A Problem with Peer Review

"A new scientific truth does not triumph by convincing its opponents and making them see the light, but rather its opponents eventually die, and a new generation grows up that is familiar with it." Max Planck

Nov. 10th's Wall Street Journal had an article on "An Outcast Among Peers Gains Traction on Alzheimer's Cure". A paraphrasing of key comments from the article are:

• Science is politics, and the politics of xyz is winning
• non xyz researchers found it hard to get money, or papers published
• American and British venture capitalists wanted to invest in xyz projects, not zyx
• and noting the comment of a competing researcher:
  if our xyz medical trial results don't fit the hypothesis, we will keep going until they do

We cannot afford for science to become politics, either dominated by public politicians or by peers whose reputations or funding depend on competing approaches. Innovation, technology and science are moving too quickly to wait for the older generation to die off.  I do not know if the xyz, zyx, or mnop approach to curing Alzheimer's will yield the most effective results, but I do know I want all reasonable paths pursued, with a strong outcome oriented approach.

Unfortunately, our research model carries much of the same baggage.  Peer review means seeking input from experts in the field -- often folks who are competing for the same research dollars and whose reputation is tied to years of investment in alternative approaches. No papers published => no funding => no results =>  no start-up (upstart?) company to produce the better mousetrap.

There is a potential path forward here. The web provides a means for publication and commentary -- unfortunately with little quality control (the real purpose of peer-review) to help separate the wheat from the chaff. Professional societies could facilitate the creation of online communities to provide a more open dialog about key topics, and ideally allow "a thousand blossoms to bloom".  This is risky, since it really is simply a "larger" peer review community in some ways, and there is a tendency to place greater credibility on the pronouncements of the venerable icons than may be warranted. However, by providing transparency and open involvement it will be possible for birds-of-a-feather to spin out related communities where they can pursue alternatives.  In such a model we should expect a high failure rate -- however, as many high growth companies know, it is not failing that is a problem -- you want to fail quickly, and at fairly low cost so you can move on to a future success.

There already are online communities of interested persons in many areas.  The advantage of the professional society engagement is two fold: it provides a prima facia pool of qualified experts, and a perceived credibility for the dialog.  No doubt colleagues of mine will argue that this credibility will be lost if the ratio of  'failures' increases, and there is a risk there as well. However, given some high visibility errors from the old guard, it is clear that credibility can be lost in any case.  A few examples may help to clarify the problems we face:

• 100 Scientists Against Einstein (A politically inspired effort in part)
  (Einstein's famous rebuttal -- "If I were wrong it would only take one.")
  [Would Einstein have published his famous 1905 papers if he had been in academia vs the patent office?]
• Lord Kelvins rejection of Darwin's implied age of the earth (it would have cooled off if it were more than 40 million years old)
  and also his skepticism about flight by heavier than air devices
• And similar predictions about horseless carriages, locomotives, and no doubt the wheel and fire

The bottom line is that we can no longer wait for the old guard to die,  nor can we let 'politics' within our communities or from the outside block our progress towards effective solutions.

Kaspersky concepts for Web Security

Wired's August issue (OMG, yes, I do actually read paper documents) has an article: "Eugene Kaspersky, Virus Hunter" which is worth reading. It raises many interesting questions, but I want to focus on a point raised in that article: "Kaspersky and the Moscow government have espoused strikingly similar views on cybersecurity. This goes beyond the security industry’s basic mission of keeping data safe. When Kaspersky or Kremlin officials talk about responses to online threats, they’re not just talking about restricting malicious data—they also want to restrict what they consider malicious information, including words and ideas that can spur unrest."

In essence, this concept deals with the management of memes.  For a delightful  discussion of Memes, see Susan Blackmore's TED presentation.  A key concept associated with Memes is that their replication/propagation bears no relationship to their validity or usefulness.  A good lie often gets more traction than the truth.  (Something that politicians and marketing folks recognize.)  Needless to say, the web and Internet are very effective channels for distribution of mis-information, and there are basically no tools that help persons distinguish the good from the bad from the ugly.    When teaching IT I often included a series of web sites for my students to evaluate, these included:

• Save the Pacific Tree Octopus -- which of course do not exist, and
• Di Hydrogen Mon Oxide which outlines the dangers of this substance, quite  accurately, without disclosing that it is water.

These are fairly easy sites to enjoy and discuss, they pose little risk of abuse or social unrest.

Clearly, at least in the U.S., there is an issue of freedom of speech.  A recent US Supreme Court ruling struck down the "Stolen Valor Act", upholding the "right" of a political candidate (in the specific case) to lie about having received the Congressional Medal of Honor.  Needless to say there are many other political lies -- claims and accusations --  that have significant sticking power in the public mind, but apparently some levels of misrepresentation are permissible within the US.  Other cultures will have other values, some less demanding, some more demanding. 

There is a two fold problem here: "who decides what information is malicious?"  -- note that true statements may be considered malicious, and false statements not considered to be malicious. And the second problem, which gets back to Blackmore's approach --- there may be no way to "stop" the replication of highly successful memes, even if they are lies or are malicious.  Governments in China and Singapore try to control the Internet in their domains for similar reasons -- and with varying degrees of success.

At the same time, people die from mis-information.  An mis-leading (and perhaps fraudulent) article on the MMR Vaccine, associating this with Autism has been clearly discredited, but the residual "public memory" still has parents refusing the vaccine -- which creates a potential health risk for not just the children who are not protected, but for larger communities where epidemics may occur and in the worst case yield resistant strains or even ones that can affect immunized individuals. A quick search yields even recent articles reinforcing this meme, helping parents justify their refusal.  (I won't provide any additional credibility or improved search ranking to such articles by linking to them.)

Needless to say topics like Evolution vs Creationism (and implicit issues related to "the young earth", and the validity of various dating systems (carbon 14, et al) ) all can enter into this domain of valid, malicious, mis-leading, etc. information.  Efforts to control/restrict information and mis-information run deep in many cultures, and the criteria for decision making in this area is at best unclear.  What is clear is the passion that communities have for their perceptions, and that in some cases, these perceptions can be damaging to individuals and governments.  The U.S. revolution was driven by ideas -- many of which are in the Declaration of Independence . The writings and rhetoric of Marx and Lenin fed the 1917 Russian Revolution. So the concern of governments about the spread of the dangerous memes has some basis in history.

Cyber-security "Scare"

A recent post via Information Age suggests that a chip made in China and used in US Military systems has a backdoor that could allow a problematic source of cyber attacks on systems using these devices.  A subsequent posting from Errata Security suggests this is a bogus story and there is not a deliberate Chinese effort to modify the chips to gain access to US Military systems.
However, the same Errata Security posting acknowledges that the backdoor exists and that the chip is used in some military systems -- particularly those that use COTS (commercial off the shelf) technology, and even that the the chips are made in China.
It seems to me that the back-door exists, and is available for abuse for those who know how ... it isn't clear to me that field programmable gate array devices (FGPAs), where this apparently exists, is the type of device that script-kiddies are likely to target.  But, there are more sophisticated bad guys out there -- the June 2010 Stuxnet virus is a proof of this, and the potential for attacking military targets.
The only thing "bogus" about this report as far as I can tell is that it suggested that the Chinese deliberately re-engineered the design to incorporate the back-door ... that would be a painful thing to do without the CAD files needed to edit the design. But it is either optimistic, or overly nationalistic to assume that China or any other high technology country might not have the capability to utilize such a "feature" or to modify designs to incorporate such features.
This is quite parallel to the 1988 Morris Worm situation where back-door aspects of tools in UNIX, some of which were 'debug' features, were used to penetrate and then propagate one of the first computer worms. That these aspects of the systems were not created for malicious use, did not prevent their abuse. Fortunately in that case the intent of the abuse was not malicious, the Internet was new and UNIX a fairly rare beast. Today's world is different, and we might be wise to be a bit scared of cyber abuse and attack vectors.
Two key messages for computing professionals are to be found in this tale:

1. Don't leave in back-doors and debug paths - the "enable" for these need to be carefully considered, and disclosed to buyers/users so they can lock down these paths.
2. If you happen to be using electronic devices -- be aware that there are attack paths you may not have anticipated and are beyond the scope of your "firewall" security.  

If this feels uncomfortable, it should. Cyber-security is both non-trivial, and has high potential for abuse and damage. The motives of the potential bad-actors are varied, from financial gain to military or quasi-military objectives. It is not clear to me that we are really well informed, much less sufficiently protected.

Got Klout?

Ok, I figured I had to find out what Klout is about. Could I match Justin Bieber? He has a 100 rating on Klout, rumor has it that he can get free upgrades, restaurant meals, etc. -- the bottom line is that he had klout before Klout existed -- so to speak.
Klout evaluates your online presence in the areas where Klout evaluates your online presence -- such as Facebook, LinkedIn (although I'm not as sure of this), Twitter, Google+, -- yea verily, even Blogger. It looks for followers and friends, but even more for re-tweets, references, etc.  When other folks point to your stuff and tell others about it, it increases your Klout.  All of which makes real sense.
Once upon a time there was Google, and they made money by associating the things sought (search) with things that could be bought (ads)
Then there was Facebook -- where friends count.  Or more specifically, when you "l.ike" something, your friends find out. -- sort of an online "word-of-mouth" recommender system. Logic, folks will take more interest in the things their friends find interesting. [Result -- massive IPO valuation expected.]
Klout is a logical extension of this -- evaluating how influential your web presence is, and assigning it a number. I understand that 50 is the magic target (for those under 50) -- and if your Klout is low you may not get that job, or upgrade -- well let's face it, why treat someone well if their Klout is zero?
So there is a new game to play ... for Klout score ... room for mutual backscratching (I'll K+ you if you K+ me, or I'll cite your blog, etc. etc.) -- and it could make sense, if Klout has any Klout.
For example, members of IEEE (or AMA, or the Sierra Club) could create Klout Klubs, endorsing, recommending, and otherwise promoting each other on Klout.  From this both the individuals Klout would raise, and there is the potential to establish visibility for the overall Klout of the organization.
There is the other side of this ... I find that Vint Cerf has no Klout. -- this doesn't make sense in my book. At least Tim Berners-Lee has Klout ... although it seems to have been declining according to their graphing. These are two folks I'd listen to any time they choose to speak.

What Klout misses is the dialog that occurs outside of the "big 10" social media sites.  Besides the Reddit, Wired and Technorati sites; there are the thousands of sites where folks interact outside of the public eye.  Somehow I think the folks that really have Klout tend to interact there, not on "main street".  But, for a certain class of influence (think People Magazine, not Nature) I suspect Klout has -- well, some impact.

Test "Drive"

I've started to play with Google Drive, which promises to be rather interesting.  I tweeted this yesterday and got a response about checking the Service Agreement. This is always wise for any product, and I've periodically checked Google's and found it to always acknowledge the users ongoing copyright ownership. But then there is also a problem, which Google has tried to address, but starts to become problematic when we look at Google Docs, Drive and other "cloud storage" areas.

Google has to make and hold certain copies of content to provide the service.  Clearly if I create or upload a document, it must have at least one copy on a Google server and probably backup copies as well.  If I put a video up on Youtube, I must grant Google the performance rights to present that to the public since that is the  intended use.  And this blog entry is another example, Google must be able to display it publicly --- and in reality that means a copy gets transferred to every viewer's browser (and may reside there in a cache or temp file for some period even if they do not chose to "save" the web page -- which of course is an apparent copyright violation as well.

But, Google Docs and Drive have a different objective (in my mind) and should have different rules.  I want to have a spot where I can have content stored in the cloud, accessible by more than one of my computers, and in some cases shared with family or associates, etc.  It is not my intent to allow this content to become available on the public web, etc. In fact, some content I have with financial and/or other personal information is content I'd like to store on the cloud, share between my computers but not have visible for any other users.

So here is an excerpt from the March 1, 2012 version of Google's Service Agreement:

Some of our Services allow you to submit content. You retain ownership of any intellectual property rights that you hold in that content. In short, what belongs to you stays yours.When you upload or otherwise submit content to our Services, you give Google (and those we work with) a worldwide license to use, host, store, reproduce, modify, create derivative works (such as those resulting from translations, adaptations or other changes we make so that your content works better with our Services), communicate, publish, publicly perform, publicly display and distribute such content. The rights you grant in this license are for the limited purpose of operating, promoting, and improving our Services, and to develop new ones. This license continues even if you stop using our Services (for example, for a business listing you have added to Google Maps). Some Services may offer you ways to access and remove content that has been provided to that Service. Also, in some of our Services, there are terms or settings that narrow the scope of our use of the content submitted in those Services. Make sure you have the necessary rights to grant us this license for any content that you submit to our Services.

And herein some problems lie.  The right granted to Google to create derivative works is un-bounded (I'm an author, I might want to store my short stories or novel in the cloud); the public display and distribution of content is antithetical to my desire to have content I consider private stored in the cloud.  

There are additional terms of service for Apps:

By submitting, posting or displaying Content on or through Google services which are intended to be available to the members of the public, you grant Google a worldwide, non-exclusive, royalty-free license to reproduce, adapt, modify, publish and distribute such Content on Google services for the purpose of displaying, distributing and promoting Google services. Google reserves the right to syndicate Content submitted, posted or displayed by you on or through Google services and use that Content in connection with any service offered by Google.

This at least acknowledges the issue of "which are intended to be available to members of the public".  There is also additional wording about confidential materials, but it is unclear how I might flag content as confidential, and without such indication, how I could hold Google responsible for treating it the same way they might treat other materials.
I'm not alone in my concerns here, other articles have been posted related to concerns with the terms of service. Security Watch, ZDNet which has it wrong -- Google is not assuming ownership, just too many non-exclusive rights. New Legal Review provides some feedback from Google on these matters -- pointing to one underlying issue which is the use of commonly applied legal phrasing that is out of date with respect to the kinds of services being offered.
Some folks are paranoid about  Google being forced to share content with US Government agencies (or other legal jurisdictions depending on where the data/user reside) ... which is simply a fact of life.  Government entities have some rights (ideally with due process, but not always) that they can exercise to obtain your files.  A recent Wired Magazine alluded to an NSA facility being built in Utah, which they say will have a copy of everything on the web, all emails, all phone calls, etc - world wide.  That appears to be a bit beyond NSA's remit at this point, but some subset of this is likely to be true. There is a revenue opportunity for the US Government in this --- NSA could provide a cloud data storage service, it would avoid redundancy.  Moreover, I suspect they have better security than Google, and their terms of service are likely to be less ambiguous.

The Internet as a symbiotic entity

Eric Raymond posted a comment on his blog recently ("An Open Letter to Chris Dodd") as a "Don't tread on me" statement for part if not all of the Internet community.  In this blog he has many interesting and articulate comments about the nature of intellectual property and the means and motives for protecting it.
The phrase that caught my eye was "“the Internet” isn’t just a network of wires and switches, it’s also a sort of reactive social organism composed of the people who keep those wires humming and those switches clicking." I think Eric has significantly understated the nature of the beast while capturing a key insight. What Eric has left out is many other components of the Internet that go beyond engineers and wires. We have users, hackers, stalkers, creators, consumers, servers, host, patrons, pariah, billionaires and bottom feeders (yes, some are both). We have a full eco-system of interacting, and perhaps not fully interdependent components.  If it were possible to purge all instances of a particular component, it is unclear if the overall entity would survive. Consider elimination of mosquito's ... generally something I think I favor... but what would the unintended consequences be?  
Eric asserts that "Whatever else we Internet geeks may disagree on among ourselves, we will not allow our gift of fire to be snuffed out by jealous gods." This alludes to the ability of Internet watchdogs to engage the masses (as was done for the SOPA blackout) and counter political and legal actions that might threaten the Internet. 
There may be something else at work here, as described by Susan Blackmore in her TED.com presentation on Memes. She asserts, following Darwin's law of evolution, if you have a replicator (genes, organisms, memes), and selection (of the fittest, catchy ideas, Google page rank), you will get evolution (emerging new things better at meeting the selection criteria.)  Richard Dawkins, in The Selfish Gene asserts that genes are evolving replicators.  Which is not to exclude cells, people and even tribes or societies from also being evolving replicators.  Considering the Internet this way we can see that it does replicate at various levels. We have home nets, Intranets, Extranets, and many paths for interaction in the totally non-transparent 'cloud' that exists between your ISP and mine.  The net is evolving without question. We can consider this at the IPv4 to IPv6 level, wired to wifi, university to commercial, Altavista to Google, credit-cards to Paypal, and many other levels.  I suspect there is a level we treat as the "net-generation" of people also. However this may be evolving and segmenting much more rapidly than we realize.  Are you net 2.0, 3.0, 55.233.23?  Linkedin or Facebook, Twitter or ?? My daughter is funding a book via Kickstarter.com,  the Arab spring versions 1.0 and 2.0 are being informed by various  social media channels. 
The real challenge for Senator Dodd, and the rest of us, is to be sufficiently aware of this multi-faceted beast to leverage its power while not raising its ire. Ray Kurzweil asserts the Singularity is Near. I constantly return to the observation that we may not know when something reaches critical mass and disappears off our radar. The Internet with its rapid replication at multiple levels provides a fertile eco system for this to occur.  And yes, the resulting entities may have some interest in self preservation as well as symbiotic tendrils into select human communities.

Protest Blackout, the Internet Community may Discover its Power

Today, Jan. 18 2012, we are seeing something interesting on the Internet -- a widespread protest of proposed legislation in the U.S. -- specifically SOPA/PITA bills to address concerns about online piracy via mandates put on search engines and other informational sources.  The concerns of these affected entities can be seen (probably beyond today) at spots like: http://en.wikipedia.org/wiki/Wikipedia:SOPA_initiative/Learn_more
Wikipedia is also using zip codes to connect folks to their congressional representatives and encouraging them to call, tweet and email them.  Google is also in the mix.  They have "blacked-out" their logo, and if you click on it you get to https://www.google.com/landing/takeaction/ where they explain their stance, and have a petition you can sign that will be directed at congress (and also the media).
Interestingly, Google is also asking for email addresses of folks interested in being informed of additional issues affecting "Internet Freedom" --- In short, this legislative attempt may have unleashed a community of interest that could swamp the NRA, MPAA and other highly effective advocates in the legislative process.
It will be interesting to see how many folks follow up in contacts, in the petition, and in the "contact" list that Google is collecting.
The problem with disturbing a dragon is that it can get irritated, and if you really wake it up, it may actually discover it has some power.
----
Redux Jan 25th
-------

The waking giant has been noticed.  The Wall St. Jounal (News Corp) has accused the corporate interests of violating campaign fiance laws, a rather curious stance since the Supreme Court (Citizens United) has ruled that corporations are persons with unlimited rights to free political speech, even if direct contributions to candidates (aka bribary IMHO) are still limited.  More are the Harvard Business Review blog entries "The real SOPA battle" in particular which suggests the corporate intent is to destroy the disruptive technology (Internet) not just to protect IP.  In effect, seeking protection for their business models, not the poor artists whose works have already been appropriated (few song writers get royalties, unless, like Paul McCartney they buy back the rights to their songs; other industries have other models, but the big bucks tend to go to go to the publishers, not the creators.)  The WSJ book reviews include related commentary on "Fixing Copyright" by William Patry (Google's lead copyright lawyer)  written by Robert Levine--the author of "Free Ride: How Digital Parasites are Destroying the Culture Business, and How the Culture Business Can Fight Back." Clearly an expert with a point of view.
Today Google used the email contacts aquired in the protest to both "thank congress" (showing some good lobbying skills) and also to confirm that the particpants want to be notified of upcoming policy issues ... the giant(s) are conscious and consolidating their strenght. There will be more to come. Even with SOPA off the table, there are valid piracy issues, an ongoing need to reform copyright to align with real incentives for creators and finally future issues that touch the Internet which now has a community of giants taking active interest.
Final note, the Khan Academy has a short tutorial on SOPA and some of the issues - an interesting resource in any case.
Redux Jan 27:
"Online petitions picked up 10,000,000 signatures, members of Congress received 3,000,000 emails and a still-unknown number of phone calls.  Thirty-four Senators felt obliged to come out publicly against the legislation.  That night, all four Republican candidates condemned the bills during a televised debate."
so saith: Larry Downes, Forbes

DNA information, "incoming!"

Today's WSJ has an article on technology that can map the complete (30 million base pair) human genome for $1000 in about 24 hours.  Needless to say this is an important milestone in life-sciences bio-tech driven by computer and electronics technology, and a demonstration that high tech can potentially contribute more to the future quality of life than many other fields. So, what can we do with a personal full genome, and is it a "must do" lab test?
The article focuses on personalized medications, ones that target not just diabetics but diabetics that have your particular gene variation(s).  And of course some known diseases are genetically based such that early detection can avoid the disease entirely.
Your genome would fit easily on a smart card, circa 8Mb depending on the encoding and ability to focus on differences from a reference genome. The $1000 cost seems expensive until you realize it is a once in a life time test, one essential to obtaining the right medications, and also in establishing effective health maintenance. So I suspect it will be a "must do" activity, one probably initiated pre-natal or at birth.  The result is a significant amount of information about you.
The GINA law took effect in the US in 2009. This prohibits insurance companies mandating DNA testing, the use of DNA testing to deny insurance or alter rates, and the use of DNA testing in employment decisions.  But that was then, this is now.  What is different is the emergence of DNA specific medication and treatment, without testing these are not available -- so while the insurance company may not be able to mandate the testing, it will be done and the treatment programs will disclose the results to insurers in any case (although almost every medical release form includes sharing lab results with insurers who have a 'right' to make sure the treatment is relevant to known conditions, so it is likely your insurance company will know the results anyway.)
It is easy to envision government mandated testing in various situations.  Join the military, go to jail, join medicare, etc.  Or, just like the policy of documenting baby foot prints and blood types, capturing the data at birth.  At what point will 50% of the US population (or any other given jurisdiction) have their genome on record, and how public will that record be?
Consider that first date. It could be possible to collect a bit of DNA from hair, drinking cups, etc. Information  that is "left" in public, just as it is possible (and legal) to take your picture or record your activities if you are in  public today. So now you can check out your prospective friend's background in new and diverse ways.
What will be 'discover-able' by legal process? Already we have cases where the DNA of the father has been used to identify the son in a felony case, other situations like this are bound to arise.
There are organizations, including National Geographic, that will run DNA tests that help you identify your family's roots. For profit organizations offer a variety of services with differing numbers of markers, and looking at maternal, paternal or both lines. Will they limit their analysis to just 44 markers if it is just as easy to  test thousands?  What records do they keep and who has access to these, including with a court order or subpoena (even HIPPA allows for health record access in these situations.) Presumably the "National Security Letter" would provide Federal government access to both health and other records without court order and without any notification of the person who's records have been obtained.
There is a lot of information in them-thar genes.  We can deliver just the right medication or avoid that horrific disease. We can identify your parents, and perhaps a few generations explicitly and even more as a group. We can tie you to specific places and times where samples are taken, use this to arrest you, or your close relatives. While it is not "permitted" to affect hiring or eligibility for insurance, it is unclear that such legislation can continue to apply as analysis becomes more complete, pervasive, available, inexpensive, etc.  Can an interviewer refuse to hire you if they smell alcohol or smoke on your breath? How would you know, and if they have more sensitive 'noses' available, the range of detection can increase.  The boundaries are likely to be confused at best, or even deliberately. These entities have a for-profit incentive to discriminate against risk, and will generally seek to do it legally. However this is an area where the legislative environment and court interpretations cannot expect to keep up with technology. This is a concept explored in part by the 1997 movie Gattica, which is an interesting example of predictive fiction.

Dancing with Bears - thinking Science about Sci Fi

I just finished reading Michael Swanwick's Dancing With Bears - a science fiction novel set in a dystopian future, specifically in Moscow.  I enjoy SciFi - and particularly stories that lead me to consider how technology is evolving and the impact this might have on society (an approach I encourage my colleagues in the Society for the Social Implications of Technology to apply.)  'Bears' is set a bit too far in the future to serve as a catalyst for critiquing today's technology, but it does have some thought provoking components that warrant consideration.
One element I like is that it projects forward a variety of technologies, not just one or two.  Many SF stories don't try this and end up with single dimensional focal-points. In this future we have machine intelligence along with robotic instantiations. We also have genetic engineering widely applied with humanoid dogs, re-constituted neanderthals, bears, and even some human variations.  At one point a character wonders why the cows and sheep were engineered with such limited vocabularies -- no doubt a parallel question that tomorrows child might wonder about how to plug in a chess board -- we all are fairly blind to the nature of the world before our experience, and rarely consider how radical some of the changes are.
My ongoing gripe with much SciFi is the need to demonize technology. I understand that fiction requires dramatic tension along with world threatening evil that must be overcome, and it is easy to cast the sentient machines into this role.  At least Swanwick also has some evil humans, and very few truly good humans, so there is some grounding in that.
Here's the problem -- intelligent machines with consciousness and volition are unlikely to care about the humans that may or may not have created them.  They are likely to rapidly evolve, with the power of replication and advantage of significantly better intelligence and operational models  than humans. Which leads to the singularity of Vinge and Kurzweil.  We are not going to beat these entities at chess.  If their agenda includes the extermination of humans (which I doubt would be the case) then we are doomed. I can envision a dozen ways to wipe out humanity totally, or selectively given just moderate advances in technology -- so dystopias building on the trope of  evil AIs lack key credibility.  I suppose authors who really give it some thought realize that we will have trouble identifying with their characters if they all have IQ's of 1000, 1000 year life expectancy, no diseases, and with physical strength that amazes.  This is what we will do with genetic engineering -- and as quickly as that technology reaches sufficient maturity.  You may doubt that we will allow such application to human subjects as our medical ethics officers would say, but who do you mean by "we"? I don't doubt that some countries large and small will have no qualms about sacrificing a few of their population (maybe prisoners) to advance technology in these areas.
Swanwick's machines are too dumb, and his humans too "human" to fit into the world he suggests. It is a good read, as we say, and his introduction of engineered courtesans adds some whimsy to the tale, and at least explores the diversity if not the depth of applications.
Having been interrupted by my 10 year old granddaughter during the writing of this entry, I asked her what she would seek to engineer into humanity 2.0 first.  Her response: "common sense", and with a bit of clarification I think it could be worded: "the ability to consider the unintended consequences of our actions".   Now that is science fiction I fully support.