Hiatus

We will be taking a break until the first week in April, when spring quarter begins. In the meantime, I will try to figure out a good time for everyone to meet and what we would like to take up when we reconvene. Suggestions most welcome.

Have a great end of the quarter and spring break.

Note son our eighth meeting

We continued our discussion of Chalmers' singularity essay, beginning with Patrick's comment on the blog post from last week's meeting.

Patrick’s comment: How are we supposed to conceive of the extensions of intelligence and/or abilities that Chalmers talks about in sec 3?

            The idea is that the AI+(+) is an intelligence of a different kind

The way that AI+ will come about seems deeply dependent on what the abilities are.

One theme in phenomenology: consciousness/the mind is destined for the world—they are tied up in the context in which they make sense. For example, consider a proper functioning view: we get an ability that distinguishes us form animals and that functions properly in a certain context.

But it’s not clear (a) how we can be said to extend these same abilities to new contexts and (b) how these extended abilities might be said to be better.

Success is always success in a context. But we do not have access to the stage relevant to the success of AI+. This is significant because it blocks our ability to predict success relevant to AI++.

A related point (perhaps the same point put another way): the Wittgensteinian idea that our concepts are built for this world, and certain kinds of counterfactuals cannot be properly evaluated because they outstrip the context of our language game

Perhaps: pick a very simple measure for evaluation (e.g., ability to generate wealth, efficiency)

Bergsson: has an argument that every creature is the best example of its kind (Matter and Memory, at the end)

Is there a distinction to be made between a difference in degree and a difference in kind?

Perhaps we are responsible for assigning differences in kind given various differences in degree.

            But does this make the distinction irrelevant or uninteresting?

There are interesting issues here about reality, whether we can experience an objective reality or only ever a subjectively conditioned reality.

Will we ever reach a consensus regarding where to draw the line for a difference in kind? Perhaps, so long as we agree to some background presuppositions—e.g., whether to take a functional perspective or a perspective regarding material constitution.

What constitutes progress?

            Paradigm shifts, death of ideas, (greater or lesser) consensus?

Bostrom (2012) just defines intelligence as something like instrumental rationality

Are bacteria intelligent in the same way as modern AI? Yes, if we define reasoning behaviorally. And this definition of intelligence is easily measurable.

But is it safe to assume that the desire to have power over oneself and one’s environment are prerequisites for success at survival?

            Is this what we think intelligent people have?

All living things modify their internal environment in order to better survive (bacteria, plants, humans, etc.)

Gray goo: a nanobot that builds a copy of itself and the apocalypse comes about because it replicates itself in an uncontrolled fashion, eating all life on earth to feed its end of copying itself.

A problem: We have AI, then pick the capacities we most care about, extend them into AI+, and then the extension to AI++ would no longer be a sort of being we would value. The idea is that the set of things extended comes to include fewer things we care about, to the point that AI++ does not contain anything that we care about.

If we assume that intelligence is instrumental rationality, then this will be ramped up to the exclusion of other interests. But we have a system of interconnected interests—we have cognitive interests, say, in individuating objects in perception. But this might not be maintained in the pursuit of maximizing instrumental rationality.

What does it mean to give a machine values? Give them ends, in the sense relvant to means-ends reasoning.

An argument that a superintelligence might be both moral and extinguish humanity:

Suppose consequentialism is right and AI++ discovers the true conception of well-being. It might be that in order to achieve this they need to wipe out human beings. This would result in a better state of affairs, but extinction for us.

How should we feel about this?

Many of these issues come to a similar problem: The production of an AI++ will involve a loss of some things we find very valuable, and this presents us with a problem. Should we pursue or should we inhibit or constrain the relevant progress in intelligence?

This is probably closely related to Chalmers’ claim that motivational obstacles are the greatest.

What sort of control do we have over the singularity?

            We could delay it, but for how long?

            We could stop it from happening on Earth, say, by blowing up the planet.

We could constrain the ways in which the possibility of the singularity occurring unfolds.

Notes on our seventh meeting

We discussed David Chalmers' "The Singularity: A Philosophical Analysis," which we will continue to discuss next time.

We began by noting Chalmers’ moderate sense of ‘singularity’ (p. 3): referring to an intelligence explosion by a recursive mechanism, where successively more intelligent machines arise that are better at producing even more intelligent machines.

We also noted a nice distinction Chalmers makes (in the spirit of Parfit): identity vs survival

Parfit on Personal Identity: identity doesn’t really matter that much; trying to persuade us to get less attached to notions of identity

Eric Schwitzgebel’s view: it is convenient to have a logic with clean lines between people for us (we don’t fission, duplicate, upload), but in weird cases, this logic does not model well, so should switch to modeling what you care about (e.g., survival).

But practical issues remain (e.g., who pays the mortgage).

Enhancement: much of this has already happened

The Flynn effect: increasing IQs across generations, requires re-calibrating the IQ test to keep the norms in a certain range

There is room for skepticism about measuring general intelligence: (i) perhaps we are better test-takers; (ii) there are multiple intelligences, and IQ-style tests don't test for many (or even most) of them.

In sec 3 of Chalmer's essay: notice the embedding of ‘what we care about’ in the characterization of the relevant capacities. This is in line with the Parfitian approach to identity.

Values: There are many, complex issues here

            How to define them

            How to identify them

            Subjective vs objective

            Universal values (e.g., cross-cultural, across times)

3 different sense of ‘objectivity’ for values: judgment-independent, choice-independent, human nature-independent

Kant vs Hume:

            An issue about whether mistakes in value are mistakes in rationality (Hume: no; Kant: yes).

            And what does this entail about the moral behavior of AI+(+)?

See the new Pinker book: where he argues that we have beome both more intelligent and more moral over time.

Two sense of morality over time: across generations vs. over the course of an individual’s life

            It seems that older people have more sophisticated moral reasoning, but this is a distinct

            question from whether different cultures have more or less sophisticated moral reasoning and

            also from the issue whether one culture is more correct in its moral practices than another.

There are important things that transcend a particular context: e.g., math, logic

            Perhaps the survival instinct is another one

A distinction: one's moral beliefs vs. one's behavior

Another distinction: immortality vs longevity

Obstacles: Chalmers claims that motivational ones are most plausible to stop singularity from coming

            Is this right? Why, exactly, does he think this?

Perhaps there are structural obstacles: the intelligence growth becomes too hard, diminishing returns

Energy needs: can be a situational obstacle, but can also be tied to a motivational obstacle

And when there is a single system, because the energy requirements become greater, this can create a single entity and then it would all depend on its motivation

Some related issues:

Matrioshka brain: concentric circles around the sun, using all the energy, Dyson sphere brain

Kurzweil’s sixth epoch

The Fermi paradox: the odds are not good that we would be the first to reach superintelligence, so we should see evidence of others, but we don’t, so perhaps the process will stall out

Take-home messages from Chalmer's essay:

1.     a broadly functionalist account of the mind, such that we could be instantiated in a computer

              -So long as you have a nomologically possible world, conscious properties go with

              functional properties

2.     the real take-home: there’s a significant enough possibility of something like the singularity that we should seriously worry about it and consider how we are going to handle it

Notes on our sixth meeting

For this meeting, we read two more chapters in Kurzweil's The Singularity Is Near. Our discussion was rather wide-ranging and did not follow the text very closely. But it was interesting nonetheless.

We began with this question: Recall that Vinge distinguishes between AI and IA. In which of these ways does Kurzweil envision the Singularity coming about? That is, does Kurzweil think that the Singularity will arise in combination with our minds (IA), or else as a result of an artificial intelligence we produce (AI)?

The significance of this question has to do with the issue of mind-uploading. Why would we have to upload our minds to the Singularity, as Kurzweil suggested in the reading from last week, if the Singularity arises in combination with our minds?

An Answer: Kurzweil envisions a combination of the two: AI will lead to IA (e.g., Google), which will lead to strong AI in the future, which will then come back and beam us up to the heavens. In any case, the two approaches very much compliment each other.

Kurzweil is suggesting that there will be an AI that is smarter than humans before the uploading. But not certain how it will occur.

Might IA involve uploading in the process of the Singularity coming about? The uploading enters the equation before the Singularity.

What exactly is uploading? A transfer. When a blow to the head no longer matters. A change in substrate. Technically: uploading means that one makes a copy, and then a copy of a copy. Not just plugging in.

One consideration against thinking that Kurzweil envisions a certain version of the IA route to the SIngularity: Kurzweil doesn’t like the single global consciousness idea, because he thinks that it would preclude him being there. He assumes that his individual self would not persist.

This brings up issues about where to draw the boundary of the individual mind: These are salient, not only for the picture where we are plugged in to a growing intelligence that eventually becomes the Singluarity, but also for the picture according to which we are uploaded to a pre-existing Singularity.

How is Kurzweil using the term ‘the Singularity’? And how does this relate to Vinge’s use?: Kurzweil uses the term to refer to an event in human history, not necessarily a particular intelligence that comes into existence, as Vinge does. But Kurzweil does seem to have the arrival of this intelligence in mind.

Kurzweil’s focus on progress in intelligence seems myopic. There have been other periods of advancement in human history that have seen the same pattern of change (perhaps not quite as fast) in different areas of human experience. Why privilege the type of change that interests Kurzweil?

Kurzweil seems to greatly underestimate two things: (1) the limits of technology (need more hardware as well as more code) and (2) the power behind biology (he assumes that technology is better because our chemical synapses slows down our thinking—but there is more going on than just transfer of electrical signals, a trade-off between speed and fine control, also not just signal transfer but also what goes on inside neurons).

Many of the signals required for higher thought don’t transfer info but rather change the way neurons behave—and even the nanobots might not be able to tell us all the ways in which the neurons are functioning

Because of the many complexities to how our brains work, in thought, it may be possible that the robot person might be slower than the human person, even though the robot is faster at transferring electrical signals that carry information. For example, what look like limitations given our biology might be mechanisms that help to achieve optimum speed, given the various operations imolicated in our minds' functioning.

Articles on creating a baby robot (one that they teach):

Stuck on certain tasks: e.g., trying to pay attention to what it is holding, and this is because its eyesight is too good and doesn’t discriminate enough

The key was to make its eyes worse

The process of life as it is may not be the most efficient way to do things, but it is hard to make certain the stronger claim that it is not the most efficient way to do things.

Record to MP3 analogy, or live music to recording analogy: Music recorded on a record (in analog) has no gaps and so has a sound quality that cannot be matched by digital means (e.g., MP3).

Might the new medium be missing some qualitative characteristics of the old medium? And might these be essential to the experience? Can the same be said for different substrates for purported conscious experience?

The challenge is to 'the substrate independence thesis' (e.g., invoked by Bostrom).

Need to be careful: need to be aware if and when nostalgia plays a role in evaluation

Is evolution slow?

            Well it might seem so, only if one assumes that the environment changes slowly

Is there a good distinction to be made between biological advancement/evolution vs technological advancement/evolution?

The main consideration in favor of the distinction is that technological advancement/evolution essentially involves intentions and design by an intelligence. Biological evolution is normally considered to be a 'blind' process in that it is not guided by an intelligent hand.

 

In biology: random mutations give rise to new features, that are more or less adaptable to the environment.

How does the environment influence the mutations?: by changing the rate, but not the kind—they are still random.

What is randomness in this context? Seems to be not by intelligent design.

So “evolution” cannot begin with an intentionally produced mutation

What exactly is evolution?

What is the difference between the other tool using animals and us, such that advancements according to our intentions are of a different category than advancements according to their intentions?

Humans make tools by reproducing things we’ve seen by making them better.

And other animals don’t pass down the acquired knowledge to future generations

In biological evolution: we are talking about the traits of a species.

In technological evolution: can also talk about traits (e.g., a computer having wifi), but then can distinguish between the processes that selected those trait.

There is a different set of useful predictions from intentional vs. unintentional adaptations. We use the label 'biological evolution' in certain contexts, and we use the label 'technological evolution' in another, and this distinction is useful. It is useful to talk about these two processes differently, because it makes certain things easier to discuss: (1) the extreme differences in the observed rates and (2) because of certain other predictions (e.g., the vastly increased capability of tech to make large jumps to break out of local maxima (small change detrimental, but large change possibly beneficial)).

In Darwinian evolution: no such things as revolutions, only evolutions; Darwinian evolution predicts unnecessary/inefficient intermediary steps that are not predicted by technological evolution. And Darwinian evolution is normally considered biological evolution.

The view in favor of the distinction seems to be that technological evolution originates in an intention. But stopping the causal chain at the intention can seem arbitrary from a certain point of view. The intention, after all, may just be a part of the event-cuasal order, and so it will have causes, and they will have causes, and so on. Thus, it seems to be an arbitrary stopping point from the perspective of causal explanation.

Notes on our fifth meeting

We started out with Patrick’s nice comment on the blog about Nietzsche. You can read it, below. This led to a discussion of related issues:

Is the Singularity a continuation of human existence? A particular human’s (i.e., Kurzweil’s) existence?

What constitutes 'fndamental' change? When is a change in degree a change in kind?

Are there limits to human progress and development?

It seems so: we can only think and extend our ideas in a human way, along a restricted range of options. These limits might not be known or knowable to us, but they are there all the same.

But: if we assume that we are essentially limited in certain ways, where do we draw the line? Before vaccines, we might have claimed that we are essentially subject to certain diseases. But now we do not think that.

One clear fundamental difference between humans and the Singularity: the Singularity will not be carbon-based.

But: There still must be matter that is a prerequisite for any existence. This is so, even if the Singularity stands to the matter that underlies it up in a different relation than we stand to the matter that underlies us. (Is 'underlie' the right relation here?)

The Singularity can move through the space of information in a different way than we can move through physical space.

But this does not mean that the relation of Singularity to matter is different than that of human to matter. It seems to be a matter of salience.

Could envision, not the Singularity, but a collection of superhuman consciousnesses

A difference between the relation of the Singularity to its physical instantiation and me to my body: the Singularity can transfer to a different physical instantiation in a way I cannot (when one portion of the computer network goes down, a very different portion can keep the consciousness that is the Singularity going—perhaps even has been all along: multiple, parallel realization).

Recall from the Chomsky piece that there are different conceptions of underlying principles: behaviorism (copying) vs Chomsky (understanding): Perhaps Kurzweil is just using the copying conception. And perhaps he is getting mileage off of trading on the ambiguity between the two  interpretations of ‘capturing underlying principles'.

An objection to the Input/output picture: it treats the mind as a black-box.

Views that call for filling in the black box: don’t need to appeal to a soul.

One might claim that mental states are strongly historical: they do not supervene on mere time-slices of functional organization; allows that physical systems count as minds partly in virtue of their past (cf. Dennett).

This is, perhaps, illustrated by Jason’s sprinx case: one imagines a sprinx thousands of years before evolution creates one. Have I seen a sprinx?

Distinction: the content of a mental state vs. something being a mental state

Less controversial to claim relevance of history to content (content externalism) than to say the same for being a mental state

A claim in physics: the universe is a state function

For any given state, future states can be predicted from it in ignorance of past states

All future time moments would be predicted the same, regardless of past staes leading to the given state

Two issues:

1.     The rise of the Singularity

2.     Its enabling us to achieve immortality

There are many sub-issues for each of these two issues.

Just given a qualitative change in the intelligence, it does not follow that it cannot be us who survive.

In the personal identity literature, there are some who think it is not a matter of whether I continue, but whether there is the right kind of continuity for me to care about the one who continues.

Kurzweil is trying to live as long as he can, so that he can be around for the Singularity in order to achieve immortality

If it is a leap to a new form of intelligence, one that transcends human limitations, then couldn’t be me, because a different form of life. (Perhaps this was Russell’s point from earlier, or in the background of what he said.)

Varia:

A different view of uploading: not me in a computer, but a child of mine in a computer.

A good distinction: logical possibility vs natural possibility

The way the brain works (parallel processing) vs the way the computer processes (logic trees, etc.)

Didn’t the IA Singularity already occur?

Notes on Our Fourth Meeting

I thought our discussion of the Vinge and Moravec pieces was really great. Thank you everyone for such interesting comments and questions. Since we will be continuing with this topic for at least a week or two longer, I hope the discussion continues to excite everyone.

Here are some of the highlights, as I recall, from this past Tuesday:

Both pieces ended on what seemed like different notes: Moravec sounded like something of a mystic or along the lines of a Buddhist or Hindu, with a much more positive slant to what he was saying, whereas Vinge seemed to express a sense of impending doom, or at least a worrisome outlook.

Some questions about motivation: What would the motivation of a superintelligent being (of the sort that the Singularity is characterized to be) be like? Human and animal motivation is shaped in a large part by the need to find food and take care of other basic needs. What about an artifical superintelligence?

Some questions about intelligence: How do we define intelligence? What characteristics are essential for a recognizable form of intelligence (e.g., creativity, inspiration, nostalgia)? Could the Singularity possess these characteristics? In what way is the form of intelligence characteristic of the Singularity supposed to be beyond our ken? The form of intelligence of a mature adult human is beyond the ken of a baby human. Is there supposed to be a difference in the case of the Singularity's being beyond our ken? What is this difference?

Some questions pertaining to our supposed inability to predict what the Singularity would be like:

1.     With a new sort of intelligence, the Turing test won’t apply. What sort of continuity is there between them?

2.     Epistemological claim about our predictions: there will be an event beyond which we cannot predict where things will go. Might the ignorance be connected to question 1?

3.     What makes the Singularity unique? We cannot predict future theories of our own even now. So what’s the difference between the uncertainties we face everyday and the ones this possibility presents?

How is the concept of the singularity already a projection into the future of what we already know? How would we recognize it? Might it already exist, and we don’t know yet?

On some conceptions, the Singularity seems to transcend individuality. Is this a difference between our conception of ourselves as humans and the kind of entity that the Singularity is supposed to be? Does it factor into issues about the desirability of the coming of the Singularity

Why the Singularity might scare us: A future where people aren’t running things anymore is fundamentally different from our present. We might no longer be at the center of things. AI would be scary because has no continuity with our current existence. A future superintelligence might be hostile toward humans.

But is the Singularity to be feared? Would a superintelligence (necessarily, most likely) respect biodiversity, the rights of other creatures, and so on? Would it recognize moral values? WOuld it be a moral exemplar?

The contrast between Aritifical Intelligence (AI) and Intelligence Amplification (IA), in Vinge, was very interesting: Which is the more plausible route to the Singularity? Which is the most desirable, from the perspective of our own well-being as humans? How discontinuous would the Singularity be with human existence if it arose in this way, as opposed to through more traditional AI? Does IA lead to something like a hive-mind or a superintelligence that takes a cue from the Gaia hypothesis?

Would the Singularity (or any other superintelligence) become bored? What characteristics might cause or prevent this? What sort of immortality would it have? What importance does the fact that even a superintelligence has a physical base have with respect to its longevity prospects?

Some different issues:

1.     Could there be a different kind of entity that is super-intelligent?

2.     Could it be immortal?

3.     Could I be immortal in the sense that I have these super-enhanced capabilities?

An irony: Psychology teaches us that those who are deeply religious live longest, so, ironically, the people who live the longest would not believe in a Singularity (on the assumption that this is not something that the religious believe in).

Nietzsche came up a few times: How does he describe the Ubermensch?How does the Ubermensch relate to the Singularity, if at all?

The notion that it might be our function to enable the development of the Singularity also came up: What sense of 'function' is in play here? What does this imply about our relationship to the Singularity (causal, normative)? What about the Singularity's relationship to us (ancestor worship, fuel)?

Notes on Our Third Meeting

We discussed an interview with Noam Chomsky, where he articulated some criticisms of the current state of several fields, including neuroscience, connectomics and AI.

One distinction we drew was between these questions:
1. Is the mimicking of some behavior  produced by statistical analysis ever going to be fully accurate?
2. Even if it were, would it provide us with an understanding of the internal processing of the relevant agent of this behavior?

It seemed that Chomsky's main criticism was of the latter kind. Predicting behavior on the basis of statistical analysis does not provide understanding of why that behavior was actually produced--it does not provide insight into the general principles according to which the relevant system functions and the ways in which those principles are instantiated. (This criticism draws on Marr's three levels for understanding a complex biological organism.)

We also distinguished between these four projects:

1.     Create a machine that performs a function that we used to need humans for.

2.     Create something that can perform the full range of human functions at least as well as humans can.

3.     Create something that does things the way humans do.

4.     Create something that does things exactly the way a specific human does.

Each project is much more difficult than the one that came before it. And each project has an intelligible goal. But we might disagree about the relative values of these goals.

We also talked about the goal of unification in science: a single theory to understand everything.


What is Chomsky's notion of success in science?
It does not seem to be mere predictive success, but rather predictive success across a wide range of contexts. This would rule out the sort of predictive success achieved by big data projects (such as the fictional one he characterizes in the case of the video camera looking out the window) and help to motivate the appeal he finds in a concept of science as seeking to understand the general principles guiding the overt behavior of different systems.

But what about outliers? How do general principles capture their behavior?

What difference does it make to this notion of success whether behavior is instinctual or not?



We also discussed Searle's Chinese Room experiment and the objection it raises for strong AI (and functionalism about the mind/consciousness). [Searle coins "strong AI" to refer to the program of trying to understand our minds by creating artificial intelligences that can behave like us.] Roughly (and we will read this piece later on, hopefully), the idea is that the mere ability to pass the Turing test (to have a behavioral output that is indistinguishable (by other humans) from a human being's) is not sufficient for understanding (or consciousness).
          One observation was that not all AI is "strong AI" in Searle's sense.
         

A question: Might Searle's argument suggestive that predictive success across all contexts may still leave out something essential? That is, might we be able to articulate general principles that govern the mental behavior characteristic of us and be able to specify what areas of the brain instantiate these algorithms and yet still not understand consciousness (say, because consciousness essentially depends on instantiating these algorithms in the type of material that makes up our brains)?