frig

Background Situation: you maintain a heavy internet presence in which you write lots of articles on various "issues" and have attracted a sizable community of hangers-on who like to argue about the issues. The "issues" you write about are things you believe in strongly.

Puzzling Development: you have a fundraiser for the "issues" you care about and are disappointed in the outcome (more on this later, see (!)). You get some amount of donations but the most-visible response is a lot of arguing and discussion, which you (correctly) conclude is counterproductive to the fundraising process.

Possible Conclusion One: there's a sizable community of people who like discussing and arguing about these issues, the same way there's a sizable community of people who enjoy discussing lost, or star trek, or whatnot. The underlying issue isn't that important to them (no more so than lost or trek is important to losties or trekkies), but they sure do enjoy the arguing and discussion.

Possible Conclusion Two: there's a sizable community of people who deeply care about the "issues" at hand, as evidenced by the argumentation and discussion they exhibit. Sadly, the fondness for argumentation leads to counterproductive behavior when (eg) we try to raise funds, b/c the argumentation is public (producing negative feedback loops) but the donations are not public.

"Possible Conclusion One" seems the most warranted; there clearly are people who love arguing and discussion the issues but there's also clearly far fewer of them who care enough to donate (or otherwise do more than post comments on teh blog).

"Possible Conclusion Two" is more flattering to the writer but requires more assumptions about the community of readers and requires drawing more inferences from observed behavior than "Possible Conclusion One".

(!) Note that the reason the author feels disappointed is b/c the author looks @ the amount of donations vis-a-vis the amount of passionate argument the fundraiser provoked and is disappointed (that so much of the passion went into arguing instead of donations). This is tantamount to assuming "Possible Conclusion Two", as it is assuming you can correctly infer actual interest in (doing something about) the "issues" just from apparent passion to argue and discuss ideas pertinent to the "issues".

If you started from the assumption you had attracted a bunch of people who like arguing and for whom your "issues" are a kind of catnip -- the way losties and trekkies find lost and trek to be their catnip -- you might still be disappointed but for a different reason; you'd be disappointed with your ability to get your readers to make the leap from enjoying your work as "entertainment" to caring about the issues it raised, instead of assuming you'd already done that and then bemoaning their seeming inability to act productively wrt those "issues".

- you step out to home depot to buy a $20 hammer

- when you get there "The Flash" (of dc comics fame) figures out you're looking to buy a hammer and buys all the $20 hammers

- he leaves a post-it note saying "if you want a hammer meet me in aisle 7"

- you get there and he offers to sell you a $20 hammer for $21

- either you buy it or you don't; once you've made your decision he returns either (N-1) or (N) hammers back to the store (30 day return policy and all that)

- if you head over to lowe's instead well he is "The Flash"

Not an exact metaphor for "flash trading" (as sketched it's riskless) but it's imho a relevant thought experiment.

At the individual level the winners and losers are pretty obvious. At the systemic level it's unclear any actual efficiency is gained; differentiating between actual "efficiency gains" and "red queen" scenarios is tricky.

Good thought.

There's a handful of companies that essentially do more in-depth versions of these things at like the level of "h1n1 comes to small town america; what emergent behavior patterns do we see in our model" (names escaping me sorry).

There's also a lot of boring-ish economics sims that've been done (names escaping me also, sorry).

To make it fun for the end-user you'd have to "borrow" tilesets from a videogame I think.

You're right though: it's more fun for the person making it than a casual observer.

You put together custom regex-processing silicon (eg: what Tarari used to make) and related hardware (eg: entropy calculators, that kind of thing) and mix in various types of content-addressed memory and you could get some nutty performance #s on specialized tasks like what's discussed there.

A noninfected can see a zombie and recognize that it is the zombified version of some "known" agent.

There's two goals with the communication:

- the first is that it opens up a lot of fun if you have pluggable behavior for noninfected (EG: LIARS!)

- the second is that it forces you to deal with interval arithmetic (or an equivalent scheme)

Basic rules:

-- VISION:

- noninfected can see agents within radius R but can only determine if a seen agent is "zombie / not zombie" within radius r < R

- infected can see agents within radius Z (r < Z < R) and can make "zombie / not zombie" determination instantly

-- VISUAL PERCEPTION:

- both agent types perceive (1) the position of other agents within their field of view and (2) the orientation of those agents (3) the current movement of the agent (movement vector + rotation)

-- SMELL:

- noninfected and infected leave (distinct) scent trails; over time these trails both (1) diffuse and (2) diminish in intensity (say: exponential with a cutoff below which you round to zero)

- noninfected agents have a cutoff N below which they can't smell infected scent

- infected agents have no such cutoff

-- NONINFECTED COMMUNICATION:

- noninfected can communicate about zombie locations in relative terms over a distance C > the visual cutoff. The contents of the message have the form of a list of agents (positions, orientations) relative to the sender's position ("zombie five meters to my 7 o clock, facing my 3 o clock"). The communications are tagged with two fields: a relative position (direction and distance relative to agent's current position) and the communicator identity.

- position rules: if communicating agent is in visual range then relative position is exact; if communicating agent is out of visual range than relative position is a range (22 to 57 degrees), (50-60 meters)

- identification rules: if communicating agent is "known" to receiving agent then receiving agent gets the message tagged with communicating agent's identity. If the communicating agent is not known to receiving agent then the message is tagged "unknown" if it originates outside of visual range and with the communicating agent's identity if it originates within visual range.

- knowing rules: an agent becomes known to another agent if the agent communicates to the other agent within visual range (eg: if you talk to me when I can see you I know your voice now)

-- INFECTION RULES:

- all agents have a variable tracking the amount of zombie toxin in their bloodstream; this ranges from 0 to 100.

- a noninfected agent becomes infected when they cross the '50' level of zombie toxin. At the moment of infection they instantly shoot up to '100' zombie toxin.

- a noninfected agent loses zombie toxin @ a linear rate (with floor @ 0)

- an infected agent loses zombie toxin @ a slow-ish exponential decay; it should take them awhile to decay from

- an infected agent becomes uninfected when they cross the '25' level of zombie toxin.

-- MOVEMENT RULES:

- noninfected agents move more quickly than infected agents

- all agents have a finite maximum turning speed

- noninfected agents move progressively slower the more zombie toxin they have (eg: base speed - (penalty x (ZombieToxin/100))

- infected agents move progressively slower the less zombie toxin they have (eg: base speed + (booster x (ZombieToxin/100))

-- INFECTED BEHAVIOR:

- infected are understood as trying to produce more infected.

- INFECTING: infected who are within some infection radius of a noninfected agent for > some minimum # of timesteps begin to infect the agent with zombie toxin at some rate (perhaps: higher rate the more zombie toxin there is); the agent doing the infecting also gets increased zombie toxin at some fractional rate (eg: if an infected agent adds Z zombie toxin to an noninfected agent's bloodstream then the infected agent gets .7 * Z zombie toxin, also). The infected agen ThInfection "stacks" on agents.

- INFECTING: infection (zombie toxin increasing) stacks on the receiving agent; "multiple zombies can attack the same survivor at the same time"

- INFECTING: zombies can't attack multiple agents at the same time; "a zombie can only be biting one survivor at a time"

- HUNTING: zombies are either "hunting" or "wandering". A zombie "wanders" (random walk) when it isn't doing anything else. A zombie can determine if some other zombie in its field of vision is hunting or wandering.

- HUNTING, LOWEST PRIORITY: if a zombie sees a sufficiently strong and unambiguous zombie scent trail it will go into "hunting" mode and follow the trail until it thins out

- HUNTING, NEXT LOWEST PRIORITY: if a zombie sees a sufficiently strong and unambiguous noninfected scent trail it will go into "hunting" mode and follow the trail until it thins out

- HUNTING, NEXT LOWEST PRIORITY: if a zombie sees other "hunting" zombies it starts chasing the nearest "hunting" zombie

- HUNTING, HIGHEST PRIORITY: if a zombie sees noninfected agent it immediately begins chasing the agent

- WANDERING: a zombie returns to wandering mode when none of the above conditions trigger

-- SURVIVOR BEHAVIOR:

- I'd make this pluggable (and then the fun really begins!) but the general behavior is to try to avoid getting infected, presumably by avoiding contact with zombies (break incoming zombie hordes into clusters with similar movement vectors; project movement vectors forward in time; aim to be elsewhere)

-- MAP:

- some tile-based system with walls and open spaces is a good starting point

Attention to small details is the greater part of friendliness.

I was thinking it'd be also like pdf and that other format adobe's been working on for ebooks (basically similar to pdf but better about reflowing text for different screen sizes) would wind up having fine grained analytics platforms built in (obviously in supported readers only) @ the chapter and paragraph levels or whatnot.

That kind of thing.

Omniture+Adobe leads to: deeply integrated content+analytics, eg the ability to author web or flash content with very fine-grained data capture and analytics.

When you hit a certain age "oh my aching joints" is far more relevant to your interests than "brains might survive motorcycle crash", though the brain stuff is admittedly cooler.

Stem cells may not necessary for soft tissue; if you could inject some time-release hormones or steroids or whatnot you might get improved results vis-a-vis what's out there now.

EG: injecting nanoparticles with suitable coatings as a way of promoting faster joint repair (hips, wrists, shoulders, etc.).

Assume you already have the fancy camera or camcorder.

http://farm3.static.flickr.com/2453/3714584029_6d37f5a64e_o....

Oval Office:

http://graphics8.nytimes.com/images/2009/01/21/us/21obama5-6...

http://en.wikipedia.org/wiki/C*-algebra

The results are better than I would've guessed, but not really so great (this article is pretty degenerate I think, versus more "human-interest" type stories).

One thing I'm curious out: the list of topics is here:

People (7.88) Algebra (4.47) Research Groups (3.41) Software (2.42) Math (2.03) Functional Analysis (1.53) Journals (1.01) Science (0.71) Publications (0.64) Logic and Foundations (0.6)

...does looking at the content give you any sense for why "People" comes in first?

Posted by: quilner | 09/10/09 | 12:28 am Some questions about Nanosolar’s $1-per-watt claim: What happens at night? What happens when its cloudy? What happens during the winter-time? Does Nanosolar’s $1 per watt investment guarantee a consistent 24/7/365 source of electricity? Does Nanosolar’s $1 figure include the cost of installing transmission lines? Does Nanosolar’s $1 figure include the cost laying cement for a power plant, assembling the cells into solar panels, installing the panels, and commissioning the power plant? Does Nanosolar’s $1 figure include the cost of building base-load generating capacity to provide electricity for when the sun doesn’t shine? Installing electrical transmission lines costs about $1 million per mile. Most PV power plants are located far away from demand centers. This sounds like vaporware.

Agree with your general thrust but it's the qualitative change that's the more interesting, as it's perhaps unknowable (in the way that your dog Fido will never understand most of your thoughts, no matter how patiently you explain them).

Better: what are the major arguments for the position that machine consciousness is not possible? What do I think of those arguments? Particularly taking care to distinguish:

- does the argument prove what the people advancing it think it proves?

- do I find myself agreeing with the argument?

Major lines of argument I've heard:

- the metaphysical argument: consciousness derives from having a soul somehow linked to your brain (and thenceforth to your body); the purported impossibility of "machine consciousness" follows from a belief that only people have souls (of the right type, at least)

- the limited-smarts argument: consciously building a conscious machine is beyond the capabilities of a conscious entity (of our type of consciousness)

- the "difference between silicon and wetware" argument: this ranges from assumptions there's quantum magic in parts of the brain to assumptions that the brain organization implements some other, unsimulable-by-silicon computing architecture (perhaps super-turing)

- the "consciousness is an illusion" argument: consciousness and intelligence per-se have little to do with each other despite the apparent overlap one perceives from reflection on one's own thoughts. Thus machine intelligence seems to be possible but that says nothing about consciousness per se.

If you're really serious enough about this to consider making it a life's work (or because you really want to make a conscious AI) I would suggest taking none of the above arguments lightly, even though it's somewhat fashionable in some circles to assume most of the above are just nebulous handwaving by anti-rational mystics.

"Taking them seriously" doesn't mean "pack your bag and go home"; it means you keep thinking scientifically and analytically and try to answer questions like:

- suppose smarts are limited, but we don't know that yet for certain. What could this intuition mean (in a more formal or more precise sense)? How could I make the intuition more precise? What would a formal proof of the intuition look like (and what would be the theorem)? Does this inquiry seem to be leading me in the direction of possible theorems or nontrivial facts about the expressive power of symbolic systems (that aren't already known, or a retread of Godel)? Does there still seem to be work I'd be interested in doing in this general field if it turns out that smarts are limited?

...as even "wrong" counterarguments can do wonders for pointing you in the direction of interesting questions

Extra Credit: of the researchers who at least seemed to think they'd solve the problem pretty quickly, is there a recurring pattern to be found in the failures those researchers encountered?

Major recurring themes: people who've thought they were within reach of making conscious machines typically assumed that the part of their own nature that they valued most was the keystone to consciousness and assumed everything else was either secondary or easy (and thus could be filled-in later).

Thus a Hofstadter-type -- who loves delving deep into various piles of work and crafting new and insightful analogies -- winds up thinking that the core capability an artificial intelligence needs is the ability to craft such analogies; people with phds in mathematical fields and a more logical bent assume that the core ability is symbolic inference, and make software that does symbolic inference; yet others assume that rational hedonism is the core and work on utility-maximizing decision-theoretic planners and agents; others still love making systems with complicated interactions and seeing what pops out and start chasing emergence.

All of that work is good work and has found applications, but the dynamic is obvious: people who get into the field with the specific goal of making AI -- instead of, say, improving algorithms for multicamera view synthesis with applications to industrial quality control -- tend to radically overvalue whatever intellectual style they happen to be good at, but so far none of those intellectual styles appears to have really scratched the surface of consciousness in the sense you're interested in. Beware your best ideas and favorite subjects!

How much of marine philosophy and organizational style makes it into your consulting work?

I had a similar experience at one point, though in different circumstances and a different organization. Hardscrabble, self-bootstrapped types tend to try and solve everything on their own b/c that's just how they've done.

It's good to have that backbone but can be sabotaging if it's your only strategy; the frame of mind it puts you in is intrinsically adversarial and limiting -- how do I do X for Y despite Z? -- instead of collaborative -- Y wants me to do X but there's Z...can Y and I figure something out?

People from more privileged backgrounds generally are more comfortable taking the latter approach; the behavior can be learned, but once you do you realize how much ground you needless lost, which is sobering.

Bravo for the excellent anecdote.

(a, minor) the dalvik vm is a closer fit to the underlying cpu (being a register-based vm iirc) which makes it simpler to implement and/or a little more performant

(b, major) the dalvik vm (contra the jvm) is designed to more-easily support the "single vm instance with multiple logically independent apps running in it" model; this reduces the vm-induced memory overhead b/c the vm and any system-wide libraries are only loaded once.

There's probably also some truth to the trademark argument.

I'm still not sure you understand what I was trying to say; regardless of whether your take on realpolitik is correct or not (and I don't have a dog in that hunt) it's effectively a non-sequitur to the discussion.

EG: If I were saying that I can't make it downtown by 5 o'clock because:

- it's too far to bike in the available time

- my car is in the shop

- and I missed the last train

...your statement would be something like:

- trains have uncomfortable seats

...which may be true or false but doesn't really have much to do with the previous thrust of the conversation, does it?

The relationship between your point (which you've made twice now) on "realpolitik" and the posts of mine that mention "realpolitik" is well-nigh isomorphic to the example I just sketched. If you don't believe this I will be happy to make the analogy more explicit.

As this has now happened twice I can't tell which of the following holds:

(i) you understood the analysis I made of the politician's situation but don't care to comment on it; you'd rather discuss what you think of realpolitik. Given the tangential nature of realpolitik to the point I was making this leaves me confused.

(ii) you didn't really get what I was writing but saw a few words in there you have opinions on and so to continue the conversation offered your opinions on those words independent of any consideration as to the structural role those words played in the point I was making.

My underlying moral intuitions as relevant to the specific situation at hand are as follows:

- generally speaking I don't expect real-world moral dilemmas to have parties unambiguously in the right and in the wrong. It is convenient when things work out that way but not to be expected; what is to be expected is that a real-world moral dilemma will have convincing arguments on all sides. Consequently making a moral judgment is not playing "spot the villain!" but, instead, deciding which position has the most merit; that a particular position has the most merit says nothing about the merits (or lack thereof) of the other positions other than that what merit the other positions have is less than the position with the most merit. tl;dr: moral dilemmas involve hard choices.

- generally speaking mortal peril -- the threat of losing one's life -- is a significant circumstance that effects how I judge someone's actions. For example: I think murder is wrong, but killing in self-defense is not wrong; most justice systems of the world recognize this principle, and I think they do so correctly.

- there's nothing magical about contracts that exempts them from the previous point: when one party to a contract can only fulfill that contract at cost of that party's own life, the situation demands heavy scrutiny; expecting someone to place adherence to past agreements above self-preservation fails most common-sense morality tests (and is for example why killing in self-defense is considered differently from murder, and why bank employees aren't usually charged with theft when a bank robber sticks a gun in their face). Just having your life at risk isn't carte-blanche to do anything you want; it's simply a relevant fact that colors how I judge a person's decision to renege on past agreements

- in the specific case of an individual suffering from a terminal illness with a single, patented cure who has tried-and-failed in good faith to negotiate a hardship discount from the patent holder I see nothing morally wrong in the individual using their own materials and knowledge to produce the patented substance thereby saving their own life (risking what consequences may come). This analysis is equal parts general moral principles -- not expecting people to go to their own graves for the general benefit of society -- and the specifics of the situation at hand (ignoring patents: fine; stealing stuff: not so fine)

- (unsurprisingly) I also think that in the case of third-world nations facing national epidemics that a similar moral conclusion obtains: if after a good-faith effort to negotiate a price the nation can afford (where "afford" means "purchase in enough quantities to halt the epidemic / mass suffering / etc") then the nation cannot be faulted for ignoring the patents and self-manufacturing the substances in question (regardless of whether production is nationalized or farmed out to private-sector actors); I don't expect the patent holders to like this -- and I don't think they're wrong to dislike it -- but I also don't see anything wrong with the state's actions in that case. The morality-of-the-state does not immediately follow from the individual situation but I think that it does follow insofar as you view the state as obligated to advance its citizens' aggregate interests. I can understand why one might believe the state would better advance its citizens' aggregate interests by respecting the patent-holder's rights, and have already explained in great detail why I don't take that position.

These are relevant moral intuitions; can you state yours?

The request that you either:

- explain how both ditches and knowledge share an identical classification along both of those axes (excludability and rivalry)

- or explain not only that your little theory doesn't recognize distinctions about the above but also why your little theory ignores the distinction

...was an invitation to stop carping and actually state your premises. You declined, and just huffed that "duh they're different but I don't care".

It's also interesting that your response to the fact that it's not the actual beneficiary who is using the knowledge, but a speculator, was to invoke realpolitik.

One problem with that is that the means produce the ends. The other is that the speculator doesn't have the characteristics that you use to justify your conclusion, but your conclusion is unchanged.

You've per usual mischaracterized what I said.

Anatoly raised a valid point: the specific scenario I drew up isn't actually an accurate match for the scenario at hand, and explained that the moral calculations would probably come out differently in a situation that more closely resembled the situation on the ground.

Unlike you, he very carefully sketched out his own situation, premises, and conclusions, and pointed out explicitly where our views of the world diverged. He then asked what the moral choice would be for Bob -- whether Bob ought to shut down Jim's off-brand pharma factory or whether Bob ought to let Jim stay in business.

Given such a well-formed approach I felt it warranted a proper reply.

I first explained why as a practical matter if Bob is an elected representative -- and thereby accountable to his constituents in some fashion -- the incentives for Bob almost certainly tip in favor of Bob's letting Jim continue.

One thing I admittedly elided somewhat was the relationship between "Bob's incentives as his constituent's elected representative" and "Bob's moral obligations as his constituent's elected representative"; there are entire fields of political philosophy that try to nail that down and getting dragged into it in this context is pretty much a derail.

Suffice to say that yes: Bob is clearly incentivized to allow Jim to continue operating due to the situation Bob is in -- whether Bob is being moral to do so -- and (unstated but implied) to the extent you think that the moral balance generally tips in favor of saving lives now (at the expense of ignoring some patents) instead of preserving the integrity of the patent system (at the expense of deaths) then Bob is, additionally, morally correct to act on those incentives.

I then pointed out that that analysis only applied if Bob was an elected representative.

The only mention of realpolitik was in one of the explanations for why Bob would be incentivized to allow Jim's explanation to continue; it was in the context of a hypothetical defense he might offer to his constituents after he shut down Jim's factory. The realpolitik-derived explanation Bob could offer was that even though he might personally approve of Jim's operation allowing the operation to continue would jeopardize other trade agreements, with worse consequences than whatever fallout derived from shutting down Jim.

I then explained this was a politically weak explanation, as it made Bob look lame for having agreed to such crappy trade terms to begin with; given that it's a weak explanation that makes Bob look bad he'd be incentivized not to wind up in a situation in which he'd have to proffer that defense.

If what you got from that response was "realpolitik! that's another word I don't like" I give up.

This is where we went off the tracks:

Yes, my example is different. I'm using that difference to demonstrate that your argument fails.

From here on out I've been saying:

- (i) your example is sufficiently different from mine that to me it proves nothing about mine (though I'm sure it's an example you know how to "win" arguments with)

- (ii) here's why I consider your example sufficiently different from mine as to no longer be relevant

- (iii) please explain why I'm wrong -- why I ought to consider your example and my example equivalent

...at which point to my eyes the rest of our "conversation" reads as you re-iterating that you prefer your example to mine blah blah blah.

And it's worth noting you can't even summarize my scenario accurately, which gives me little faith in your reading ability:

* You think that should be able to benefit from work that he does before you have a need but not from work that he does after you find yourself in need.*

...is wrong; the work was finished well before I was in need...regardless of whether I die or not he's doing no further work, full stop.