Vernor Vinge on the Turing Test, Artificial Intelligence

/1 Comment/in , /by

Preface

the_imitation_game_bOn the coat-tails of a the blockbuster film “The Imitation Game” I saw quite a bit of buzz on the internet about Alan Turing, and the Turing Test.  The title of the movie refers to the idea of the Turing Test may someday show that machines would ostensibly be (at least in controlled circumstances) indistinguishable from humans.
Vernor Vinge is a mathematician and science fiction author who is well known for many Hugo Award-winning novels and novellas*   and his 1993 essay “The Coming Technological Singularity”, in which he argues that the creation of superhuman artificial intelligence will mark the point at which “the human era will be ended”, such that no current models of reality are sufficient to predict beyond it.

 

Alan Turing and the Computability of Intelligence

Adam Ford: Alan Turing is considered the “Father of Theoretical Computer Science and Artificial Intelligence” – his view about the potential of AI contrasts with much of the skepticism that has subsequently arose.  What is at the root of this skepticism?

Vinge_Singularity_Omni_face250x303Vernor Vinge: The emotional source of the skepticism is the ineffable feeling that many (most?)  people have against the possibility that self-awareness could arise from simple, constructed devices.

 

AF: Many theorists feel that the combined talents of pure machines and humans will always produce more creative and therefore useful output – what are your thoughts?

VV: When it comes to intelligence, biology just doesn’t have legs. _However_ in the near term, teams of people plus machines can be much smarter than either — and this is one of the strongest reasons for being optimistic that we can manage the new era safely, and project that safety into the farther future.

 

AF: Is the human brain essentially a computer?

VV: Probably yes, but if not the lack can very likely be made up for with machine improvements that we humans can devise.

 

AF: Even AI critics John Searle and Hubert Dreyfus (i.e. “What Computers (Still) Can’t Do”) agree that a brain simulation is possible in theory, though they argue that merely mimicking the functioning brain would in itself be an admission of ignorance (concerning intelligence) – what are your thoughts?

VV: The question of whether there is self-awareness behind a mimick may be the most profound issue, but for almost all practical purposes it isn’t relevant: in a few years, I think we will be able to make machines that can run circles around any human mind by all externally measured criteria. So what if no one is really home inside that machine?

Offhand, I can think of only one practical import to the answer, but that _is_ something important: If such minds are self-aware in the human sense, then uploads suddenly become very important to us mortality-challenged beings.

For reductionists interested in _that_ issue, some confidence might be achieved with superintelligence architectures that model those structures in our minds that reductionists come to associate with self-awareness. (I can imagine this argument being carried on by the uploaded supermind children of Searle and Moravec — a trillion years from now when there might not be any biological minds around whatsoever.)

 

AF: Do you think Alan Turing’s reasons for believing in the potential of AI are different from your own and other modern day theorists?  If so in what ways?

VV: My guess is there is not much difference.

 

AF: Has Alan Turing and his work influenced your writing? If it has, how so?

VV: I’m not aware of direct influence. As a child, what chiefly influenced me was the science-fiction I was reading! Of course, those folks were often influenced by what was going in science and math and engineering of the time.

Alan Turing has had a multitude of incarnations in science fiction…   I think that besides being a broadly based math and science genius, Turing created accessible connections between classic philosophical questions and current issues.

 

AF: How do you think Alan Turing would respond to the specific concept of the Technological Singularity as described by you in your paper “The Coming Technological Singularity: How to Survive in the Post-Human Era“?

VV: I’d bet that Turing (and many AI pioneers) had extreme ideas about the consequences of superhuman machine intelligence. I’m not sure if Turing and I would agree about the potential for Intelligence Amplification and human/machine group minds.

I’d be _very_ interested in his reaction to modern analysis such as surveyed in Bostrom’s recent _Superintelligence_ book.

 

AF: In True Names, agents seek to protect their true identity. The guardian of the Coven’s castle is named ‘Alan Turing’ – what was the reason behind this?

It was a tip of the hat in Turing’s direction. By the time I wrote this story I had become quite aware of Alan Turing (contrasting with my childhood ignorance that I mentioned earlier).

 

AF: Your first novella Bookworm Run! was themed around brute forcing simpler-than-human-intelligence to super-intelligence (in it a chimpanzee’s intelligence is amplified).  You also explore the area of intelligence amplification in Marooned in Realtime.
Do you think it is possible for a Singularity to bootstrap from brute forcing simple cognitive models? If so do you think Super-Intelligence will be achieved through brute-forcing simple algorithms?

VV: I view “Intelligence Amplification” (IA) as a finessing of the hardest questions by building on top of what already exists. Thus even UI design lies on the path to the Singularity. One could argue that Intelligence Amplification is the surest way of insuring humanity in the super-intelligence (though some find that a very scary possibility in itself).

 

The Turing Test and Beyond

AF: Is the Turing Test important? If so, why, and how does it’s importance match up to tracking progress in Strong AI?

VV: In its general form, I regard the Turing Test as a marvelous, zen-like, bridge between reductionism and the inner feelings most people have about their own self-awareness.  Bravo Dr. Turing!

 

AF: Is a text conversation is ever a valid test for intelligence? Is blackbox testing enough for a valid test for intelligence?

VV: “Passing the Turing Test” depends very much on the setup:
a) The examining human (child? adult? fixated or afflicted adult? –see Sherry Turkle’s examples of college students who passed a chatbot).
b) The duration of the test.
c) The number of human examiners participating.
d) Restrictions on the examination domain.

In _The Emperor’s New Mind_, Penrose has a (mostly negative) critique of the Turing Test. But at the end he says that if the test was very broad, lasting years, and convincing to him (Penrose), then it might be meaningful to talk about a “pass grade”.

 

AF: The essence of Roger Penrose’s argument (in the Emperor’s New Mind)
–  It is impossible for a Turing machine to enumerate all possible Godel sentences. Such a program will always have a Godel sentence derivable from its program which it can never discover
–  Humans have no problem discovering these sentences and seeing the truth of them
And he concludes that humans are not reducible to turing machines.  Do you agree with Roger’s assessment  – Are humans not reducible to turing machines?

VV: This argument depends on comparing a mathematical object (the Turing Machine) with whatever kind of object the speaker considers a “human mind” to be.  As a logical argument, it leaves me dubious.

 

AF: Are there any existing interpretations of the Turing Test that you favour?

VV: I think Penrose’s version (described above) is the most important.

In conversation, the most important thing is that all sides know which flavor of the test they are talking about 🙂

 

AF: You mentioned it has been fun tracking Turing Test contests, what are your thoughts on attempts at passing the Turing Test so far?

VV: So far, it seems to me that the philosophically important thresholds are still far away. Fooling certain people, or fooling people for short periods of time seems to have been accomplished.

 

AF: Is there any specific type of intelligence we should be testing machines for?

VV: There are intelligence tests that would be very interesting to me, but I rather not call them versions of the Turing Test. For instance, I think we’re already in the territory where more and more [forms->sorts] of superhuman forms of creativity and “intuition” are possible.

I think there well also be performance tests for IA and group mind projects.

 

AF: Some argue that testing for ‘machine consciousness’ is more interesting – what are your thoughts?

VV: Again, I’d keep this possibility separate from Turing Test issues, though I do think that a being that could swiftly duplicate itself and ramp intellect up or down per resource and latency constraints would have a vastly different view of reality compared to the severe and static time/space/mortality restrictions that we humans live with.

 

AF: The Turing Test seems like a competitive sport.  Though some interpretations of the Turing Test have conditions which seem to be quite low.  The competitive nature of how the Turing Test is staged seems to me to select for the cheapest and least sophisticated methods to fool judges on a Turing Test panel.

VV: Yes.

 

AF: Should we be focusing on developing more complex and adaptive Turing style tests (more complex measurement criteria? more complex assessment)? What alternatives to a Turing Test competition (if any) would you suggest to motivate regular testing for machine intelligence?

VV: The answers to these questions may grow out of hard engineering necessity more than from the sport metaphor. Going forward, I imagine that different engineering requirements will acquire various tests, but they may look more like classical benchmark tests.

 

Tracking Progress in Artificial Intelligence

AF: Why is tracking progress towards AI important?

VV: Up to a point it could be important for the sort of safety reasons Bostrom discusses in _Superintelligence_. Such tracking could also provide some guidance for machine/human/society teams that might have the power to guide events along safe paths.

 

AF: What do you see as the most useful mechanisms for tracking progress towards a) human equivalence in AI, b) a Technological Singularity?

VV: The approach to human equivalence might be tracked with a broad range of tests. Such would also apply to the Singularity, but for a soft takeoff, I imagine there would be a lot of economic effects that could be tracked. For example:
–  trends in employment of classic humans, augmented humans, and computer/human teams;
–  trends in what particular jobs still have good employment;
–  changes in the personal characteristics of the most successful CEOs.

Direct tests of automation performance (such as we’ve discussed above) are also important, but as we approach the Singularity, the center of gravity shifts from the programmers to the programs and how the programs are gaming the systems.

 

AF: If you had a tardis and you could bring Alan Turing forward into the 21st century, would he be surprised at progress in AI?  What kinds of progress do you think he would be most interested in?

VV: I don’t have any special knowledge of Turing, but my guess is he would be pleased — and he would want to _understand_ by becoming a super himself.

 

AF: If and when the Singularity becomes imminent – is it likely that the majority of people will be surprised?

VV: A hard takeoff would probably be a surprise to most people. I suspect that a soft takeoff would be widely recognized.

 

Implications

AF: What opportunities could we miss if we are not well prepared (This includes opportunities for risk mitigation)?

VV: Really, the risk mitigation is the serious issue. Other categories of missed opportunities will probably be quickly fixed by the improving tech.  For pure AI, some risk mitigation is the sort of thing MIRI is researching.

For pure AI, IA, and group minds, I think risk mitigation involves making use of the human equivalent minds that already exist in great numbers (namely, the human race). If these teams and early enhancements recognized the issues, they can form a bridge across to the more powerful beings to come.

 

AF: You spoke about an AI Hard Takeoff as being potentially very bad – can you elaborate here?

VV: A hard takeoff is too fast for normal humans to react and accommodate to.  To me, a Hard Takeoff would be more like an explosion than like technological progress. Any failure in mitigation planning is suddenly beyond the ability of normal humans to fix.

 

AF: What stood out for you after reading Nick Bostrom’s book ‘Superintelligence – paths, dangers, strategies’?

VV: Yes. I think it’s an excellent discussion especially of the pure AI path to superintelligence. Even people who have no intense interest in these issues would find the first few chapters interesting, as they sketch out the problematic issues of pure AI superintelligence — including some points that may have been missed back in the twentieth century. The book then proceeds to a fascinating analysis of how to cope with these issues.

My only difference with the analysis presented is that while pure AI is likely the long term important issue, there could well be a period (especially in the case of a Soft Takeoff) where the IA and groupmind trajectories are crucial.

vernor_vinge_LosCon

Vernor Vinge at Los Con 2012

Notes:
* Hugo award winning novels & novellas include: A Fire Upon the Deep (1992), A Deepness in the Sky (1999), Rainbows End (2006), Fast Times at Fairmont High (2002), and The Cookie Monster (2004), and The Peace War (1984).

Also see video interview with Vernor Vinge on the Technological Singularity.

Cognitive Biases & In-Group Convergences with Joscha Bach

/0 Comments/in , , , /by

True & false vs right & wrong – People converge their views to set of rights and wrongs relative to in-group biases in their peer group.
As a survival mechanism, convergence in groups is sometimes more healthy than being right – so one should optimize for convergence sometimes even at the cost of getting stuff wrong – so humans probably have an evolutionary propensity to favor convergence over truth.
However by optimizing for convergence may result in the group mind being more stupid than the smartest people in the group.

 

 
Joscha highlights the controversy of Yonatan Zunger being fired for sending out an email about biological differences between men and women effecting abilities as engineers – where Zunger’s arguments may be correct – now regardless of what the facts are about how biological differences effect differences in ability between men & women, google fired him because they thought supporting these arguments would make for a worse social environment.

This sort of thing leads to an interesting difference in discourse, where:
* ‘nerds’ tend to focus on ‘content‘, on imparting ideas and facts where everyone can judge these autonomously and form their own opinions – in view that in order to craft the best solutions we need to have the best facts
* most people the purpose of communication is ‘coordination‘ between individuals and groups (society, nations etc) – where the value on a ‘fact’ is it’s effect on the coordination between people

So is Google’s response to the memo controversy about getting the facts right, or about how Google at this point should be organised?

What’s also really interesting is that different types of people read this ‘memo’ very differently – making it very difficult to form agreement about the content of this memo – how can one agree on whats valuable about communication – whether it’s more about imparting ideas and facts or whether it’s more about coordination?

More recently there has been a lot of talk about #FakeNews – where it’s very difficult to get people to agree to things that are not in their own interests – and including, as Joshca points out, the idea that truth matters.

Joscha Bach, Ph.D. is an AI researcher who worked and published about cognitive architectures, mental representation, emotion, social modeling, and multi-agent systems. He earned his Ph.D. in cognitive science from the University of Osnabrück, Germany, and has built computational models of motivated decision making, perception, categorization, and concept-formation. He is especially interested in the philosophy of AI and in the augmentation of the human mind.

Discussion points:
– In-group convergence: thinking in true & false vs right & wrong
– The group mind may be more stupid than the smartest individuals in the group

The Generative Universe Hypothesis

/0 Comments/in , /by

Remembering Lee Smolin’s theory of the dynamical evolution of the universe  where through a form of natural selection, black holes spawn new universes, I thought that if a superintelligent civilization understood its mechanics, they may try to control it, and engineer or bias the spawned universe – say that they found out how to talk along the parent/child relations between universes, it may be a an energy efficient way achieve some of the outcomes of simulations (as described in Nick Bostrom’s Simulation Hypothesis).

Why leave the future evolution of universes up to blind natural selection?  An advanced post-singularity alien civs might hypothesize an extremely strict set of criteria to allow for the formation of the right kinds of matter and energy in child universes to either mirror our own universe, or take it to another level – where the limits of computation is limited by the laws of this universe, it may be that spawning a new universe could allow for more efficient computation.

It may also be a great way to escape the heat death of the universe 🙂

I spoke about the idea with Andrew Arnel recently while out for a drink, where I came up with a really cool name for this idea – though I can’t remember what it was 🙂  perhaps it only sounds good after a few beers – perhaps it was something like the ‘generative’ or’ spawnulation or ‘genulation’ hypothesis…

Why Technology Favors a Singleton over a Tyranny

/0 Comments/in , /by

Is democracy loosing its credibility, will it cede to dictatorship?  Will AI out-compete us in all areas of economic usefulness – making us the future useless class?

It’s difficult to get around the bottlenecks of networking and coordination in distributed democracies. In the past quite naturally distributed systems being scattered are more redundant wer in many ways fault tolerant and adaptive – though these payoffs for most of us may dwindle if humans become less and less able to compete with Ex Machina. If the relative efficiency of democracies to dictatorships tips towards the latter nudging a transition to centralized dictatorships, while solving some distribution & coordination problems, the concentration of resource allocation may be exaggerated beyond historical examples of tyranny. Where the ‘once was proletariat’ now new ‘useless class’ have little to no utility to the concentration of power – the top 0.001% – the would be tyrants will likely give up on ruling and tyrannizing – and instead find it easier to cull the resource hungry and rights demanding horde – more efficient that way. Ethics is fundamental to fair progress – ethics is philosophy with a deadline creeping closer – what can we do to increase the odds of a future where the value of life is evaluated beyond it’s economic usefulness?
I found ‘Why Technology Favors Tyranny by Yuval Noah Harari‘ was a good read – I enjoy his writing, and it provokes me to think.  About 5 years ago I did the ‘A Brief History of Humankind’ course via coursera – urging my friends to join me.  Since then Yuval has taken the world by storm.
The biggest and most frightening impact of the AI revolution might be on the relative efficiency of democracies and dictatorships. […] We tend to think about the conflict between democracy and dictatorship as a conflict between two different ethical systems, but it is actually a conflict between two different data-processing systems. Democracy distributes the power to process information and make decisions among many people and institutions, whereas dictatorship concentrates information and power in one place. Given 20th-century technology, it was inefficient to concentrate too much information and power in one place. Nobody had the ability to process all available information fast enough and make the right decisions. […]Why Technology Favors Tyranny
I assume AI superintelligence is highly probable if we don’t go extinct first.  For the same reason that the proletariat’s become useless I think ultimately the AI-Human combination will likely become useless too, and cede to Superintelligent AI – so all humans becomes useless. The bourgeoisie elite may initially feel safe in the idea that they don’t need to be useful, they just need to maintain control of power. Though the sliding relative dumbness of bourgeoisie to superintelligence will worry them.. perhaps not long after wiping out the useless class, the elite bourgeoisie will then see the importance of the AI control problem, and that their days are numbered too – at which point will they see ethics and the value of life beyond economic usefulness as important?
However, artificial intelligence may soon swing the pendulum in the opposite direction. AI makes it possible to process enormous amounts of information centrally. In fact, it might make centralized systems far more efficient than diffuse systems, because machine learning works better when the machine has more information to analyze. If you disregard all privacy concerns and concentrate all the information relating to a billion people in one database, you’ll wind up with much better algorithms than if you respect individual privacy and have in your database only partial information on a million people. An authoritarian government that orders all its citizens to have their DNA sequenced and to share their medical data with some central authority would gain an immense advantage in genetics and medical research over societies in which medical data are strictly private. The main handicap of authoritarian regimes in the 20th century—the desire to concentrate all information and power in one place—may become their decisive advantage in the 21st century.Why Technology Favors Tyranny
Yuval Noah Harari believes that we could be heading for a technologically enabled tyranny as AI automates all jobs away – and we become the useless class. Though if superintellignece is likely, then human’s will likely to be a bottleneck in any AI/Human hybrid use case – if tyranny happens, it won’t last for long – what use is a useless class to the elite?

Technology without ethics favors singleton utility monsters – not a tyranny – what use is it to tyrannize over a useless class?

Uncovering the Mysteries of Affective Neuroscience – the Importance of Valence Research with Mike Johnson

/0 Comments/in /by

Valence in overview

Adam: What is emotional valence (as opposed to valence in chemistry)?

Mike: Put simply, emotional valence is how pleasant or unpleasant something is. A somewhat weird fact about our universe is that some conscious experiences do seem to feel better than others.

 

Adam: What makes things feel the way they do? What makes some things feel better than others?

Mike: This sounds like it should be a simple question, but neuroscience just don’t know. It knows a lot of random facts about what kinds of experiences, and what kinds of brain activation patterns, feel good, and which feel bad, but it doesn’t have anything close to a general theory here.

..the way affective neuroscience talks about this puzzle sometimes sort of covers this mystery up, without solving it.

And the way affective neuroscience talks about this puzzle sometimes sort of covers this mystery up, without solving it. For instance, we know that certain regions of the brain, like the nucleus accumbens and ventral pallidum, seem to be important for pleasure, so we call them “pleasure centers”. But we don’t know what makes something a pleasure center. We don’t even know how common painkillers like acetaminophen (paracetamol) work! Which is kind of surprising.

In contrast, the hypothesis about valence I put forth in Principia Qualia would explain pleasure centers and acetaminophen and many other things in a unified, simple way.

 

Adam: How does the hypothesis about valence work?

Mike: My core hypothesis is that symmetry in the mathematical representation of an experience corresponds to how pleasant or unpleasant that experience is. I see this as an identity relationship which is ‘True with a capital T’, not merely a correlation.  (Credit also goes to Andres Gomez Emilsson & Randal Koene for helping explore this idea.)

What makes this hypothesis interesting is that
(1) On a theoretical level, it could unify all existing valence research, from Berridge’s work on hedonic hotspots, to Friston & Seth’s work on predictive coding, to Schmidhuber’s idea of a compression drive;

(2) It could finally explain how the brain’s so-called “pleasure centers” work– they function to tune the brain toward more symmetrical states!

(3) It implies lots and lots of weird, bold, *testable* hypotheses. For instance, we know that painkillers like acetaminophen, and anti-depressants like SSRIs, actually blunt both negative *and* positive affect, but we’ve never figured out how. Perhaps they do so by introducing a certain type of stochastic noise into acute & long-term activity patterns, respectively, which disrupts both symmetry (pleasure) and anti-symmetry (pain).

 

Adam: What kinds of tests would validate or dis-confirm your hypothesis? How could it be falsified and/or justified by weight of induction?

Mike: So this depends on the details of how activity in the brain generates the mind. But I offer some falsifiable predictions in PQ (Principia Qualia):

  • If we control for degree of consciousness, more pleasant brain states should be more compressible;
  • Direct, low-power stimulation (TMS) in harmonious patterns (e.g. 2hz+4hz+6hz+8hz…160hz) should feel remarkably more pleasant than stimulation with similar-yet-dissonant patterns (2.01hz+3.99hz+6.15hz…).

Those are some ‘obvious’ ways to test this. But my hypothesis also implies odd things such as that chronic tinnitus (ringing in the ears) should product affective blunting (lessened ability to feel strong valence).

Note: see https://qualiacomputing.com/2017/06/18/quantifying-bliss-talk-summary/ and http://opentheory.net/2018/08/a-future-for-neuroscience/ for a more up-to-date take on this.

 

Adam: Why is valence research important?

Mike Johnson: Put simply, valence research is important because valence is important. David Chalmers famously coined “The Hard Problem of Consciousness”, or why we’re conscious at all, and “The Easy Problem of Consciousness”, or how the brain processes information. I think valence research should be called “The Important Problem of Consciousness”. When you’re in a conscious moment, the most important thing to you is how pleasant or unpleasant it feels.

That’s the philosophical angle. We can also take the moral perspective, and add up all the human and non-human animal suffering in the world. If we knew what suffering was, we could presumably use this knowledge to more effectively reduce it and make the world a kinder place.

We can also take the economic perspective, and add up all the person-years, capacity to contribute, and quality of life lost to Depression and chronic pain. A good theory of valence should allow us to create much better treatments for these things. And probably make some money while doing it.

Finally, a question I’ve been wondering for a while now is whether having a good theory of qualia could help with AI safety and existential risk. I think it probably can, by helping us see and avoid certain failure-modes.

 

Adam: How can understanding valence could help make future AIs safer? (How to help define how the AI should approach making us happy?, and in terms of a reinforcement mechanism for AI?)

Mike: Last year, I noted a few ways a better understanding of valence could help make future AIs safer on my blog. I’d point out a few notions in particular though:

  • If we understand how to measure valence, we could use this as part of a “sanity check” for AI behavior. If some proposed action would cause lots of suffering, maybe the AI shouldn’t do it.
  • Understanding consciousness & valence seem important for treating an AI humanely. We don’t want to inadvertently torture AIs- but how would we know?
  • Understanding consciousness & valence seems critically important for “raising the sanity waterline” on metaphysics. Right now, you can ask 10 AGI researchers about what consciousness is, or what has consciousness, or what level of abstraction to define value, and you’ll get at least 10 different answers. This is absolutely a recipe for trouble. But I think this is an avoidable mess if we get serious about understanding this stuff.

 

Adam: Why the information theoretical approach?

Mike: The way I would put it, there are two kinds of knowledge about valence: (1) how pain & pleasure work in the human brain, and (2) universal principles which apply to all conscious systems, whether they’re humans, dogs, dinosaurs, aliens, or conscious AIs.

It’s counter-intuitive, but I think these more general principles might be a lot easier to figure out than the human-specific stuff. Brains are complicated, but it could be that the laws of the universe, or regularities, which govern consciousness are pretty simple. That’s certainly been the case when we look at physics. For instance, my iPhone’s processor is super-complicated, but it runs on electricity, which itself actually obeys very simple & elegant laws.

Elsewhere I’ve argued that:

>Anything piped through the complexity of the brain will look complex, regardless of how simple or complex it starts out as. Similarly, anything will look irreducibly complex if we’re looking at it from the wrong level of abstraction.

 

Adam: What do you think of Thomas A. Bass’s view of ITheory – he thinks that (at least in many cases) it has not been easy to turn data into knowledge. That there is a pathological attraction to information which is making us ‘sick’ – he calls it Information Pathology. If his view offers any useful insights to you concerning avoiding ‘Information Pathology’ – what would they be?

Mike: Right, I would agree with Bass that we’re swimming in neuroscience data, but it’s not magically turning into knowledge. There was a recent paper called “Could a neuroscientist understand a microprocessor?” which asked if the standard suite of neuroscience methods could successfully reverse-engineer the 6502 microprocessor used in the Atari 2600 and NES. This should be easier than reverse-engineering a brain, since it’s a lot smaller and simpler, and since they were analyzing it in software they had all the data they could ever ask for, but it turned out that the methods they were using couldn’t cut it. Which really begs the question of whether these methods can make progress on reverse-engineering actual brains. As the paper puts it, neuroscience thinks it’s data-limited, but it’s actually theory-limited.

The first takeaway from this is that even in the age of “big data” we still need theories, not just data. We still need people trying to guess Nature’s structure and figuring out what data to even gather. Relatedly, I would say that in our age of “Big Science” relatively few people are willing or able to be sufficiently bold to tackle these big questions. Academic promotions & grants don’t particularly reward risk-taking.

 

Adam: Information Theory frameworks – what is your “Eight Problems” framework and how does it contrast with Giulio Tononi’s Integrated Information Theory (IIT)? How might IIT help address valence in a principled manner? What is lacking IIT – and how does your ‘Eight Problems’ framework address this?

Mike: IIT is great, but it’s incomplete. I think of it as *half* a theory of consciousness. My “Eight Problems for a new science of consciousness” framework describes what a “full stack” approach would look like, what IIT will have to do in order to become a full theory.

The biggest two problems IIT faces is that (1) it’s not compatible with physics, so we can’t actually apply it to any real physical systems, and (2) it says almost nothing about what its output means. Both of these are big problems! But IIT is also the best and only game in town in terms of quantitative theories of consciousness.

Principia Qualia aims to help fix IIT, and also to build a bridge between IIT and valence research. If IIT is right, and we can quantify conscious experiences, then how pleasant or unpleasant this experience is should be encoded into its corresponding mathematical object.

 

Adam: What are the three principles for a mathematical derivation of valence?

Mike: First, a few words about the larger context. Probably the most important question in consciousness research is whether consciousness is real, like an electromagnetic field is real, or an inherently complex, irreducible linguistic artifact, like “justice” or “life”. If consciousness is real, then there’s interesting stuff to discover about it, like there was interesting stuff to discover about quantum mechanics and gravity. But if consciousness isn’t real, then any attempt to ‘discover’ knowledge about it will fail, just like attempts to draw a crisp definition for ‘life’ (elan vital) failed.

If consciousness is real, then there’s a hidden cache of predictive knowledge waiting to be discovered. If consciousness isn’t real, then the harder we try to find patterns, the more elusive they’ll be- basically, we’ll just be talking in circles. David Chalmers refers to a similar distinction with his “Type-A vs Type-B Materialism”.

I’m a strong believer in consciousness realism, as are my research collaborators. The cool thing here is, if we assume that consciousness is real, a lot of things follow from this– like my “Eight Problems” framework. Throw in a couple more fairly modest assumptions, and we can start building a real science of qualia.

Anyway, the formal principles are the following:

  1. Consciousness can be quantified. (More formally, that for any conscious experience, there exists a mathematical object isomorphic to it.)
  2. There is some order, some rhyme & reason & elegance, to consciousness. (More formally, the state space of consciousness has a rich set of mathematical structures.)
  3. Valence is real. (More formally, valence is an ordered property of conscious systems.)

 

Basically, they combine to say: this thing we call ‘valence’ could have a relatively simple mathematical representation. Figuring out valence might not take an AGI several million years. Instead, it could be almost embarrassingly easy.

 

Adam: Does Qualia Structuralism, Valence Structuralism and Valence Realism relate to the philosophy of physics principles of realism and structuralism? If so, is there an equivalent ontic Qualia Structuralism and Valence Structuralism?….

Mike: “Structuralism” is many things to many contexts. I use it in a specifically mathematical way, to denote that the state space of qualia quite likely embodies many mathematical structures, or properties (such as being a metric space).

Re: your question about ontics, I tend to take the empirical route and evaluate claims based on their predictions whenever possible. I don’t think predictions change if we assume realism vs structuralism in physics, so maybe it doesn’t matter. But I can get back to you on this. 🙂

 

Adam: What about the Qualia Research Institute I’ve also recently heard about :D! It seems both you (Mike) and Andrés Gómez Emilson are doing some interesting work there

Mike: We know very little about consciousness. This is a problem, for various and increasing reasons– it’s upstream of a lot of futurist-related topics.

But nobody seems to know quite where to start unraveling this mystery. The way we talk about consciousness is stuck in “alchemy mode”– we catch glimpses of interesting patterns, but it’s unclear how to systematize this into a unified framework. How to turn ‘consciousness alchemy’ into ‘consciousness chemistry’, so to speak.

Qualia Research Institute is a research collective which is working on building a new “science of qualia”. Basically, we think our “full-stack” approach cuts through all the confusion around this topic and can generate hypotheses which are novel, falsifiable, and useful.

Right now, we’re small (myself, Andres, and a few others behind the scenes) but I’m proud of what we’ve accomplished so far, and we’ve got more exciting things in the pipeline. 🙂

Also see the 2nd part, and the 3rd part of this interview series. Also this interview with Christof Koch will likely be of interest.

 

Mike Johnson is a philosopher living in the Bay Area, writing about mind, complexity theory, and formalization. He is Co-founder of the Qualia Research Institute. Much of Mike’s research and writings can be found at the Open Theory website.
‘Principia Qualia’ is Mike’s magnum opus – a blueprint for building a new Science of Qualia. Click here for the full version, or here for an executive summary.
If you like Mike’s work, consider helping fund it at Patreon.

Physicalism & Materialism – John Wilkins

/0 Comments/in , , , , /by

Materialism was a pre-socratic view that for something to be real it has to be matter – physical stuff made of atoms (which at the time were considered hard like billiard balls – fundametal parts of reality).  The reason these days the term physicalism is used is because it can describe things that aren’t matter – like forces, or aren’t observable matter – like dark matter, or energy or fields, or spacetime etc..  Physicalism is the idea that all that exist can be described in the language of some ‘ideal’ physics – we may never know what this ideal physics is, though people think that it is something close to our current physics (as we can make very accurate predictions with our current physics).

If magic, telepathy or angels were real, there would be a physics that could describe them – they’d have patterns and properties that would be describable and explainable.  A physicist would likely think that even the mind operates according to physical rules.  Being a physicalist according to John means you think everything is governed by rules, physical rules – and that there is an ideal language that can be used to describe all this.

Note John is also a deontologist.  Perhaps there should exist an ideal language that can fully describe ethics – does this mean that ideally there is no need for utilitarianism?  I’ll leave that question for another post.

Interview with John Wilkins on Materialism & Physicalism.

Here are some blog posts about physicalism by John Wilkins:

Is physicalism an impoverished metaphysics?

Every so often, we read about some philosopher or other form of public intellectual who makes the claim that a physicalist ontology – a world view in which only things that can be described in terms of physics are said to exist – is impoverished. That is, there are things whereof science cannot know, &c. A recent example is that made by Thomas Nagel [nicely eviscerated here by the physicist Sean Carroll], whose fame in philosophy rests with an influential 1974 paper that there is something like being a bat that no amount of physics, physiology or other objective science could account for.

Recent, Nagel has argued that the evolutionary view called (historically misleadingly) neo-Darwinism, is “almost certainly” false. One of the reasons is that “materialism” (which Nagel should know is an antiquated world view replaced by physicalism defined above; there are many non-material things in physics, not least fields of various kinds) does not permit a full account of consciousness; the subjective facts of being a particular individual organism. Another is that the chance that life would emerge from a lifeless universe is staggeringly unlikely. How this is calculated is somewhat mysterious, given that at best we only have (dare I say it?) subjective estimates anyway, but there it is.

But Nagel is not alone. Various nonreligious (apparently) thinkers have made similar assertions, although some, like Frank Jackson, who proposed the Knowledge Argument, have since backed down. What is it that physicalism must account for that these disputants and objectors say it cannot?

It almost entirely consists of consciousness, intentions, intelligence or some similar mental property which is entirely inexplicable by “reductionist” physicalism. [Reductionism is a term of abuse that means – so far as I can tell – solely that the person who makes such an accusation does not like the thing or persons being accused.] And that raises our question: is physicalism lacking something?

I bet you are dying to know more… you’ll just have to follow the link…
See more at Evolving Thoughts>>

Is Physicalism Coherent?

In my last post I argued that physicalism cannot be rejected simply because people assert there are nonphysical objects which are beyond specification. Some are, however, specifiable, and one commentator has identified the obvious ones: abstract objects like the rules of chess or numbers. I have dealt with these before in my “Pizza reductionism” post, which I invite you to go read.

Done? OK, then; let us proceed.

It is often asserted that there are obviously things that are not physical, such as ideas, numbers, concepts, etc., quite apart from qualia, I once sat with a distinguished philosopher, who I respect greatly and so shall not name, when he asserted that we can construct natural classifications because we can deal first with the natural numbers. I asked him “In what sense are numbers natural objects?”, meaning, why should we think numbers are entities in the natural world. He admitted that the question had not occurred to him (I doubt that – he is rather smart), but that it was simply an axiom of his philosophy. I do not think such abstract objects are natural.

This applies to anything that is “informational”, including all semantic entities like meanings, symbols, lexical objects, and so on. They only “exist” as functional modalities in our thoughts and language. I have also argued this before: information does not “exist”; it is a function of how we process signals. Mathematics is not a domain, it is a language, and the reason it works is because the bits that seriously do not work are not explored far[*] – not all of it has to work in a physical or natural sense, but much of it has to, or else it becomes a simple game that we would not play so much.

So the question of the incoherence of physicalism is based on the assumption (which runs contrary to physicalism, and is thus question begging) that abstract objects are natural things. I don’t believe they are, and I certainly do not think that a thought, or concept, for example, which can be had by many minds and is therefore supposed to be located in none of them (and thus transcendental), really is nonphysical. That is another case of nouning language. The thought “that is red” exists, for a physicalist, in all the heads that meet the functional social criteria for ascriptions of red. It exists nowhere else – it just is all those cognitive and social behaviours in biological heads…

Yes, I know, it’s a real page turner…
See more at Evolving Thoughts>>

In philosophy, physicalism is the ontological thesis that “everything is physical”, that there is “nothing over and above” the physical, or that everything supervenes on the physical. Physicalism is a form of ontological monism—a “one substance” view of the nature of reality as opposed to a “two-substance” (dualism) or “many-substance” (pluralism) view. Both the definition of physical and the meaning of physicalism have been debated. Physicalism is closely related to materialism. Physicalism grew out of materialism with the success of the physical sciences in explaining observed phenomena. The terms are often used interchangeably, although they are sometimes distinguished, for example on the basis of physics describing more than just matter (including energy and physical law). Common arguments against physicalism include both the philosophical zombie argument and the multiple observers argument, that the existence of a physical being may imply zero or more distinct conscious entities. “When I lost my belief in religion I had to decide what I needed to accept as a bare minimum. I decided that I needed to believe in the physical world. I never found the slightest reason to accept the existence of anything else. To this day I am a physicalist only because I never found the need to be anything else. The principle of parsimony suggests that one should not believe in more than one needs to. Even if it does make you feel comfortable.”

 

Let’s get physicalism!

See John Wilkin’s Blog ‘Evolving Thoughts‘

#philsci #philosophy #science #physics

On Consciousness, Qualia, Valence & Intelligence with Mike Johnson

/4 Comments/in , , /by

Andrés L. Gómez Emilsson

Andrés Gómez Emilsson joined in to add very insightful questions for a 3 part interview series with Mike Johnson, covering the relationship of metaphysics to qualia/consciousness/hedonic valence, and defining their terms, whether panpsychism matters, increasing sensitivity to bliss, valence variance, Effective Altruism, cause prioritization, and the importance of consciousness/valence research .
Andrés Gómez Emilsson interviews Mike Johnson

Carving Reality at the Joints

Andrés L. Gómez Emilsson: Do metaphysics matter for understanding qualia, consciousness, valence and intelligence? Mike Johnson: If we define metaphysics as the study of what exists, it absolutely does matter for understanding qualia, consciousness, and valence. I think metaphysics matters for intelligence, too, but in a different way. The big question is whether terms like qualia, consciousness, and valence “carve reality at the joints” or whether they’re emergent linguistic constructs that don’t reflect the structure of the universe. And if these things are ‘real’ in some sense, the follow-up question is: how can we formalize these concepts? Intelligence seems different: it seems like a ‘fuzzy’ concept, without a good “crisp”, or frame-invariant, definition. Andrés: What about sources of sentient valence outside of human brains? What is the “minimum viable valence organism”? What would you expect it to look like?

Mike Johnson

Mike: If some form of panpsychism is true- and it’s hard to construct a coherent theory of consciousness without allowing panpsychism- then I suspect two interesting things are true.
  1. A lot of things are probably at least a little bit conscious. The “minimum viable valence experiencer” could be pretty minimal. Both Brian Tomasik and Stuart Hameroff suggest that there could be morally-relevant experience happening at the level of fundamental physics. This seems highly counter-intuitive but also logically plausible to me.
  2. Biological organisms probably don’t constitute the lion’s share of moral experience. If there’s any morally-relevant experience that happens on small levels (e.g., quantum fuzz) or large levels (e.g., black holes, or eternal inflation), it probably outweighs what happens on Earth by many, many, many orders of magnitude. Whether it’ll outweigh the future impact of humanity on our light-cone is an open question.

The big question is whether terms like qualia, consciousness, and valence “carve reality at the joints” or whether they’re emergent linguistic constructs that don’t reflect the structure of the universe. And if these things are ‘real’ in some sense, the follow-up question is: how can we formalize these concepts?

In contrast with Brian Tomasik on this issue, I suspect (and hope) that the lion’s share of the qualia of the universe is strongly net positive. Appendix F of Principia Qualia talks a little more about this. Andrés: What would be the implications of finding a sure-fire way to induce great valence for brief moments? Could this be used to achieve “strategic alignment” across different branches of utilitarianism? Mike: A device that could temporarily cause extreme positive or negative valence on demand would immediately change the world. First, it would validate valence realism in a very visceral way. I’d say it would be the strongest philosophical argument ever made. Second, it would obviously have huge economic & ethical uses. Third, I agree that being able to induce strong positive & negative valence on demand could help align different schools of utilitarianism. Nothing would focus philosophical arguments about the discount rate between pleasure & suffering more than a (consensual!) quick blast of pure suffering followed by a quick blast of pure pleasure. Similarly, a lot of people live their lives in a rather numb state. Giving them a visceral sense that ‘life can be more than this’ could give them ‘skin in the game’. Fourth, it could mess a lot of things up. Obviously, being able to cause extreme suffering could be abused, but being able to cause extreme pleasure on-demand could lead to bad outcomes too. You (Andres) have written about wireheading before, and I agree with the game-theoretic concerns involved. I would also say that being able to cause extreme pleasure in others could be used in adversarial ways. More generally, human culture is valuable and fragile; things that could substantially disrupt it should be approached carefully. A friend of mine was describing how in the 70s, the emerging field of genetic engineering held the Asilomar Conference on Recombinant DNA to discuss how the field should self-regulate. The next year, these guidelines were adopted by the NIH wholesale as the basis for binding regulation, and other fields (such as AI safety!) have attempted to follow the same model. So the culture around technologies may reflect a strong “founder effect”, and we should be on the lookout for a good, forward-looking set of principles for how valence technology should work. One principle that seems to make sense is to not publicly post ‘actionable’ equations, pseudocode, or code for how one could generate suffering with current computing resources (if this is indeed possible). Another principle is to focus resources on positive, eusocial applications only, insofar as that’s possible– I’m especially concerned about addiction, and bad actors ‘weaponizing’ this sort of research. Another would be to be on guard against entryism, or people who want to co-opt valence research for political ends. All of this is pretty straightforward, but it would be good to work it out a bit more formally, look at the successes and failures of other research communities, and so on.

A question I find very interesting is whether valence research is socially disruptive or socially stabilizing by default. I think we should try very hard to make it a socially stabilizing force.

A question I find very interesting is whether valence research is socially disruptive or socially stabilizing by default. I think we should try very hard to make it a socially stabilizing force. One way to think about this is in terms of existential risk. It’s a little weird to say, but I think the fact that so many people are jaded, or feel hopeless, is a big existential risk, because they feel like they have very little to lose. So they don’t really care what happens to the world, because they don’t have good qualia to look forward to, no real ‘skin in the game’. If valence tech could give people a visceral, ‘felt sense’ of wonder and possibility, I think the world could become a much safer place, because more people would viscerally care about AI safety, avoiding nuclear war, and so on. Finally, one thing that I think doesn’t make much sense is handing off the ethical issues to professional bioethicists and expecting them to be able to help much. Speaking as a philosopher, I don’t think bioethics itself has healthy community & dresearch norms (maybe bioethics needs some bioethicsethicists…). And in general, I think especially when issues are particularly complex or technical, I think the best type of research norms comes from within a community. Andrés: What is the role of valence variance in intelligence? Can a sentient being use its consciousness in any computationally fruitful way without any valence variance? Can a “perfectly flat world(-simulation)” be used for anything computational?   Mike: I think we see this today, with some people suffering from affective blunting (muted emotions) but seemingly living functional lives. More generally, what a sentient agent functionally accomplishes, and how it feels as it works toward that goal, seem to be correlated but not identical. I.e., one can vary without the other. But I don’t think that valence is completely orthogonal to behavior, either. My one-sentence explanation here is that evolution seems to have latched onto the

Why we seek out pleasure: the Symmetry Theory of Homeostatic Regulation

property which corresponds to valence- which I argue is symmetry– in deep ways, and has built our brain-minds around principles of homeostatic symmetry. This naturally leads to a high variability in our valence, as our homeostatic state is perturbed and restored. Logically, we could build minds around different principles- but it might be a lot less computationally efficient to do so. We’ll see. 🙂 One angle of research here could be looking at people who suffer from affective blunting, and trying to figure out if it holds them back: what it makes them bad at doing. It’s possible that this could lead to understanding human-style intelligence better. Going a little further, we can speculate that given a certain goal or computation, there could be “valence-positive” processes that could accomplish it, and “valence-negative” processes. This implies that there’s a nascent field of “ethical computation” that would evaluate the valence of different algorithms running on different physical substrates, and choose the one that best satisfices between efficiency and valence. (This is of course a huge simplification which glosses over tons of issues…)
Andrés: What should we prioritize: super-intelligence, super-longevity or super-happiness? Does the order matter? Why? Mike: I think it matters quite a bit! For instance, I think the world looks a lot different if we figure out consciousness *before* AGI, versus if we ignore it until AGI is built. The latter seems to involve various risks that the former doesn’t. A risk that I think we both agree is serious and real is this notion of “what if accelerating technology leads to Malthusian conditions where agents don’t- and literally can’t, from a competitive standpoint- care about qualia & valence?” Robin Hanson has a great post called “This is the Dream Time” (of relaxed selection). But his book “Age of Em” posits a world where selection pressures go back up very dramatically. I think if we enter such an era without a good theory of qualia, we could trade away a lot of what makes life worth living.  
Andrés: What are some conceptual or factual errors that you see happening in the transhumanist/rationalist/EA community related to modeling qualia, valence and intelligence? Mike: First, I think it’s only fair to mention what these communities do right. I’m much more likely to have a great conversation about these topics with EAs, transhumanists, and rationalists than a random person off the street, or even a random grad student. People from this community are always smart, usually curious, often willing to explore fresh ideas and stretch their brain a bit, and sometimes able to update based on purely abstract arguments. And there’s this collective sense that ideas are important and have real implications for the future. So there’s a lot of great things happening in these communities and they’re really a priceless resource for sounding out theories, debating issues, and so on. But I would highlight some ways in which I think these communities go astray.

Computationalism, functionalism, fun theory, ‘hedonic brain regions’, ‘pleasure neurochemicals’, the reinforcement learning theory of valence, and so on all give the illusion of explanatory depth but don’t actually explain things in a way which allows us to do anything useful.

First, people don’t realize how bad most existing models of qualia & valence are. Michael Graziano argues that most theories of consciousness are worse than wrong- that they play to our intuitions but don’t actually explain anything. Computationalism, functionalism, fun theory, ‘hedonic brain regions’, ‘pleasure neurochemicals’, the reinforcement learning theory of valence, and so on all give the illusion of explanatory depth but don’t actually explain things in a way which allows us to do anything useful. Second, people don’t realize how important a good understanding of qualia & valence are. They’re upstream of basically everything interesting and desirable. Here’s what I think has happened, at least in the rationalist community: historically, consciousness research has been a black hole. Smart people go in, but nothing comes out. So communities (such as physicists and LessWrong) naturally have an interest in putting up a fence around the topic with a sign that says

historically, consciousness research has been a black hole. Smart people go in, but nothing comes out. So communities .. naturally have an interest in putting up a fence around the topic with a sign that says ‘Don’t go here!’

‘Don’t go here!’ – But over time, people forgot why the mystery was blocked off, and started to think that the mystery doesn’t exist. This leads to people actively avoiding thinking about these topics without being able to articulate why. Andrés: Is there value in studying extreme cases of valence? E.g. Buddhist monks who claim to achieve extreme sustainable bliss, or people on MDMA? Mike: ‘What science can analyze, science can duplicate.’ And studying outliers such as your examples is a time-honored way of gathering data with high signal-to-noise. So yes, definitely. 🙂
Also see the 1st part, and the 2nd part of this interview series. Also this interview with Christof Koch will likely be of interest.
 
Mike Johnson is a philosopher living in the Bay Area, writing about mind, complexity theory, and formalization. He is Co-founder of the Qualia Research Institute. Much of Mike’s research and writings can be found at the Open Theory website. ‘Principia Qualia’ is Mike’s magnum opus – a blueprint for building a new Science of Qualia. Click here for the full version, or here for an executive summary. If you like Mike’s work, consider helping fund it at Patreon.

Ethics, Qualia Research & AI Safety with Mike Johnson

/2 Comments/in , , /by

What’s the relationship between valence research and AI ethics?

Hedonic valence is a measure of the quality of our felt sense of experience, the intrinsic goodness (positive valence) or averseness (negative valence) of an event, object, or situation.  It is an important aspect of conscious experience; always present in our waking lives. If we seek to understand ourselves, it makes sense to seek to understand how valence works – how to measure it and test for it.

Also, might there be a relationship to the AI safety/friendliness problem?
In this interview, we cover a lot of things, not least .. THE SINGULARITY (of course) & the importance of Valence Research to AI Friendliness Research (as detailed here). Will thinking machines require experience with valence to understand it’s importance?

Here we cover some general questions about Mike Johnson’s views on recent advances in science and technology & what he sees as being the most impactful, what world views are ready to be retired, his views on XRisk and on AI Safety – especially related to value theory.

This one part of an interview series with Mike Johnson (another section on Consciousness, Qualia, Valence & Intelligence). 

 

Adam Ford: Welcome Mike Johnson, many thanks for doing this interview. Can we start with your background?

Mike Johnson

Mike Johnson: My formal background is in epistemology and philosophy of science: what do we know & how do we know it, what separates good theories from bad ones, and so on. Prior to researching qualia, I did work in information security, algorithmic trading, and human augmentation research.

 

Adam: What is the most exciting / interesting recent (scientific/engineering) news? Why is it important to you?

Mike: CRISPR is definitely up there! In a few short years precision genetic engineering has gone from a pipe dream to reality. The problem is that we’re like the proverbial dog that caught up to the car it was chasing: what do we do now? Increasingly, we can change our genome, but we have no idea how we should change our genome, and the public discussion about this seems very muddled. The same could be said about breakthroughs in AI.

 

Adam: What are the most important discoveries/inventions over the last 500 years?

Mike: Tough question. Darwin’s theory of Natural Selection, Newton’s theory of gravity, Faraday & Maxwell’s theory of electricity, and the many discoveries of modern physics would all make the cut. Perhaps also the germ theory of disease. In general what makes discoveries & inventions important is when they lead to a productive new way of looking at the world.

 

Adam: What philosophical/scientific ideas are ready to be retired? What theories of valence are ready to be relegated to the dustbin of history? (Why are they still in currency? Why are they in need of being thrown away or revised?)

Mike: I think that 99% of the time when someone uses the term “pleasure neurochemicals” or “hedonic brain regions” it obscures more than it explains. We know that opioids & activity in the nucleus accumbens are correlated with pleasure– but we don’t know why, we don’t know the causal mechanism. So it can be useful shorthand to call these things “pleasure neurochemicals” and whatnot, but every single time anyone does that, there should be a footnote that we fundamentally don’t know the causal story here, and this abstraction may ‘leak’ in unexpected ways.

 

Adam: What have you changed your mind about?

Mike: Whether pushing toward the Singularity is unequivocally a good idea. I read Kurzweil’s The Singularity is Near back in 2005 and loved it- it made me realize that all my life I’d been a transhumanist and didn’t know it. But twelve years later, I’m a lot less optimistic about Kurzweil’s rosy vision. Value is fragile, and there are a lot more ways that things could go wrong, than ways things could go well.

 

Adam: I remember reading Eliezer’s writings on ‘The Fragility of Value’, it’s quite interesting and worth consideration – the idea that if we don’t get AI’s value system exactly right, then it would be like pulling a random mind out of mindspace – most likely inimicable to human interests. The writing did seem quite abstract, and it would be nice to see a formal model or something concrete to show this would be the case. I’d really like to know how and why value is as fragile as Eliezer seems to make out. Is there any convincing crisply defined model supporting this thesis?

Mike: Whether the ‘Complexity of Value Thesis’ is correct is super important. Essentially, the idea is that we can think of what humans find valuable as a tiny location in a very large, very high-dimensional space– let’s say 1000 dimensions for the sake of argument. Under this framework, value is very fragile; if we move a little bit in any one of these 1000 dimensions, we leave this special zone and get a future that doesn’t match our preferences, desires, and goals. In a word, we get something worthless (to us). This is perhaps most succinctly put by Eliezer in “Value is fragile”:

“If you loose the grip of human morals and metamorals – the result is not mysterious and alien and beautiful by the standards of human value. It is moral noise, a universe tiled with paperclips. To change away from human morals in the direction of improvement rather than entropy, requires a criterion of improvement; and that criterion would be physically represented in our brains, and our brains alone. … You want a wonderful and mysterious universe? That’s your value. … Valuable things appear because a goal system that values them takes action to create them. … if our values that prefer it are physically obliterated – or even disturbed in the wrong dimension. Then there is nothing left in the universe that works to make the universe valuable.”

If this frame is right, then it’s going to be really really really hard to get AGI right, because one wrong step in programming will make the AGI depart from human values, and “there will be nothing left to want to bring it back.” Eliezer, and I think most of the AI safety community assumes this.

But– and I want to shout this from the rooftops– the complexity of value thesis is just a thesis! Nobody knows if it’s true. An alternative here would be, instead of trying to look at value in terms of goals and preferences, we look at it in terms of properties of phenomenological experience. This leads to what I call the Unity of Value Thesis, where all the different manifestations of valuable things end up as special cases of a more general, unifying principle (emotional valence). What we know from neuroscience seems to support this: Berridge and Kringelbach write about how “The available evidence suggests that brain mechanisms involved in fundamental pleasures (food and sexual pleasures) overlap with those for higher-order pleasures (for example, monetary, artistic, musical, altruistic, and transcendent pleasures).” My colleague Andres Gomez Emilsson writes about this in The Tyranny of the Intentional Object. Anyway, if this is right, then the AI safety community could approach the Value Problem and Value Loading Problem much differently.

 

Adam: I’m also interested in the nature of possible attractors that agents might ‘extropically’ gravitate towards (like a thirst for useful and interesting novelty, generative and non-regressive, that might not neatly fit categorically under ‘happiness’) – I’m not wholly convinced that they exist, but if one leans away from moral relativism, it makes sense that a superintelligence may be able to discover or extrapolate facts from all physical systems in the universe, not just humans, to determine valuable futures and avoid malignant failure modes (Coherent Extrapolated Value if you will). Being strongly locked into optimizing human values may be a non-malignant failure mode.

Mike: What you write reminds me of Schmidhuber’s notion of a ‘compression drive’: we’re drawn to interesting things because getting exposed to them helps build our ‘compression library’ and lets us predict the world better. But this feels like an instrumental goal, sort of a “Basic AI Drives” sort of thing. Would definitely agree that there’s a danger of getting locked into a good-yet-not-great local optima if we hard optimize on current human values.

Probably the danger is larger than that too– as Eric Schwitzgebel notes​, ​

“Common sense is incoherent in matters of metaphysics. There’s no way to develop an ambitious, broad-ranging, self- consistent metaphysical system without doing serious violence to common sense somewhere. It’s just impossible. Since common sense is an inconsistent system, you can’t respect it all. Every metaphysician will have to violate it somewhere.”

If we lock in human values based on common sense, we’re basically committing to following an inconsistent formal system. I don’t think most people realize how badly that will fail.

 

Adam: What invention or idea will change everything?

Mike: A device that allows people to explore the space of all possible qualia in a systematic way. Right now, we do a lot of weird things to experience interesting qualia: we drink fermented liquids, smoke various plant extracts, strap ourselves into rollercoasters, and parachute out of plans, and so on, to give just a few examples. But these are very haphazard ways to experience new qualia! When we’re able to ‘domesticate’ and ‘technologize’ qualia, like we’ve done with electricity, we’ll be living in a new (and, I think, incredibly exciting) world.

 

Adam: What are you most concerned about? What ought we be worrying about?

Mike: I’m worried that society’s ability to coordinate on hard things seems to be breaking down, and about AI safety. Similarly, I’m also worried about what Eliezer Yudkowsky calls ‘Moore’s Law of Mad Science’, that steady technological progress means that ‘every eighteen months the minimum IQ necessary to destroy the world drops by one point’. But I think some very smart people are worrying about these things, and are trying to address them.

In contrast, almost no one is worrying that we don’t have good theories of qualia & valence. And I think we really, really ought to, because they’re upstream of a lot of important things, and right now they’re “unknown unknowns”- we don’t know what we don’t know about them.

One failure case that I worry about is that we could trade away what makes life worth living in return for some minor competitive advantage. As Bostrom notes in Superintelligence,

“When it becomes possible to build architectures that could not be implemented well on biological neural networks, new design space opens up; and the global optima in this extended space need not resemble familiar types of mentality. Human-like cognitive organizations would then lack a niche in a competitive post-transition economy or ecosystem. We could thus imagine, as an extreme case, a technologically highly advanced society, containing many complex structures, some of them far more intricate and intelligent than anything that exists on the planet today – a society which nevertheless lacks any type of being that is conscious or whose welfare has moral significance. In a sense, this would be an uninhabited society. It would be a society of economic miracles and technological awesomeness, with nobody there to benefit. A Disneyland with no children.”

Nick Bostrom

Now, if we don’t know how qualia works, I think this is the default case. Our future could easily be a technological wonderland, but with very little subjective experience. “A Disneyland with no children,” as Bostrom quips.

 

 

Adam: How would you describe your ethical views? What are your thoughts on the relative importance of happiness vs. suffering? Do things besides valence have intrinsic moral importance?

Mike: Good question. First, I’d just like to comment that Principia Qualia is a descriptive document; it doesn’t make any normative claims.

I think the core question in ethics is whether there are elegant ethical principles to be discovered, or not. Whether we can find some sort of simple description or efficient compression scheme for ethics, or if ethics is irreducibly complex & inconsistent.

The most efficient compression scheme I can find for ethics, that seems to explain very much with very little, and besides that seems intuitively plausible, is the following:

  1. Strictly speaking, conscious experience is necessary for intrinsic moral significance. I.e., I care about what happens to dogs, because I think they’re conscious; I don’t care about what happens to paperclips, because I don’t think they are.
  2. Some conscious experiences do feel better than others, and all else being equal, pleasant experiences have more value than unpleasant experiences.

Beyond this, though, I think things get very speculative. Is valence the only thing that has intrinsic moral importance? I don’t know. On one hand, this sounds like a bad moral theory, one which is low-status, has lots of failure-modes, and doesn’t match all our intuitions. On the other hand, all other systematic approaches seem even worse. And if we can explain the value of most things in terms of valence, then Occam’s Razor suggests that we should put extra effort into explaining everything in those terms, since it’d be a lot more elegant. So– I don’t know that valence is the arbiter of all value, and I think we should be actively looking for other options, but I am open to it. That said I strongly believe that we should avoid premature optimization, and we should prioritize figuring out the details of consciousness & valence (i.e. we should prioritize research over advocacy).

Re: the relative importance of happiness vs suffering, it’s hard to say much at this point, but I’d expect that if we can move valence research to a more formal basis, there will be an implicit answer to this embedded in the mathematics.

Perhaps the clearest and most important ethical view I have is that ethics must ultimately “compile” to physics. What we value and what we disvalue must ultimately cash out in terms of particle arrangements & dynamics, because these are the only things we can actually change. And so if people are doing ethics without caring about making their theories cash out in physical terms, they’re not actually doing ethics- they’re doing art, or social signaling, or something which can serve as the inspiration for a future ethics.

Perhaps the clearest and most important ethical view I have is that ethics must ultimately “compile” to physics. What we value and what we disvalue must ultimately cash out in terms of particle arrangements & dynamics, because these are the only things we can actually change.

The analogy I’d offer here is that we can think about our universe as a computer, and ethics as choosing a program to run on this computer. Unfortunately, most ethicists aren’t writing machine-code, or even thinking about things in ways that could be easily translated to machine-code. Instead, they’re writing poetry about the sorts of programs that might be nice to run. But you can’t compile poetry to machine-code! So I hope the field of ethics becomes more physics-savvy and quantitative (although I’m not optimistic this will happen quickly).

Eliezer Yudkowsky refers to something similar with his notions of “AI grade philosophy”, “compilable philosophy”, and “computable ethics”, though I don’t think he quite goes far enough (i.e., all the way to physics).

 

Adam: What excites you? What do you think we have reason to be optimistic about?

Mike: The potential of qualia research to actually make peoples’ lives better in concrete, meaningful ways. Medicine’s approach to pain management and treatment of affective disorders are stuck in the dark ages because we don’t know what pain is. We don’t know why some mental states hurt. If we can figure that out, we can almost immediately help a lot of people, and probably unlock a surprising amount of human potential as well. What does the world look like with sane, scientific, effective treatments for pain & depression & akrasia? I think it’ll look amazing.

 

Adam: If you were to take a stab at forecasting the Intelligence Explosion – in what timeframe do you think it might happen (confidence intervals allowed)?

Mike: I don’t see any intractable technical hurdles to an Intelligence Explosion: the general attitude in AI circles seems to be that progress is actually happening a lot more quickly than expected, and that getting to human-level AGI is less a matter of finding some fundamental breakthrough, and more a matter of refining and connecting all the stuff we already know how to do.

The real unknown, I think, is the socio-political side of things. AI research depends on a stable, prosperous society able to support it and willing to ‘roll the dice’ on a good outcome, and peering into the future, I’m not sure we can take this as a given. My predictions for an Intelligence Explosion:

  • Between ~2035-2045 if we just extrapolate research trends within the current system;
  • Between ~2080-2100 if major socio-political disruptions happen but we stabilize without too much collateral damage (e.g., non-nuclear war, drawn-out social conflict);
  • If it doesn’t happen by 2100, it probably implies a fundamental shift in our ability or desire to create an Intelligence Explosion, and so it might take hundreds of years (or never happen).
 

If a tree falls in the forest and no one is around to hear it, does it make a sound? It would be unfortunate if a whole lot of awesome stuff were to happen with no one around to experience it.  <!–If a rainbow appears in a universe, and there is no one around to experience it, is it beautiful?–>

Also see the 2nd part, and 3nd part (conducted by Andrés Gómez Emilson) of this interview series conducted by Andrés Gómez Emilson and this interview with Christof Koch will likely be of interest.

 

Mike Johnson is a philosopher living in the Bay Area, writing about mind, complexity theory, and formalization. He is Co-founder of the Qualia Research Institute. Much of Mike’s research and writings can be found at the Open Theory website.
‘Principia Qualia’ is Mike’s magnum opus – a blueprint for building a new Science of Qualia. Click here for the full version, or here for an executive summary.
If you like Mike’s work, consider helping fund it at Patreon.

Science, Mindfulness & the Urgency of Reducing Suffering – Christof Koch

/2 Comments/in , , , , /by

In this interview with Christof Koch, he shares some deeply felt ideas about the urgency of reducing suffering (with some caveats), his experience with mindfulness – explaining what it was like to visit the Dali Lama for a week, as well as a heart felt experience of his family dog ‘Nosey’ dying in his arms, and how that moved him to become a vegetarian. He also discusses the bias of human exceptionalism, the horrors of factory farming of non-human animals, as well as a consequentialist view on animal testing.
Christof Koch is an American neuroscientist best known for his work on the neural bases of consciousness.

Christof Koch is the President and Chief Scientific Officer of the Allen Institute for Brain Science in Seattle. From 1986 until 2013, he was a professor at the California Institute of Technology. http://www.klab.caltech.edu/koch/

Towards the Abolition of Suffering Through Science

/0 Comments/in , , , , /by

An online panel focusing on reducing suffering & paradise engineering through the lens of science.

Panelists: Andrés Gómez Emilsson, David Pearce, Brian Tomasik and Mike Johnson

Note, consider skipping to to 10:19 to bypass some audio problems in the beginning!!


Topics

Andrés Gómez Emilsson: Qualia computing (how to use consciousness for information processing, and why that has ethical implications)

  • How do we know consciousness is causally efficacious? Because we are conscious and evolution can only recruit systems/properties when they do something (and they do it better than the available alternatives).
  • What is consciousness’ purpose on animals?  (Information processing).
  • What is consciousness’ comparative advantage?  (Phenomenal binding).
  • Why does this matter for suffering reduction? Suffering has functional properties that play a role in the inclusive fitness of organisms. If we figure out exactly what role they play (by reverse-engineering the computational properties of consciousness), we can substitute them by equally (or better) functioning non-conscious or positive hedonic-tone analogues.
  • What is the focus of Qualia Computing? (it focuses on basic fundamental questions and simple experimental paradigms to get at them (e.g. computational properties of visual qualia via psychedelic psychophysics)).

Brian Tomasik:

  • Space colonization “Colonization of space seems likely to increase suffering by creating (literally) astronomically more minds than exist on Earth, so we should push for policies that would make a colonization wave more humane, such as not propagating wild-animal suffering to other planets or in virtual worlds.”
  • AGI safety “It looks likely that artificial general intelligence (AGI) will be developed in the coming decades or centuries, and its initial conditions and control structures may make an enormous impact to the dynamics, values, and character of life in the cosmos.”,
  • Animals and insects “Because most wild animals die, often painfully, shortly after birth, it’s plausible that suffering dominates happiness in nature. This is especially plausible if we extend moral considerations to smaller creatures like the ~1019 insects on Earth, whose collective neural mass outweighs that of humanity by several orders of magnitude.”

Mike Johnson:

  • If we successfully “reverse-engineer” the patterns for pain and pleasure, what does ‘responsible disclosure’ look like? Potential benefits and potential for abuse both seem significant.
  • If we agree that valence is a pattern in a dataset, what’s a good approach to defining the dataset, and what’s a good heuristic for finding the pattern?
  • What order of magnitude is the theoretical potential of mood enhancement? E.g., 2x vs 10x vs 10^10x
  • What are your expectations of the distribution of suffering in the world? What proportion happens in nature vs within the boundaries of civilization? What are counter-intuitive sources of suffering? Do we know about ~90% of suffering on the earth, or ~.001%?
  • Valence Research, The Mystery of Pain & Pleasure.
  • Why is it such an exciting time round about now to be doing valence research?  Are we at a sweet spot in history with this regard?  What is hindering valence research? (examples of muddled thinking, cultural barriers etc?)
  • How do we use the available science to improve the QALY? GiveDirectly has used change in cortisol levels to measure effectiveness, and the EU (what’s EU stand for?) evidently does something similar involving cattle. It seems like a lot of the pieces for a more biologically-grounded QALY- and maybe a SQALY (Species and Quality-Adjusted Life-Year)- are available, someone just needs to put them together. I suspect this one of the lowest-hanging highest-leverage research fruits.

David Pearce: The ultimate scope of our moral responsibilities. Assume for a moment that our main or overriding goal should be to minimise and ideally abolish involuntary suffering. I typically assume that (a) only biological minds suffer and (b) we are probably alone within our cosmological horizon. If so, then our responsibility is “only” to phase out the biology of involuntary suffering here on Earth and make sure it doesn’t spread or propagate outside our solar system. But Brian, for instance, has quite a different metaphysics of mind, most famously that digital characters in video games can suffer (now only a little – but in future perhaps a lot). The ramifications here for abolitionist bioethics are far-reaching.

 

Other:
– Valence research, Qualia computing (how to use consciousness for information processing, and why that has ethical implications),  animal suffering, insect suffering, developing an ethical Nozick’s Experience Machine, long term paradise engineering, complexity and valence
– Effective Altruism/Cause prioritization and suffering reduction – People’s practical recommendations for the best projects that suffering reducers can work on (including where to donate, what research topics to prioritize, what messages to spread). – So cause prioritization applied directly to the abolition of suffering?
– what are the best projects people can work on to reduce suffering? and what to work on first? (including where to donate, what research topics to prioritize, what messages to spread)
– If we successfully “reverse-engineer” the patterns for pain and pleasure, what does ‘responsible disclosure’ look like? Potential benefits and potential for abuse both seem significant
– If we agree that valence is a pattern in a dataset, what’s a good approach to defining the dataset, and what’s a good heuristic for finding the pattern?
– What order of magnitude is the theoretical potential of mood enhancement? E.g., 2x vs 10x vs 10^10x

Panelists

David Pearce: http://hedweb.com/
Mike Johnson: http://opentheory.net/
Andrés Gómez Emilsson: http://qualiacomputing.com/
Brain Tomasik: http://reducing-suffering.org/
‪

 

#hedweb ‪#EffectiveAltruism ‪#HedonisticImperative ‪#AbolitionistProject

The event was hosted on the 10th of August 2015, Venue: The Internet

Towards the Abolition of Suffering Through Science was hosted by Adam Ford for Science, Technology and the Future.

Towards the Abolition of Suffering Through Science

Towards the Abolition of Suffering Through Science

The End of Aging

/0 Comments/in , , /by

Aging is a technical problem with a technical solution – finding the solution requires clear thinking and focused effort. Once solving aging becomes demonstrably feasible, it is likely attitudes will shift regarding its desirability. There is huge potential, for individuals and for society, in reducing suffering through the use of rejuvenation therapy to achieve new heights of physical well-being. I also discuss the looming economic implications of large percentages of illness among aging populations – and put forward that focusing on solving fundamental problems of aging will reduce the incidents of debilitating diseases of aging – which will in turn reduce the economic burden of illness. This mini-documentary discusses the implications of actually solving aging, as well as some misconceptions about aging.


‘The End of Aging’ won first prize in the international Longevity Film Competition *[1] in 2018.


The above video is the latest version with a few updates & kinks ironed out.

‘The End of Aging’ was Adam Ford’s submission for the Longevity Film Competition – all the contestants did a great job. Big thanks to the organisers of competition, it inspires people to produce videos to help spread awareness and understanding about the importance of ending aging.

It’s important to see that health in old age is desirable at population levels – rejuvenation medicine – repairing the bodies ability to cope with stressors (or practical reversal of the aging process), will end up being cheaper than traditional medicine  based on general indefinite postponement of ill-health on population levels (especially in the long run when rejuvenation therapy becomes efficient).

According to the World Health Organisation:

  1. Between 2015 and 2050, the proportion of the world’s population over 60 years will nearly double from 12% to 22%.
  2. By 2020, the number of people aged 60 years and older will outnumber children younger than 5 years.
  3. In 2050, 80% of older people will be living in low- and middle-income countries.
  4. The pace of population ageing is much faster than in the past.
  5. All countries face major challenges to ensure that their health and social systems are ready to make the most of this demographic shift.

The End of Aging – WHO 1 – 2020 portion of world population over 60 will double
The End of Aging – WHO 2 – Elderly outnumbering Infants
The End of Aging – WHO 3 – Pace of Population Aging Faster than in Past
The End of Aging – WHO 4 – 80 perc elderly in low to middle income countries
The End of Aging – WHO 5 Demographic Shifts

 

Happy Longevity Day 2018! 😀

[1] * The Longevity Film Competition is an initiative by the Healthy Life Extension Society, the SENS Research Foundation, and the International Longevity Alliance. The promoters of the competition invited filmmakers everywhere to produce short films advocating for healthy life extension, with a focus on dispelling four usual misconceptions and concerns around the concept of life extension: the false dichotomy between aging and age-related diseases, the Tithonus error, the appeal to nature fallacy, and the fear of inequality of access to rejuvenation biotechnologies.