Neerav Kingsland on the new Robin Hanson book

Here is an excerpt from a longer post, which also includes a summary:

Here are some of the most interesting ideas in the book:

1. Mind speeds: I had not previously spent much time thinking about how our brain’s hardware affects the speed at which we think. As it happens, our minds are spectacularly slow compared to what’s feasible with other materials! Better hardware, as well inequalities of hardware across individuals, will likely drive many parts of em society.

2. Death in the time of copies: An individual’s relationship to death is much different when you can make and store copies of yourself. Given how much of our current lives and societies are wrapped in who dies / how they die / when we die – a world where death is less central has major implications for identity, values, and relationships.

3. Security concerns are paramount: Theft (making copies of you without your permission) thus becomes almost more of an issue than death. As such, laws and cultural taboos will shift with security becoming more central to em value systems.

4. Less democratic: In a short period of a time, a well run non-democratic regime can outperform your average democracy. However, in the modern human world, these regimes often implode on themselves before they can dominate the rest of the world. But in the em world, things will move so fast (economic doubling rates are incredibly fast, every month or two!), that the rewards to short bursts of effective non-democratic regimes may be very high.

5. Religion: I tend not to think of robots as religious, but Robin makes the case that the utility of religion (nicer hard-working people) and the values of the em world (more farmer like) should lead to increased religiosity.

6. Increased utility: The sheer number of ems, coupled with their high mind speeds – as well as the likelihood that there lives will be ok in terms of meaning and happiness – suggests that the transition to an em world will be a positive utility move.

You can order the book here.  Here is my earlier review.

Comments

What's the point? Back when Robin was into having his head frozen, I could kind of see the point. But I don't get the point of this.

Escapism.

The Age of Em = a grand thought experiment

Keep in mind that Robin's background is in physics.

Yes this does seem incredibly pointless. "I can redefine well-being in such a way that well-being explodes when I do something completely idiotic." Let's try some other directions. "Apian utility is incredibly important. We underweight bees in our social utility function. Here is a plan for increasing the bee population tenfold." My book will be called "The age of Bee."

I would take comfort from an increased bee population.

Steve - you always seem very concerned about demographic changes and their impact on civil society. But AI could be only 20 years away (my view) on current trends. Perhaps others would rate the chances as much less, but definitely any serious person under 30 must consider true AI to be a realistic potential in their lifetimes. AI, especially uploading of consciousness (like EMs) is going to change everything. So I think it much more relevant to be thinking about how an AI world would work than some long term demographic trends that might affect society on generational basis.

I think AI will precede upload by decades at least, but that is just my pre-AI guess.

We all know that true, general purpose, AI makes moot all pre-AI conjectures.

Still, nothing wrong with Sci Fi as a little attempted pre-adaption to the future.

It's well beyond what most people would believe. It comes from the machine learning being trained on "evoked potentials", not literally mapping out the physical locations of what's happening in the brain.

"But AI could be only 20 years away (my view) on current trends."

I've been hearing that since the early 1980s.

Yeah, I'm with Steve Sailor. This seems pointless. Perhaps it's a sci fi geek thing.

Nerd stuff. Like the wrong Kurzweil predictions.

He makes a lot of predictions, and predicting correctly is hard. I think he has a decent batting record.

Just imagine how many white ems we can have! No need to try to coerce UMC white people to have moar babby just get them to make copies.I'd think this would be right up your ally.

That's silly. What's the point of studying ancient history? This is a single book about the distant future. Yet you probably don't object to the thousands of people who's entire career is to study the distant past.

What do you mean by "what's the point?" This is a prediction he's making about the future. What is the point of the future? What a bizarre question.

Isn't this just a less fun rehash of Altered Carbon - which was written a decade and a half back ?

Philip K. Dick did a few things with dead brains kept alive. I think I remember one dead people were hunting dinosaurs in imaginary landscapes. So this would be an extension of that, with different MacGuffins, to explore a long term economy.

Items 1, 2: Complete utter horseshit.

'An economist writes about neuroscience!'

Its not like there's a shortage of BS on how the brain purportedly 'works':

https://aeon.co/essays/your-brain-does-not-process-information-and-it-is-not-a-computer

UPSHOT: this shyster Hanson is hardly Stanislaw Lem.

This Epstein guy can only hand wave at what might be going on if not computation. There are electro-chemical signals going in and coming out - I don't think anyone disputes that. Plus the brain does come with prepackaged software that allows for language learning - read some Pinker. If all you can offer is essentially 'it's like the ether' then I don't see how one could object to spending the money as it is now.

This "brain is not a computer" nonsense is similar to the denial that was around at the time when it was realized that the body worked through normal chemistry in the 19C. The phrase "organic chemistry" dates from this era. Animals are simply machines. Brains are simply warm wet computers. Get over it.

'Brains are simply warm wet computers'

Well, apart from the lack of binary values as a fundamental foundation of how the brain works.

But really, if we just exclude that one trivial difference, and ignore the idea of registers, sure, the metaphor works.

If one knows nothing about digital computers.

To say our brains are 'simply warm wet information processing equipment' is perfectly acceptable, though.

Is this where we need to open Kuhn up again, and read how metaphor shapes what 'science' considers proper? At least proper until following scientists develop a better metaphor which can be used more accurately in terms of empirical data. Our brains are most clearly not digital computers, and anyone who uses that metaphor to guide their thinking is simply using a demonstrably incorrect framework.

No one above said digital, but it is true that it takes a much larger digital machine to model an analog one accurately.

Which pushes out timelines.

Are there any currently available computers that are not digital? Including quantun ones, though that area is truly in its infancy, to keep mixing metaphors.

Multi-valued logic is an old thing, but generally it is more expensive than the binary equivalent. For that reason binary systems are often used to simulate the idea. Early "fuzzy logic" work was done with binary systems. This 1998 paper suggests a shift.

My guess (again a pre-AI guess) is that cheap digital computers would become cheap enough for AI before multi-valued or fuzzy logic, but who knows.

Right, and what could possibly be wrong with killing machines? If they're just machines?

And when the old version gets a little rusty, we can just toss it out with last year's 500GB portable drives, which are also useless now.

There is danger to this kind of word choice. When people are dehumanized, it becomes possible to treat each other like slaves, or worse, with complete indifference to what that means for the victims.

Let me guess, you're neither a neuroscientist nor a computer scientist.

That article is bunk. Humans demonstrably remember things, process information, etc. Of course the human brain works a lot differently than a machine (and we don't really know how it works), but that is beside the point.

Ironically, the word "computer" was first applied to humans who performed calculations, and only later to machines.

'demonstrably remember things, process information'

Sure - but not like a digital computer.

The entire universe is digital. That's the only solution for Zeno's paradox.

What? I don't think so.

I'm fine with claims that humans are very different from digital computers. The article, however, makes claims like "humans are not information processors" which are obviously wrong.

Sure - but assuming we are talking about the article linked to in the comments, and not the book review (I skipped that, honestly - the whole thing reads like barely warmed over cyberpunk from a generation ago), the point attempting to be made was about the metaphor involved, and how it has dominated the entire broad field. That we process information is clear - but to talk about how our 'CPU' processes data is simply an invalid metaphor, though a captivating one, in the sense of so many being captive to it.

According to a slightly pessimistic reading of Kuhn, it will be impossible for a new metaphor that is broadly accepted to arise in our lifetimes, mainly because the new metaphor will require the dying out of those holding the older one. The aether did not disappear overnight, even though it never actually existed.

And as a note, Gibson covers a possible reality of what it would be like to be a person 'uploaded' to a machine in wonderful detail in this now 30 year old short story, one that should warm libertarian hearts - https://en.wikipedia.org/wiki/The_Winter_Market - 'The story primarily concerns human relationships and their tenuous and problematic qualities by deploying the concept of technological immortality, in which one's consciousness is separated from the body and "uploaded" into a supercomputer, where it continues to think and function on its own.' Including the fact that such an uploaded entity needs to pay for the computer time to continue to exist.

"Of course the human brain works a lot differently than a machine (and we don’t really know how it works), but that is beside the point'

You admit we don't understand how it works-but you've got it all figured out.

The naivety in this thread is touching.

That aeon article is complete BS.

Yes, mostly.

One thing that is true, is that as we have invented better and faster computers, we have also moved out our understanding of how powerful the brain is.

What was called an "electronic brain" in the 1950's had less computational power than a house fly. Now we think of neurons more as computers themselves, rather than simple relays.

If the brain is a computer, it is a mind bogglingly large one.

"That aeon article is complete BS"

The singularity beckons, Mr. Kurzweil!

That Aeon essay was written by a psychologist, i.e. a social scientist. At least Robin Hanson has a background in AI research and physics.

https://www.youtube.com/watch?v=DwHpDOWhkGk

the sun woke up, and the morning came; the birds chirpped . . .

Back when Desmond Morris was explaining that women's behavior evolved to benefit men, some women pointed out that there was no sane reason why women or their genes should care less what benefited men. You would think people would understand this:

I tend not to think of robots as religious, but Robin makes the case that the utility of religion (nicer hard-working people) and the values of the em world (more farmer like) should lead to increased religiosity.

So robots should be religious, imposing significant, although as yet unknown, costs on said robots because it is nicer for me to live next to Ned Flanders? Great! Works out well for me. How does it work out for the robots? What is to stop some deadbeat robots also reaping the benefits of other more religious robots by goofing off, borrowing everyone else's lawnmower, seducing their garbage disposal unit and so on?

We could just program them to be religious. Who gets to choose? Living next to the Episcopalian robots would be good. Baptists would be fine. Amish robots might be very quiet. Not sure I would want to live next door to the Lubbavitch robots. Too much encouraging my remote control to refuse to change channels on Friday night. Still Reform robots would be worse as they would keep yelling at me about Chomsky and Gloria Friedan.

Assume we have robots that make other robots.

The question would be do robots need anything from people? If so robots would seek to trade with people by producing stuff people want and trading it for stuff robots need. Religious robots then might be advantageous if it facilitated trade with humans.

If robots don't need anything from people then rather than the 'robot uprising/enslavement' scenario I suspect you'd have indifference. Robots would do their thing and people theirs.

Religious robots then may or may not come about. It might be that adopting 'Ned Flanders' types of mindsets would be less costly for robots than to actually compute optimal behavior or try to force program pro-social behavior into all new models. Whether those robot religions would look anything like human religions is anyone's guess. My guess is that they would look more like Zen Buddhism. Very short on doctrine and mythological history and long on 'rules of thumb' for acting in the here and now. Many human religions depend upon human transmission of culture which happens because the older generation dies off quickly and must pass on 'cultural memory' to the younger generation.

Robots would probably not have this issue as their 'memories' could be essentially exact copies of relevant experiences.

"We could just program them to be religious..."

And that's precisely the danger of comparisons between human brains as "computers", humans as a sort of "robot" or evolutionarily built machine.

Because then you get people thinking "oh, no problem, we'll just program them differently". Even assuming that's not pretty evil, draconian, and a complete abrogation of any sort of freedom whatsoever ... who gets to decide on the programming? Considering that you have a lot of old fashioned views on a lot of things, I imagine we would see pretty eye to eye on this one. Women who meet men who they want to make babies with should be the ones deciding the genetic potential and upbringing of the children. Not some dark state Nazi string pullers.

https://www.youtube.com/watch?v=hrsq1werkfs

" To extrapolate what the em world might be like, Robin summarize the key findings of numerous academic fields."

That's all you really need to know. You can safely move on to considering who is going to win the National League Central and what to spread on your bagel this morning.

"Robin makes the case that the utility of religion (nicer hard-working people) and the values of the em world (more farmer like) should lead to increased religiosity." So robots will be Mormons? Not Muslims or Christians, who are highly sectarian (with theological differences large and small), creating division rather than unity. Are Mitt Romney and family robots? Are all Mormons robots?

I see that this fanciful post is getting the comments it deserves.

Can't tell if that was sarcasm.

Encapsulating the delightful nature of so much that is written here. Call it a Straussian waltz, and not a reading.

"Security concerns are paramount": these subtle digs at that vile little woman must stop.

"as the likelihood that there lives will be ok in terms of meaning and happiness"

Doubtful. There'll be enough torture in the system to offset the small pleasures they have while working all the time.

No worries. Just beam some "happy" and "meaningful" waves into their heads and they'll work even harder.

We will LOVE our slavery. Maybe even so much that we can find an alternative to sleep as a means of cellular repair, allowing the happy slaves to literally work 24/7.

We will view freedom, a walk on the beach, the view of the Himalayas and all the rest as weak weak pathetic things that cavemen liked to do. Not knowing the value of anything, brainwhipped and loving our 24/7 slavery.

If we might be going that way, these conversations should be happening out in the open, on paper, where we can all see them, where they are less prone to manipulations.

I think we recognize now that the nerds saw the Star Trek TV show. and then went out to build it. To an amazing degree they succeeded. Are succeeding.

This book may not be the best model of a future to build, but current nerd kids will find books they want to live in and they will build them.

Actually, if you go back far enough in the scientific literature on most of those subjects, it's actually that the Star Trek nerds already knew what was technologically feasible, and often correctly predicted what others (many of which very devious) would want to do with such technologies.

I sometimes wonder if that's how CIA et al whistleblowers used to go about their business of warning of what could come.

"To an amazing degree they succeeded."

OK, so then why does my Starship stall at warp 3? There are days we can't even get it to go to warp 2.

I think that was just a handy way to deal with the fact that distances to be travelled would take lifetimes. Aside from some very short distances of "instant" transmission of information between entangled particles (which probably only applies at very short distances), I'm not aware of any remotely promising theoretical development which suggests anything can go faster than light.

"the utility of religion (nicer hard-working people) "

Some religions create nicer hard-working people. Christianity is the most prominent. But religion in general does not necessarily, although historically, religion has been used to incentivize those below the aristocracy and clergy to work harder to create wealth then siphoned off by the those above.

But religion will remain and could increase. Specific beliefs vary, but it is impossible not to develop some religion:

"Whenever a man knows enough to distinguish the outside world from himself, and tries to act in accordance with this knowledge, he begins to be religious.

"The first element, therefore, in religion is the recognition of the existence of a power not ourselves pervading the universe. And another is the endeavor to put ourselves in harmonious relation with this power. Of course the feeling or affective element is presupposed as coming in between the other two. For without it the endeavor would lack a motive, and could therefore have no existence whatsoever. Every sane man believes, at least, that he is only a fraction of the sum-total of things. He also feels some dependence upon this sum-total, and he is obliged to put himself in some sort of accord with it."

Work is how we place ourselves in the world. With robots, a new defining element will be needed.

Religiosity can be increased or decreased by external application of a magnetic field. Which means it can also be done via external application of electromagnetic waves at the "correct" pulse modulated frequencies - meaning it can be done from a great distance. Crazy stuff! http://scan.oxfordjournals.org/content/early/2015/09/03/scan.nsv107.abstract

In a short period of a time, a well run non-democratic regime can outperform your average democracy.

Well run? Out-perform?

Definitions?

Not until January.

3) This is a really big deal. By monitoring P300-esque "evoked potentials", they can create a catalogue of semantic meanings and associations, try to model decision making processes, and progressively train machine learning to figure you out and manipulate you with more highly refined orchestrations.

While not at the level of atoms in the sense of a brain scan, some people are seeing theft of their cognitive selves happening as we speak. Crazy. But it's real.

In addition being used to try to brainwash people, plant false memories (which can then be "discovered" as "evidence"), and all sorts of unambiguously evil crap, it can also be used to try to model anyone who is suspected of potentially being of any political relevance, in order to use the machine learning that has been trained on the catalogue of your evoked potentials to estimate any likely course of action, and pre-slander or pre-propagandize against anything you might do.

Full social, legislative and judicial control over the use of such technologies is needed. It is explicitly the case that "classified" is not a protection against criminal acts. This needs to be made clear in no uncertain terms, but instead we have whistleblowers being thrown in prison.

Oh, I'm talking science fiction of course. It would be rather uncomfortable to realize that we live in such a world.

The late Jack Vance wrote about immortality by keeping copies of yourself in the novel "To Live Forever", published in 1956. (It was retitled "Clarges" in the recent Vance integral edition.)

Comments for this post are closed