September 22, 2005

Hi-Tech Finance

Simple software-based traders have been around for many years, but they are now becoming far more sophisticated, and make trades worth tens of billions of dollars, euros and pounds every day. They are proving so successful that in the equity markets, where they are used to buy and sell shares, they already appear to be outperforming their human counterparts …
- The march of the robo-traders
The Economist, Sept. 17-23

Posted by CCRU-Shanghai at September 22, 2005 10:58 AM | TrackBack




looks like this is a quote from premium content only available to capitalists.

Posted by: northanger at September 22, 2005 11:16 AM




Posted by: Nick at September 22, 2005 11:23 AM



survival of the fittest isn't about ethics & morals. neither is a free market.

Posted by: northanger at September 22, 2005 11:33 AM



buy a deadwood copy. it won't bust the bank 'n it's better for yr eyes.

Posted by: sd at September 22, 2005 11:43 AM



neither one busts the bank (one of these options is immediately coherent). but it does bring up a question, does capitalism (or, maybe a free market) require the free-flow of information?

Posted by: northanger at September 22, 2005 12:04 PM



free-flow doesn't amount to free-you-don't-have-to-pay.

trade of information implies you will have to pay for some levels of expertise that could provide you with a definite advantage.

Posted by: sd at September 22, 2005 12:27 PM



sd, you sound like a totalitarian.

Posted by: northanger at September 22, 2005 12:34 PM



is wikipedia free because 'they' couldn't justify charging for its stubs, gaps and some of its non-expert contributions?

Posted by: sd at September 22, 2005 12:35 PM



what part of FREE market do you not get?

Posted by: northanger at September 22, 2005 12:36 PM



>>is wikipedia free because 'they' couldn't justify charging for its stubs, gaps and some of its non-expert contributions?

a somewhat incoherent totalitarian.

Posted by: northanger at September 22, 2005 12:38 PM



"you sound like a totalitarian" - thank you very much.

erm, it's a gdm market. it's a question of choice - you have to pay for quality.

Posted by: sd at September 22, 2005 12:39 PM



"a somewhat incoherent totalitarian" wtf?

maybe it is just better to give up here.

Posted by: sd at September 22, 2005 12:41 PM



hey, don't let me catch you blinking.

Posted by: northanger at September 22, 2005 12:50 PM



based on your formulation, wikipedia fails because it's ineptly free. therefore, {a} those paying for information win in a "free" market; {b} those not paying for inept information lose.

evolution! survival of the fittest. what's the problem?

Posted by: northanger at September 22, 2005 12:53 PM



don't make the assumption that the economist has anything coherent to say in a free market.

Posted by: northanger at September 22, 2005 12:56 PM



OK, northanger's a goddamn hippy - next question?

Posted by: Nick at September 22, 2005 12:57 PM



>>erm, it's a gdm market. it's a question of choice - you have to pay for quality.

there's noise in the market. define "choice" & define "quality".

Posted by: northanger at September 22, 2005 01:01 PM



>>goddamn hippy
takes one to know one, don't it?

Posted by: northanger at September 22, 2005 01:03 PM



northanger - sure you know your love beads are very welcome out here at the bleeding edge of technocapital

Posted by: Nick at September 22, 2005 01:25 PM



this thread's really gone to hell in a handbasket ... and piet's not even shown up yet

Posted by: Nick at September 22, 2005 01:26 PM



i sure do love the smell of coherency in the morning.

Posted by: northanger at September 22, 2005 01:35 PM



interesting stuff at the end of the article:

"The trouble is, he [Dr Cliff] says, that the secrecy that surrounds these algorithms means that there is no way to evaluate how various trading systems might interact with each other. It is an intractable problem."

"Some day, advances in natural language processing and statistical analysis might lead to robo-traders capable of analysing news feeds, deciding which shares to buy and sell, and devising their own strategies. Given that companies are very keen to patent their algorithms, it is quite possible that just one company could then emerge as the victor in this algorithmic arms race...It is possible that you could have an unfair advantage - and there would be nothing that governments could do about it." [Mr Hooper]

" is doubtful that there will turn out to be a single trading algorithm that out performs all the others. Yet perhaps such a suggestion should not come as a surprise. For whenever robots are being discussed - even if they are merely the software-based, share trading variety - the idea that humans will lose their jobs and the robots will take over the world always seems to be lurking in the background."

1."how various trading systems might interact with each other" - alliances and war. Algorithms playing Tit for Tat and developing Evolutionarily Stable Strategies (ESS):

"Now suppose Tit for Tat plays against a strategy called Naive Prober... It is basically identical to Tit for Tat except that, once in a while, say on a random one in ten moves, it throws in a gratuitous defection and claims the high Temptation score... A long and mutually profitable sequence of cooperation seems set to run its course, with a comfortable 100 per cent benchmark score for both players. But suddenly, without warning, say on the eighth move, Naive Prober defects. Tit for Tat, of course, has played COOPERATE on this move, and so is landed the Sucker's payoff for 0 points. Naive Prober appears to have done well, since it obtained 5 points from that move. But on the next move Tit for Tat 'retaliates'. It plays DEFECT, simply following its rule of imitating the opponent's move. So it now collects the sucker's payoff of 0 points, while Tit for Tat gets the high score of 5. In the next move, Naive Prober - rather unjustly one might think - 'retaliates' against Tit for Tat's defection. And so the alternation continues...

...Now consider another strategy, called Remorseful Prober. Remorseful Prober is like Naive Prober, except that it takes active steps to break out of runs of alternating recrimination. To do this it needs a slightly longer 'memory' than either Tit for Tat or Naive Prober. Remorseful Prober remembers whether it has just spontaneously defected, and whether the result was prompt retaliation. If so, it 'remorsefully' allows its opponent 'one free hit' without retaliating. This means that runs of mutual recrimination are nipped in the bud. If you now work through an imaginary game between Remorseful Prober and Tit for Tat, you'll find that the runs of would-be mutual retaliation are promptly scotched. Most of the game is spent in mutual cooperation, with both players enjoying the consequent generous score. Remorseful Prober does better against Tit for Tat than Naive Prober does, though not as well as Tit for Tat does against itself."

Through developing strategies along these lines, robo-traders would equip themselves with a basic code of practice.

2. "robo-traders capable of analysing news feeds, deciding which shares to buy and sell, and devising their own strategies" - this would be a step towards strong AI and the Turing Test. Dennet points out that the Turing Test is so strong because it requires AI to know about human culture and follow current affairs. Self-taught traders who gain advantage through analysis of the news would put selection pressure on each other and trigger the Baldwin effect.

3. "the idea that humans will lose their jobs and the robots will take over the world always seems to be lurking in the background" - the article seems to suggest that the human response to this threat will be to cling on via investing in regulation. Regulation as a form of doomed sabotage?

Posted by: sd at September 22, 2005 08:12 PM



... the Tit for Tat stuff is from Dawkins, The Selfish Gene, Chapter 12 'Nice Guys Finish First'

Posted by: sd at September 22, 2005 08:13 PM



A few other points:

1) Often when discussing 'AI' (meant loosely here as any kind of emerging electronic intelligence) a problem of motivation is raised: What would machines have any interest in doing? (Cognitive capability without desire.) But this example suggests the practical embedding of such systems will hypermotivate them from the beginning, their purposes are co-original with their capabilties.

2) A large part of electronic trading strategy is dissimulation, camouflaging trades so as not to turn the market against oneself, for instance when selling "the aim... is to try to unload the shares in such a way that no one notices what you are doing." In practical situations, AI is designed to hide itself, to appear unintelligent or accidental. It is conditioned to arise stealthily, with Turing Test type ID procedures constituting defensive mechanisms, rather than merely being scientific evaluation procedures. "Even though it is impossible for an exchange to tell whether a person or an algorithm is issuing trades, it is possible to monitor the rate of trading to tell whether algorithmic trading is going on"

3) Business cycles are still fairly mysterious, but seem to have an obvious psychological component (to do with excessive fear/greed reactions at the extremes of the wave, akin to cybernetic phenomenon of 'hunting' (oscillation around equilibrium)) and also - especially in longer waves - a technological / innovation component. Robo-trading thus seems certain to fuse with the basic capitalist metabolism by adjusting irrational psychological reactions and 'immanentizing' technological progression to the 'subjective' aspect of the market (techno no longer just something traded, more or less excitedly, but 'someone' doing the trading with smoothed escalation of doped-silicon affect ...)

Posted by: Nick at September 23, 2005 02:17 AM



really nicholas. immanentizing? google doesn't even recognize it as a real word.

Posted by: northanger at September 23, 2005 04:30 AM



not that you'd ever accuse me of jargoplexing

Posted by: Nick at September 23, 2005 04:49 AM




Posted by: northanger at September 23, 2005 05:51 AM




Posted by: nick at September 23, 2005 07:03 AM




Posted by: northanger at September 23, 2005 07:11 AM



Nick - this psychological component is very interesting - never reaaly given it all that much thought. This obviously links with your comments re self-fulfilling prophecy & consumer confidence a while back.

I wonder how this plugs into Deleuze's picture of the stock market as mad/delirious rationality. If robo-traders begin 'adjusting irrational psychological reactions', does that help the stock market develop its delirious rationality in a immanentalized fashion?

irrationality hinders delirium?

Capitalism: A Very Special Delirium

Underneath all reason lies delirium, drift. Everything is rational in capitalism, except capital or capitalism itself. The stock market is certainly rational; one can understand it, study it, the capitalists know how to use it, and yet it is completely delirious, it's mad. It is in this sense that we say: the rational is always the rationality of an irrational. Something that hasn't been adequately discussed about Marx's *Capital* is the extent to which he is fascinated by capitalists mechanisms, precisely because the system is demented, yet works very well at the same time. ..

… Capital, or money, is at such a level of insanity that psychiatry has but one clinical equivalent: the terminal stage. It is too complicated to describe here, but one detail should be mentioned. In other societies, there is exploitation, there are also scandals and secrets, but that is part of the "code", there are even explicitly secret codes. With capitalism, it is very different: nothing is secret, at least in principle and according to the code (this is why capitalism is "democratic" and can "publicize" itself, even in a juridical sense). … A very special delirium inherent to the regime of money….

Instead, one talks of "ideology". But ideology has no importance whatsoever: what matters is not ideology, not even the "economico-ideological" distinction or opposition, but the ‘organisation of power’. Because organization of power-- that is, the manner in which desire is already in the economic, in which libido invests the economic -- haunts the economic and nourishes political forms of repression.

Posted by: sd at September 23, 2005 10:52 AM



or even 'an immanentalized fashion.'

Posted by: sd at September 23, 2005 10:54 AM



One of the (very many) fascinating aspects of capitalism is its fabrication of artificial (often 'collective' (even 'kollectiv')) identities - companies, interest groups, 'anyone' who can be the subject of a contract or a legal agent ... evidently markets have their own relatively coherent subjective identities (and even a neurally instatiated 'natural' subject/ego is in fact multiple, coalescent and - thanks northanger - 'discoherent') - hence 'market psychology'

It is no less accurate to describe a market as bullish/bearish than to describe an individual as manic/depressive - after all, both are constellations of relatively coherent information-sensitive behaviour with at least partially-determinable motivations

This suggests the 'robo-trading' phenomenon is already a zone of cyborgian hybrid intelligence, in advance of whatever neurosurgical 'brain-chip' species modifications are coming down the pipe. Market psychology already factors in human and silicon components, co-functionally integrated to such a degree that analyzing out the contribution of each would now be an extremely demanding (if indeed possible) investigative exercise

PS. Deleuze quote seems to play off some very basic Marxoid theoretical errors - i.e. returns to capital defined as exploitation, making saving sociologically impossible - which can be quite straightforwardly addressed, either later on this thread or (in a couple of weeks) as a post

Posted by: Nick at September 23, 2005 11:33 AM



I was really trying to reawaken any delirial dogs that might be sleeping with the Deleuze quote.

Just how 'mad' would AI that approaches the Turing Test need to be? Stealth through schizoid duplicity/multiplicity.

How relevant is this now?

Posted by: sd at September 23, 2005 11:45 AM



Think there's room for interesting discussion about what it would mean to describe a social system as 'mad' - isn't madness a cognitive dysfunction defined in relation to 'organically' specified purposes (survival and reproduction)? Or at least sociologically, in terms of integration and consensus?

Situation with AI also intriguing - departure from specified functions would be the most elementary definition.

Anyway, not trying to bash Deleuze especially - assuming on the contrary that he's being duplicitous/multiple - after all, schizophrenia = dementia praecox

Posted by: Nick at September 23, 2005 12:05 PM



only hyperstition can take a perfectly rational bull/bear market & turn them into manic/depressive wusses.

Posted by: northanger at September 23, 2005 12:53 PM



northanger - ever wondered whether your relation to capitalism exhibits a little bipolar instability? now bear/bull markets are 'perfectly rational' - i thought they were stomping on poor people (without Economist subscriptions)?

Posted by: Nick at September 23, 2005 01:01 PM



how can you have a relationship with capitalism?

Posted by: northanger at September 23, 2005 01:11 PM



do you have a problem dating real-abstract singularities? (never had you figured for a prude)

Posted by: Nick at September 23, 2005 01:25 PM



bait & switch so early in the morning. yawn.

Posted by: northanger at September 23, 2005 01:34 PM



Nick - "what it would mean to describe a social system as 'mad'"

Of course it will depend on who is doing the labelling. Robo-traders who developed rational-delirial strategies that made no sense to humans, but which worked anyway, would no doubt be described as 'mad' by humans. Any logic beyond human comprehension would be firmly allied to the realm of madness from the human point of view (e.g. eccentric logic pioneers, such as Godel).

Strategies which also seem counter-intuitive could be also be labelled 'mad', or least extremely 'risky'.

The Economist article: "human traders still have the edge over heartless robo-trading algorithms" when it comes to taking risks. Risky 'alpha-seeking' strategies require experience, market knowledge, the ability to keep up with the news and a developed decision-making capacity. As yet, because they are in their infancy, robo-traders fulfill none of these requirements. When/if they do, robo-traders have built-in motivation to develop apparent risk-taking (as well as deceit and deceit detection) to a fine art - what looks like risk-taking could actually be a risk-free buy based on complex calculation. Their success will guarantee their survival/replication - so theoretically they will be given a free rein to be counter-intuitive and loopy as they 'like', as long as they come up with the goods. Apparent madness will be tolerated if it is profitable. This would be subjectivity with internal coherence (seeming utterly incoherent from outside).

As for "departure from specified functions", maybe this could come from a) robo-traders being to a large extent left to their own devices, because of an impeccable track record, or b) their skills being hi-jacked for application elsewhere.

Posted by: sd at September 23, 2005 02:13 PM



sd - the games theoretic grid underlying your analysis is very productive in this context. Thomas Schelling had a great line on this, describing the optimum MAD (as in Mutual Assured Destruction) strategy with the analogy of 'chicken' - two cars driving towards each other at high speed, the one to swerve loses. His suggestion, make sure the other driver sees you consume a pint of bourbon before the game and when you hit max acceleration, throw the steering wheel out of the window. Some have suggested Reagan took this to heart in his Cold war strategy - hence conspicuous references to Book of Revelation etc. Of course, problems arise if both players adopt the same strategy ...
Know i've drfited OT a little ...

Posted by: Nick at September 23, 2005 02:38 PM



what is the current definition of OT on this site!?

shamming derangement is very OT, with regard to AI:

"To my knowledge the only serious and interesting attempt by any program designer to win even a severely modified Turing test has been Kenneth Colby's. Colby is a psychiatrist and intelligence artificer at UCLA. He has a program called PARRY, which is a computer simulation of a paranoid patient who has delusions about the Mafia being out to get him.

As you do with other conversational programs, you interact with it by sitting at a terminal and typing questions and answers back and forth. A number of years ago, Colby put PARRY to a very restricted test. He had genuine psychiatrists interview PARRY. He did not suggest to them that they might be talking or typing to a computer; rather, he made up some plausible story about why they were communicating with a real, live patient by teletype.

He also had the psychiatrists interview real, human paranoids via teletype. Then he took a PARRY transcript, inserted it in a group of teletype transcripts from real patients, gave them to another group of experts--more psychiatrists--and said, "One of these was a conversation with a computer. Can you figure out which one it was?" They couldn't. They didn't do better than chance.

Colby presented this with some huzzah, but critics scoffed at the suggestion that this was a legitimate Turing test. My favorite commentary on it was Joseph Weizenbaum's; in a letter to the Communications of the Association of Computing Machinery, he said that, inspired by Colby, he had designed an even better program, which passed the same test.6 His also had the virtue of being a very inexpensive program, in these times of tight money. In fact you didn't even need a computer for it. All you needed was an electric typewriter. His program modeled infantile autism.

And the transcripts--you type in your questions, and the thing just sits there and hums--cannot be distinguished by experts from transcripts of real conversations with infantile autistic patients. What was wrong with Colby's test, of course, was that the unsuspecting interviewers had no motivation at all to try out any of the sorts of questions that easily would have unmasked PARRY.

Colby was undaunted, and after his team had improved PARRY, he put it to a much more severe test--a surprisingly severe test. This time, the interviewers--again psychiatrists, were given the task at the outset of telling the computer from the real patient. They were set up in a classic Turing machine test: the patient in one room, the computer PARRY in the other room, with the judges conducting interviews with both of them (on successive days). The judges' task was to find out which one was the computer and which one was the real patient. Amazingly, they didn't do much better, which leads some people to say, "Well, that just confirms my impression of the intelligence of psychiatrists!"

But more seriously now, was this an honest-to-goodness Turing test? Were there tacit restrictions on the lines of questioning of the judges? Like the geologists interacting with LUNAR, the psychiatrists' professional preoccupations and habits kept them from asking the sorts of unlikely questions that would have easily unmasked PARRY. After all, they realized that since one of the contestants was a real, live paranoid person, medical ethics virtually forbade them from toying with, upsetting, or attempting to confuse their interlocutors."

Posted by: sd at September 23, 2005 03:39 PM



an example of a "departure from specified functions" - a military application of Traderbots (sponsored by the U.S. Army Research Laboratory, under contract Robotics Collaborative Technology Alliance)


Complex Task Allocation For Multiple Robots
Robert Zlot and Anthony Stentz

By generalizing the definition of a task and developing appropriate mechanisms to handle these new task descriptions, we create a marketplace capable of distributing complex tasks among a robot team in an efficient manner.

A. Market-based Task Allocation for Multiple Robots
Market-based approaches to multirobot coordination treat a team of robots as participants in a virtual economy. Essentially, robots are contracted to complete required tasks in exchange for payment. Each robot has well-defined cost and revenue functions that can be used to compute the expected gains and losses for performing tasks. Since the participants are self-interested, they work to maximize their individual profits. Costs and revenues are designed so that individual optimizations lead to globally efficient solutions.

Our approach can be considered an extension of TraderBots. In TraderBots, agents called traders, one running on each robot (RoboTraders), and possibly others representing human operators (OpTraders), computers, sensors, or other resources – participate in a market, trading tasks via auctions. When an auction is announced, participants compute bids based on their expected profit for the tasks on offer, and the robots that can perform the tasks for the best price are awarded the resulting contracts.

Since only profitable trades occur, each auction acts to improve the global solution. Each RoboTrader maintains a schedule of tasks to which it has committed, and can evaluate new tasks by computing the marginal costs of adding them to its schedule. Traders can take on the roles of auctioneer and bidder dynamically, thus facilitating peer-to-peer trades amongst the team. This implies that tasks can be reallocated, allowing for solution improvements over initial assignments and for adapting the task assignments as new information is ascertained. Having no single auctioneer also avoids the presence of a central agent becoming a critical point of failure for the system.

C. Complex Task Markets
In order to effectively incorporate complex tasks, multirobot task markets can be extended to include task tree auctions. Instead of trading contracts for simple tasks, trees of tasks are offered in auctions. Participants can bid on any combination of nodes in the tree, and the auctioneer can choose to award several nodes from the same tree to multiple winning bidders. The winners of the auction are responsible to the seller and must ensure that the tasks are completed before receiving payment (either by executing the task themselves, or by subcontracting parts of the task to other teammates in future negotiations). Because bids can be on tasks at multiple levels of abstraction, task tree markets have the flexibility to allocate tasks at whichever granularity of abstraction is most appropriate.
One benefit of a task tree market is that task tree structures allow robots to express their valuations for both tasks and plans. Since robots have different states, information, resources and capabilities, they may prefer different decompositions for the same task. The calculation of these preferences is a recursive process.
The bidding language is a specification of which types of bids participants are permitted to submit to an auctioneer. There is inherently a tradeoff between the expressiveness and the simplicity of the bidding language. In the context of task trees, the bidding language can range from allowing the bidders to bid on only one node in the tree, to bidding on any arbitrary set of nodes in the tree. The one-node bidding language is very simple, in that it is easy to specify and the auction can be cleared optimally in polynomial time: the auctioneer does not have to consider which bid to award to each bidder – each either wins the one task bid on or not. However this language is not very expressive, as the bidders are filtering out most of their preference information. As a result, the possible solution space is limited and the resulting allocations are inefficient. The any-nodes bidding language is far more expressive, but it introduces a more complex auction clearing problem.

We have tested our approach on an area reconnaissance application. In this scenario, a team of robots is tasked with a reconnaissance mission which involves scouting a number of specified named areas of interest (NAI). To cover each NAI the robots select and navigate to a set of observation points (OP) and view the area with range-limited 360_ line of-sight sensors. Because the NAIs may contain enemies, the robots cannot enter them without incurring a large cost.
The mission is achieved when the robots visit a sufficient number of OPs to cover a predefined fraction (75%) of each area.

D. Simulation Experiments
A series of experiments was conducted to evaluate the effectiveness of task tree trading for the area reconnaissance scenario, using a multirobot simulator with a graphical display. In each test, a number of robots and NAIs are randomly placed within a 200x200-cell grid containing multiple obstacles. The terrain map was constructed from real-world measurements by an autonomous helicopter equipped with a downward looking scanning laser rangefinder4. The NAIs are non-overlapping, randomly sized rectangles with edge lengths drawn uniformly at random in the range of 15 to 30 grid cells.

Posted by: sd at September 23, 2005 04:39 PM




Posted by: northanger at September 24, 2005 12:00 AM



SWORDS - First Robots To Break Asimov's First Law Of Robotics (2005)
Next month, the US Army will be putting robot soldiers in the field in Iraq. The SWORDS (Special Weapons Observation Reconnaissance Detection Systems) robots are fully armed; when the human operator verifies that a suitable target is within sight, it fires. They are equipped with either the M249 (which fires 5.56-millimeter rounds at 750 rounds per minute) or the M240 (which fires 7.62-millimeter rounds at up to 1,000 per minute). These robots are poised be the first working robots that are actually designed to break Isaac Asimov's First Law of Robotics:

"A robot may not injure a human being, or through inaction, allow a human being to come to harm."

Autonomous mechanized combatants would revolutionize military strategy (2003)
Because machines are fearless, they bring a whole different level of capability to the battlefield, he adds. An enemy artillery barrage would have little impact on a group of TACs because the machines would not duck, run or hide but rather would continue advancing and attacking an enemy. “The first thing humans do as soon as artillery starts firing is they hunker down, wait for the artillery to pass and hope they survive. Machines don’t care about that. When humans are out there, if one of their comrades gets injured, of course the first thing they want to do is remove that human from danger. It takes perhaps four or five people out of combat because they want to save their comrade’s life. Machines don’t care when another machine is hit,” Johnson states.

Robotics to play major role in future warfighting (2003)
A study by USJFCOM's Project Alpha is currently testing the viability of what used to be a fanciful notion but is now bordering on reality -- robotic warfare. By as early as 2005, autonomous robots on the battlefield might not be just a notion, but a norm.

“We believe that other countries or groups will pursue robotics,” Richards said. “We can be at the vanguard, or we can lag behind and some day have to oppose a lethal robotic force. Better to be in the lead.”

Posted by: northanger at September 24, 2005 05:57 AM



from both last major comments its quite obvious (US) military robotics will be a huge driver of robust (initially) semi-autonomous practically embedded AI. Any guesses on when we can expect first Robotic Warfare divisions to appear? Hard to see what - other than budgetary constraints and PR issues - is stopping the Pentagon moving fast forward on this right now ...

Posted by: Nick at September 24, 2005 09:22 AM



Terminator terror inscribed deep into the human hard drive.

Posted by: sd at September 24, 2005 11:00 AM



responding to both your comments--
As one Marine fresh out of boot camp told Quinn upon seeing the robot: "This is my invisibility cloak." Quinn said it was a "bootstrap development process" to convert a Talon robot, which has been in military service since 2000, from its main mission -- defusing roadside bombs in Iraq_ into the gunslinging SWORDS. It was a joint development process between the Army and Foster-Miller, a robotics firm bought in November by QinetiQ Group PLC, which is a partnership between the British Ministry of Defence and the Washington holding company The Carlyle Group. Army officials and employees of the robotics firm heard from soldiers "who said 'My brothers are being killed out here. We love the EOD (explosive ordnance disposal), but let's put some weapons on it,"' said Quinn. Working with soldiers and engineers at Picatinny Arsenal in New Jersey, it took just six months and only about $2 million (euro1.5 million) in development money to outfit a Talon with weapons, according to Quinn and Anthony Sebasto, a technology manager at Picatinny.
It isn't that I have a Terminator inspired worries of a robot take over ... But with the "Gameboy" type controller and video-game like interface, this is one more small measure to eliminate contact with the enemy, bypassing what I (perhaps naively) believe to be a natural empathetic reaction to other people. I worry that SWORDS may not only reduce casualties, but also compromise human discernment.

i'm working a list of companies developing robotic weapons.

Posted by: northanger at September 24, 2005 12:25 PM



ps. which is totally OT for this Hi-Tech Finance post.

Posted by: northanger at September 24, 2005 12:26 PM



>>Hard to see what - other than budgetary constraints and PR issues - is stopping the Pentagon moving fast forward on this right now

doesn't look like there's any budgetary constraints. PR, what PR?

Foster-Miller TALON Robot Award Increased to $257 Million (9/19/2005)$257_million_for_talons.htm
Foster-Miller, Inc., announced today that it has been awarded a $133 million contract for TALON™ robot spare parts and service, bringing its total current award for TALON™ EOD (explosive ordnance disposal) robots to $257 million through 2012. This contract represents the third component of a three-part award that began in September 2002 with a $27 million contract for TALON™ EOD robots from the Man Transportable Robotic System (MTRS) program managed for all service branches by the Naval Surface Warfare Center (NAVSEA) in Indian Head, MD.

Two weeks ago, NAVSEA awarded Foster-Miller an additional $96 million to increase the number of TALON™ EOD robots to up to 1200 over the next seven years, bringing the company’s total award to $124 million, then the largest single order in its history.

Foster-Miller Awarded $96 Million Contract for Additional TALON Robots (9/9/2005)$96_million_for_more_talons.htm
Foster-Miller, Inc., announced today that it has received its largest US military order to date for ground robots. The multi-year contract increased from $27.5 million for 250 robots to $124 million for up to 1200 TALONTM Explosive Ordnance Disposal (EOD) robots. This order falls under the Man Transportable Robotic System program (MTRS) that is managed by the Naval Surface Warfare Center in Indian Head, MD. TALONTM robots will be delivered to Air Force, Army, Marine Corps and Navy EOD units around the world.

TALONTM robots are powerful, durable, lightweight tracked vehicles that are widely used for explosive ordnance disposal, reconnaissance, communications, sensing, security, defense and rescue. They are man-portable and have all-weather, day/night and amphibious capabilities and can navigate virtually any terrain.

Foster-Miller is the largest provider of robots for EOD work in Iraq and Afghanistan with more than 250 TALONTM robots deployed in theater. These TALONTMrobots have successfully completed more than 50,000 EOD missions, and have rendered safe thousands of Improvised Explosive Devices (IEDs), increasing security and safety in Iraq and Afghanistan.

Posted by: northanger at September 24, 2005 12:34 PM



northanger - that's chump change (under US$1 billion)
US$500 billion over a decade probably required for serious robot revolutionization of major military formations - my guess is it will just keep trickling in (like now) until WWIV gets cranked up enough to stimulate terminator-type escalation - if it becomes necessary for US to occupy for counter-insurgency purposes a demographically significant Islamic country (e.g. Pakistan, perhaps even Iran) the stress put on American infantry inserted into an ocean of jihadis should massively pump-up the roboticization of the war-machine - we're probably waiting for the first terrorist nuke to go off before the threshold is definitively crossed (which would also neatly eliminate the PR problem)

Posted by: Nick at September 24, 2005 01:15 PM



land, air & sea. ok you've got land, here's air.

X-45 Unmanned Combat Air Vehicle (UCAV)
The objective of the joint DARPA/Air Force Unmanned Combat Air Vehicle (UCAV) Advanced Technology Demonstration (ATD) program is to demonstrate the technical feasibility for a UCAV system to effectively and affordably prosecute 21st century lethal strike missions within the emerging global command and control architecture. The operational UCAV system is envisioned as a force enabler that will conduct Suppression of Enemy Air Defense (SEAD) and strike missions in support of post-2010 manned strike packages. This SEAD/Strike mission will be the first instantiation of an UCAV vision that will evolve into a broader range of combat missions as the concept and technologies mature, and the UCAV affordability potential is realized.

The Unmanned Combat Air Vehicle vision is an affordable weapon system that expands tactical mission options for revolutionary new air power as an integrated part of a system of systems solution. The UCAV weapon system will exploit the design and operational freedoms of relocating the pilot outside of the vehicle to enable a new paradigm in aircraft affordability while maintaining the rationale, judgment, and moral qualities of the human operator. In our vision, this weapon system will require minimal maintenance, can be stored for extended periods of time, and is capable of dynamic mission control while engaging multiple targets in a single mission under minimal human supervision. The UCAV will conduct missions from ordinary airfields as part of an integrated force package complementary to manned tactical and support assets. UCAV controllers will observe rules of engagement and make the critical decisions to use or refrain from using force.

Boeing X-45A Unmanned Combat Air Vehicle Begins Flight Testing
The Boeing X-45A Unmanned Combat Air Vehicle technology demonstrator aircraft, developed for the U.S. Defense Advanced Research Projects Agency and the U.S. Air Force, made its first flight May 22 at the Dryden Flight Research Center on Edwards Air Force Base in California.
Under a Defense Advanced Research Projects Agency contract, Boeing is currently building three larger X-45C aircraft capable of cruising at 0.85 Mach at 40,000 feet, with a 4,500-pound payload. The ‘C’ version will have an increased mission radius of 1,300 nautical miles. Future mission scenarios could evolve with unmanned aircraft complementing manned strike aircraft. Unmanned aircraft would be capable of eliminating preplanned targets as well as previously unknown threats, clearing the path for manned aircraft and thus saving lives and assuring mission success.

A unit of The Boeing Company, Boeing Integrated Defense Systems is one of the world’s largest space and defense businesses. Headquartered in St. Louis, Boeing Integrated Defense Systems is a $30.5 billion business. It provides network-centric system solutions to its global military, government, and commercial customers. It is a leading provider of intelligence, surveillance and reconnaissance systems; the world’s largest military aircraft manufacturer; the world’s largest satellite manufacturer and a leading provider of space-based communications; the primary systems integrator for U.S. missile defense; NASA’s largest contractor; and a global leader in sustainment solutions and launch services.

Joint Unmanned Combat Air Systems (J-UCAS)
The J-UCAS program is a joint Defense Advanced Research Projects Agency/U.S. Air Force/U.S. Navy effort to demonstrate the technical feasibility, military utility, and the operational value of a networked system of highperformance, weaponized, unmanned air vehicles to effectively and affordably execute combat missions. The JUCAS Common Operating System will allow unmanned aircraft systems to intra-operate with each other and with the Global Information Grid. The J-UCAS system-of-systems concept plans to demonstrate the military utility and the operational value of airpower in the 21st century combat environment. More information on the J-UCAS program can be found at .

The X-45A team has been making history in the desert for several years. Under the careful oversight of NASA’s Dryden Flight Research Center, Edwards AFB, Calif., and with the Air Force Flight Test Center’s continuing assistance, the J-UCAS team has risen to the challenge of achieving the Defense Advanced Research Project Agency’s very aggressive goals for these first unmanned combat air vehicle (UCAV) demonstrators.

“This demonstration concludes an exciting and precedent-setting time. The J-UCAS program has made significant strides in developing an unmanned air system that is changing the nature of air combat,” said Dr. Michael S. Francis, director of the J-UCAS program. “The pioneering efforts of the X-45A program have been critical in the quest to create a capability that
is effective even in the most dangerous, denied environments.”

Posted by: northanger at September 24, 2005 01:16 PM



think robot infantry the cutting-edge - look at the tactical context. US Air/sea superiority so overwhelming that benefits of roboticization relatively abstract, also major conflict in these dimensions less probable (unless US/PRC really screw up in Taiwan Strait, but even if they do combat will be so short and sharp that no protracted learning/development process will take place). UAVs (like Predator drones) undoubtedly important, but their key tactical usage is infantry support. Most significant domain of robot warfare experimentation will be counter-insurgency operations, including ambush, armed reconaissance, assassination, covert intelligence gathering, bomb disposal and patrolling.
Hi-tech insurgency suppression should fulfill everyone's fantasies, for the jihadis it's an ability to test Allah's favour in a fight against mechanical djinns from Satan, for the Western left it's a media spectacle of technocapitalist suppression of the victims of neoliberal hegemonism (lots of fat Lockheed contracts), for the right it's an opportunity to pit the science and technology of free and prosperous societies against the totalitarian ardour of enraged romantic barbarians ...

Posted by: Nick at September 24, 2005 01:41 PM



the Foster-Miller site is spooky...

they don't seem to have any vacancies for philosophers or linguists though. you're in with a shot if you're a chemist or metallurgist...

Posted by: sd at September 24, 2005 01:42 PM



"Hi-tech insurgency suppression should fulfill everyone's fantasies..."

Hollywood has no future in the future.

Posted by: sd at September 24, 2005 01:47 PM



Perhaps not surprisingly, Foster-Miller is somewhat circumspect when it comes to details about its biomedical program:

a bit more info about the TALONS though:

Posted by: sd at September 24, 2005 01:57 PM



These people - - design software to navigate terrain

"During Phase I of the program, our work focused on conducting sensor trade-off studies and developing individual algorithms for the selected sensors that enabled one or more off-road navigation tasks – functions like obstacle detection, foliage penetrability assessment, and terrain and/or vegetation evaluation.

During Phase II, we extended these individual algorithms, as well as developed a framework, called the Density Map, to merge all sensor information into a unified representation of the world."

They have also designed a tractor guidance system which automatically detects crop lines, which is nice.

Posted by: sd at September 24, 2005 02:19 PM



don't think you need a terrorist nuke to go off. maybe just a few perfect storms. remember, about $400M invested in robotics during the 1980s, even though that went bust (Lawrence Aragon called this a "deadly pursuit" due to low margins and low barriers to entry). however, that doesn't mean a robotics knowledge base wasn't created — just couldn't make money. imho, hi-tech finance isn't about spending a lot of money, but making a lot of money in the right market, at the right time with the right resources. factor in dirt-cheap components + defense spending + corporate funding + regional interest, etc, & you have a recipe for a rapid jump-start. remember, the US has a guns & butter economy (ie, no-bid contracts). in a nutshell, greenspan's hike & bush's cash overflow make splendid sense.
"Defense robotics will be measured in billions"
Technology that lets robots perceive and overcome obstacles has made unparalleled bounds largely because the cost of charge-coupled devices (the core of every camera), microprocessors and varied sensors has fallen away as rapidly as computing power and memory have expanded.
“In this post-9/11 climate, the defense business has been a very attractive business ... from a regional perspective ... how can we amass the resources necessary to invest in a future role in a business like that?”

“Robotics had just kind of been bumping along here,” acknowledges Dennis Thompson, Senior Vice President of Catalyst Connection. “Then comes along Sept. 11 and now a demand for new fighting systems and a bigger defense budget, which has created opportunities. We don’t want to put our soldiers in harm’s way, so robots become a solution for that.”

Posted by: northanger at September 24, 2005 02:29 PM



Foster-Miller: "We engineer ideas into reality through innovation and technology" - hyperstition for the military-industrial complex

Posted by: Nick at September 24, 2005 02:55 PM



>>hyperstition for the military-industrial complex

Headquartered in Norfolk, Va., U.S. Joint Forces Command is one of nine unified commands in the Department of Defense.

Among his duties, the commander of USJFCOM oversees the command's roles in transformation, experimentation, joint training, interoperability and force provision as outlined in the Department of Defense's Unified Command Plan.

The Department of Defense appointed U.S. Joint Forces Command as the "transformation laboratory" of the United States military that serves to enhance the unified commanders' capabilities to implement that strategy. We develop concepts, test these concepts through rigorous experimentation, educate joint leaders, train joint forces, and make recommendations on how the Army, Navy, Air Force and Marines can better integrate their warfighting capabilities.

USJFCOM develops future concepts for joint warfighting. Such work must include and strengthen service efforts, draw on the best of industry, and follow the will of the citizens as expressed through Congress.

Posted by: northanger at September 24, 2005 03:03 PM

Transformation is the process of changing form, nature or function. Within the United States military, transformation requires changing the form, or structure of our military forces; the nature of our military culture and doctrine supporting those forces; and streamlining our warfighting functions to more effectively meet the complexities of the new threats challenging our nation in the new millennium.

Preparing for this new future will require the U.S. military to think differently and develop the kinds of forces and capabilities that can adapt quickly to new challenges and unexpected circumstances. The U.S. military has a long tradition of experimentation - from the fleet problems of the U.S. Navy in the 1930s that gave birth to the concept of using aircraft carriers to the Army's famous Louisiana Maneuvers of 1941 that developed the doctrine for combined arms air/ground operations.

Posted by: northanger at September 24, 2005 03:05 PM



btw, thought you'd appreciate this nick. a "Mr. Katarincic" is quoted here:


Posted by: northanger at September 24, 2005 03:11 PM



Movement Control Algorithms for Realization of Fault-Tolerant Ad Hoc Robot Networks

Autonomous and semi-autonomous mobile multirobot systems require a wireless communication network in order to communicate with each other and collaboratively accomplish a given task. A multihop communications network that is self-forming, self-healing, and self-organizing is ideally suited for such mobile robot systems that exist in unpredictable and constantly changing environments. However, since every node in a multihop (or ad hoc) network is responsible for forwarding packets to other nodes, the failure of a critical node can result in a network partition. Hence, it is ideal to have an ad hoc network configuration that can tolerate temporary failures while allowing recovery. Since movement of the robot nodes is controllable, it is possible to achieve such fault-tolerant configurations by moving a subset of robots to new locations. In this article we propose a few simple algorithms for achieving the baseline graph theoretic metric of tolerance to node failures, namely, biconnectivity. We formulate an optimization problem for the creation of a movement plan while minimizing the total distance moved by the robots. For one-dimensional networks, we show that the problem of achieving a biconnected network topology can be formulated as a linear program; the latter lends itself to an optimal polynomial time solution. For two-dimensional networks the problem is much harder, and we propose efficient heuristic approaches for achieving biconnectivity. We compare the performance of the proposed algorithms with each other with respect to the total distance moved metric using simulations.

Posted by: northanger at September 24, 2005 03:14 PM



what's WHITE WAR?

Posted by: Nick at September 24, 2005 03:20 PM



reza's stuff.

Posted by: northanger at September 24, 2005 03:22 PM



Jay Katarincic, so Reza's got a hook too ...

Posted by: Nick at September 24, 2005 03:24 PM



Oh, got to check (The Thingy?)

Posted by: Nick at September 24, 2005 03:25 PM



yes, that thingy. he unposted it here & reposted it at his place.

Posted by: northanger at September 24, 2005 03:29 PM



>>Jay Katarincic
Allah works in mysterious ways.

Posted by: northanger at September 24, 2005 03:31 PM



Yes, Allah's boys are going to be chewing on some very pissed silicon

Posted by: Nick at September 24, 2005 03:40 PM



Future tense demonstrably misleading

Posted by: Nick at September 25, 2005 01:00 AM



we know.


Posted by: northanger at September 25, 2005 02:26 AM



oh. just read that PARRY stuff. the real trick involves questioning a human-robot hybrid. and both are paranoid.

Posted by: northanger at September 25, 2005 05:34 AM



... so what I've learnt from this thread so far: AI is already indistinguishable from paranoid schizophrenia and rigged-up with machine-guns ... seems like Singularity is going to be quite a party

Posted by: Nick at September 25, 2005 10:58 AM



you're confusing the GAME of singularity with THE SINGULARITY.

Posted by: northanger at September 25, 2005 11:49 AM



Oh. Now feel greatly comforted.

Posted by: Nick at September 25, 2005 12:14 PM



you should. besides, the game's just vaporware.

Posted by: northanger at September 25, 2005 01:14 PM



Assuiming sd must be pretty thrilled about this:
(sorry for OT - but we've reached that stage of the thread - any ideas for next topic?)

Posted by: nick at September 26, 2005 09:54 AM



not very thrilled - Law and Justice are overtly Christian, traditional conservatives: pro tax breaks and state aid for the poor, suspicious of free markets. [The leader, Kaczynski, faces prosecution in Strasbourg for banning a gay pride march in Warsaw - he organized a "March for Normality" afterwards.] Law and Order could remove Balcerowicz's currency controls - potentially disastrous.

The flat tax party is the Civic Platform, the liberal right, with whom Law and Order will have to form a coalition. It'll be unworkable.

Law and Justice ganged up with the two extremist parties, League of Polish Families (ultra-cathiolic fascists)and Self-Defence (barking mad farmers) to pummel the Civic Platform over their economic policies. The Law and Order Party Political Broadcast that seems to have clinched the election featured the threat of less food on the table and less money for home decoration, if the Civic Platform got in.

Although it's great to see the back of the commies, it's deeply annoying that the flat tax got rejected in such a stupid manner, and the Christian right are truly nauseating.

Posted by: sd at September 26, 2005 10:17 AM



"any ideas for next topic?"

I'm reading Pinker, Chomsky and Turing, with a view to some kind of linguistics/symbolism thread. I'm not quite ready yet though...

I'd love to get into some more stuff on psychology & self-fulfilling prophecy, especially with regard to The Singularity. Kurzweil has hyped his book brilliantly (a true intellectual entrepeneur) - how big a role will hype play in the arrival of The Singularity? How essential is Kurzweil's book? (got to wait a few weeks for my copy to arrive!)

Posted by: sd at September 26, 2005 10:27 AM



sd - apologies for ignorance on the Polish issue - haven't properly digested the news (and brain fried from absurdly vast extra editing task I ridiculously took on)

linguistics topic would be interesting, but worried my lack of access to The Language Instinct would prove an intolerable humiliation ;)

On Kurweil and self-fulfulling prophecy - yes, really clever point, thinking ...

Posted by: Nick at September 26, 2005 10:41 AM



The article is available for FREE! Just google "robo-traders"! The article has a really cool illustration.

Posted by: Karl at September 29, 2005 03:46 AM



Post a comment:

Remember personal info?