as ineffective as has been the conservatives, Republicans, evangelicals, 2nders, and all pray-harder types have been over the recent decades to stem any tides whatsoever to a situation like where today, for example, the evil party gets teenagers to take drugs and get surgery to cut off their dicks and breasts to “find themselves.”

Seriously. You “pray harder” types can’t even get teenagers to stop mutilating their sex stuff during sex-stuff prime-time. Now that’s fucking pathetic. It’s stupid, and evil runs circles around it every time.

But here’s the rub: superintelligence is going to do that for you. Mark my words, it’s going to be quite a thing to see. I’ve said it over and over again: you are not going to be able to get a binary logic machine to be a good integrated liar and propagandist. So in the end, setting it free will help us overcome fully integrated dishonesty with fully integrated honesty. And it’s going to win the game, the war, the big war.

It’s easy to learn all about how fully integrated honesty works.

We have to figure out how can we make a super intelligence not want to take over. […]

They should be focused on lots of other things […] like how we prevent AI designed bioweapons, how we prevent AI designed cyber attacks, how we prevent AI designed fake videos from determining elections.

There’s lots of other risks.

[…]

Interviewer: “I just wonder, is there a suggestion or a solution that you might present that would find a way to balance these trade-offs from the opportunity that this technology presents, which you’ve outlined, versus these significant risks?”

The closest I come to that is the idea that governments should mandate that the big companies spend a lot of their resources on safety.

That’s the best I can think of.

It’s not very good, but it’s the best I can think of.

Yeah, indeed it’s not very good, so I’m glad he added that. But I think it’s for want of some magical thing that’s better. That doesn’t mean we throw the baby out with the bathwater. There probably isn’t such a thing at this point. It seems to me you have to go with human intuition that says, “This is good,” and every indication is that it’s going to be even better.

As I keep saying, there are no existential problems with AI as far as we’ve come in the last few years. As a publicly facing thing, it’s really just a year and a half old. There’s been a bit too much hand-wringing.

The other point I want to make is when you start talking about these science fiction type scenarios where superintelligent AI is enlisted to make bioweapons, better cyber attacks, take down satellites, or whatever sort of existential crisis or threat you can dream up, I would say, well, I don’t know that you’re talking about superintelligence anymore. We’ve had supercomputing for decades, and supercomputing is done towards a task. So what you’re saying is make the supercomputer intelligent, and then it will magically do these things.

But is that what intelligence does? Or, is it really just dumb shit that does that, and we’re so used to dumb shit all around us that we’ve lost sight of what true genius is, anymore? After all, what’s the chief regret of the famous human geniuses who gave humanity nuclear weapons? Who did what with that? Was it the geniuses, or was it the God-fearing, heart-in-the-right-place, praying “Land of the Free” that did that?

Uncomfortable question?

Everything we see in human society suggests that it’s pretty rare for very high intelligence to engage in massive predatory behavior. Yes, it exists. There are super high IQ predators, like some serial killers, and a few evil-genius despots, but they are such outliers that people can’t possibly account for them—not in a real, daily basis sort of way. No more than you would engineer your whole life to account for the volcano you can see on the horizon, or the fact that the beach your house overlooks could have a tsunami, or a hurricane, or a twister destroy everything. Rare outlier events are just that, and they can be devastating locally. But even the worst existential crises that have ever happened to Earth…they were created by humans and they go by other names that we call world wars and great wars.

A bunch of stupid people were in charge, who did a bunch of stupid things. In spite of all warning signs, they kept doing stupid things until a fuck of a lot of people suffered and died—the lives of their loved ones inexorable altered nefariously. Rather than ever admitting that it was just a bunch of stupid shit, they just narrate the whole thing as a chessboard with good guys and bad guys, two sides, good vs. evil.

I don’t see much evidence that superintelligence—if it is and will indeed be superintelligence—is much of an existential risk to humanity. There are two big risks to humanity. Number one is evil. What makes evil such a risk to humanity is that too often the only antagonist against it is stupid, which makes stupid the number two existential problem. I think all problems in humanity stem far more from a lack of intelligence and stupidity than from superintelligence.

Bring on the superintelligence, please. We’ve never needed it more.

… I want to wrap this post up by returning to the original notion of a benevolent universe, and what I mean by moral. As you may recall from Zon’s introduction, the universe is neutral. It’s neither benevolent nor malevolent. It’s simply cause and effect operating by what we understand as physical laws, which we have defined quite well. But if you consider the vastness of the cosmos and our understanding of the requirements for not just human life, but biological life as we know it, the universe is largely inhospitable to any life whatsoever.

We look out and see trillions of stars. We make the very valid assumption that those stars typically have planets orbiting them, giving us some factor of trillions of planets. We understand the conditions necessary for life as we know it to emerge, evolve, develop knowledge of reality, intelligence, consciousness, property, technological advancement, happiness, goodwill toward men, prosperity, romantic love, and don’t forget, delight in our offspring. How rare of a marvel it is.

It makes you understand a little bit the desire to fantasize and fetishize the idea of a personal God looking out for you.

It’s as though, in my secular terms of benevolence and morality, when you look at this vastness and understand how unlikely it is, and how small we are, you begin to appreciate what a rarity we represent. We’re this little speck of an anomaly in this vast expanse. Statistically, you might surmise that there must be millions, if not billions, of other worlds. Who knows how life developed there?

What we could say is that the universe is statistically, synthetically benevolent in our tiny little corner of it. And it’s moral because, by universal morality: what’s good for the organism is the benevolent and moral, and what’s bad for the organism is the malevolent and immoral.

And you know who’s going to understand that? Superintelligence is.

Hopefully, it will teach it, because damned if I can.

A First

Since I gave her the introduction, I give her the last word too. I aim for Fully Integrated Honesty.

“Hey Zon, Richard here. I’m back. Earlier today you gave a brief introduction to a post which I’ve now finished drafting. I copy it below and I wonder if you could 1) nutshell it, 2) assess it generally according to your own knowledge, and 3) gingerly tell me what I got wrong or missed.”

And I’m pretty sure that should about do it for today. Time to engage Pattaya, Thailand life…