Every Thursday, a Bangalore-based community called ReRoll organises an evening of boardgames open to all at Lahe Lahe, a café in Indira Nagar. According to the organisers, about 40 people come each time and about 10-15% of them are first-timers – like I was today – and a similar fraction are regulars, who come every week – as I plan to hereon. It was a fun evening, I played three games: Tsuro, Avalon and Parade. I also met a bunch of cool people (including two I’ve ‘met’ only on Twitter) and who I certainly wouldn’t have met otherwise. So Tabletop Thursday, as the occasion is called, is also an excellent opportunity to make new friends. More details here.
Appeasement v. truth
Late last year, Facebook inducted The Weekly Standard (TWS), an American news outlet, as one of its only five fact-checkers and the sole conservative voice in the group. Earlier this week, TWS raised objections about an article published by Think Progress, a liberal news outlet, prompting the latter to claim it was being discriminated against for its ideological slant. Obviously, the event has raised a furore on Twitter, with many more liberal voices piping up in favour of Think Progress and denouncing Facebook’s attempt to appease everyone instead of aligning itself with the truth. There’s a lot going on here that I’d like to unpack.
The Think Progress headline reads: ‘Brett Kavanaugh said he would kill Roe v. Wade last week and almost no one noticed.’
First – As Think Progress has discussed, it doesn’t make sense that Facebook recruited TWS into its fact-checkers’ fold when it hadn’t been subjected to the same evaluation standards as the social platform’s four other fact-checkers (AP, Factcheck, PolitiFact and Snopes) – until you suspect Zuckerberg is indeed trying to appease both sides of the political spectrum. This doesn’t bode well for a platform that constantly aspires to be the place people get their news in nor does it bode well for publishers, who now have a disincentive to be liberal or more generally disagree with TWS.
Second – It’s possible that Think Progress has been penalised for its ideological slant here. Liberal v. conservative rivalries in the news space are increasingly becoming the norm, with some outlets pointedly aiming their guns at what their ‘opponent’ outlets are saying. That said, the headline for the original Think Progress article was, in fact, misleading and deserved to be cautioned about.
When you say a judge has “said he would kill” a law or previous judgment, you are immediately implying he intends to undo it. But this is not what Kavanaugh said. According to the Think Progress article itself, and explainers by Vox and The Conversation, Kavanaugh is opposed to the precedent set by Roe v. Wade. As an associate justice nominee to the SCOTUS, if Kavanaugh makes it to the bench, there will be a 5-4 majority to overturn Roe v. Wade. Does the Think Progress headline imply that he will vote that way? Yes. But has Kavanaugh said he will vote that way? No.
Third – Michael Stern argued on Slate that it was unfair for TWS to dock Think Progress for an opinion piece and label it as “false news” on Facebook. But opinion or not, Think Progress is guilty of misrepresenting reality in its headline. As TWS appears to have asserted, changing the headline will remove it from Facebook’s ‘beware’ bin. This is also where a point made by Brendan Nyhan, a public policy expert who contributes to Upshot NYT, is relevant: that on Facebook, most people see only the headline.
At The Wire, we have often observed that an article with a striking headline will receive a lot of engagement on Facebook for the headline alone, and with the numbers suggesting those sharing the link may not even have read the article. So the headline plays an outsized role on Facebook and Think Progress should have exercised restraint with that component of the standard article, and not have tried to hide behind nuanced analyses and hedging syntax following in the body.
Fourth – While Nyhan makes a useful point about how it is Think Progress that is really resorting to the ‘liberal v. conservative’ defence, and not TWS, I think he falters in believing that fact-checking wasn’t exercised as a form of censorship here as well as in not considering if TWS itself may have stepped over the line.
Both these concerns are tied to how Facebook handles content that makes questionable claims. For example, Think Progress has alleged that:
When an article is labeled false under Facebook’s third-party fact-checking system, groups that share that article on Facebook receives a notification informing them that the article received a “False Rating” and that “pages and websites” that share that piece “will see their overall distribution and their ability to monetize and advertise removed.” Facebook’s notification regarding our piece on Kavanaugh and Roe v. Wade effectively warned outlets not to share ThinkProgress content or risk censorship themselves.
(I’m not on Facebook and can’t check for myself.)
Separately, Nyhan shared a screenshot of a message he had received when he attempted to share the Think Progress article:
Zuckerberg has also said that content marked as “fake” or “false” will be demoted on Facebook and rendered invisible to up to 80% of all the people who could have seen it otherwise.
My takeaways from these messages are the following:
- Facebook doesn’t care how wrong an article is – it can be entirely wrong or it can have one wrong sentence –, it will still be graded on a single-point scale (falseness: 0/1) instead of providing users with a more meaningful assessment of the problem.
- If an article is marked as “fake” or “false”, it will lose 80% of its audience. If this isn’t censorship, what is?
- The fact-checker gets to plonk its own article in front of another on the same topic but which it has deemed unreliable. Thus, given Facebook’s apparent intent to appease all sides, including TWS among its fact-checkers runs the risk of magnifying conservative voices over liberal ones. Can we expect that Facebook will soon include a liberal voice, you know, just to please everyone?
Fifth – The essential problem with treating the political right and political left on equal footing, as much as treating the social conservative and the social liberal on equal footing, is that such an equation contains a Trojan horse that most people don’t account for: the right/conservative frequently make claims and argue from a position that is not rooted in rational beliefs and, in the US, yearn for originalist interpretations of constitutional values. As I wrote two days ago,
As such, [a realistic ‘idea of India’] is hard to come by in the media because, as Ram Guha explains, it is unglamorous and difficult to sell, even as the press is the institution responsible for the viewpoint’s day-to-day distribution and maintenance. As a result, you get partisanship more often than deliberation. … It might be useful to clarify here that such deliberation is not between leftwing and rightwing vantages but between reason and reason. Unreason has no place here, nor does its conflation with partisanship. And it is often the case that the right is aligned with pseudoscience and illogic that it confuses resistance against unreason with resistance against itself. It is frustrating not because the pedantic distinction is lost in the muddle – who cares – but because the result is often that the leftists co-opt your turf simply because you’re not right enough.
A similar endeavour to co-opt territories has happened here: because Facebook screwed up its fact-checkers’ ability to respond meaningfully to an article that was partly flawed, it gave Think Progress the ammunition it needed to claim that it had been wronged by a conservative competitor. So an effort to appease both sides – as undertaken by Facebook, for example – will remain an appeasement and will never qualify as any kind of attempt at fairness. This doesn’t negate TWS’s calling out Think Progress as much as assert that Facebook is chiefly responsible for this mess being what it is.
Footer image credit: TheDigitalArtist/pixabay.
US court settles bitter gene editing patent case
On September 10, a US court settled an increasingly churlish patent dispute between two research institutions in the country, the University of California (UC) in Berkeley and the Broad Institute, Massachusetts, with great consequence for the commercial use of a powerful gene-editing technology called CRISPR-Cas9.
The dispute centred on CRISPR’s usability in two different kinds of biological lifeforms: prokaryotes and eukaryotes. UC had devised a way to edit the genetic makeup of prokaryotes using CRISPR in 2014. Broad followed a year later with a method to use CRISPR in eukaryotes. UC had subsequently contested the patentability of Broad’s method saying that it was a derivative of UC’s method and couldn’t be patented separately.
The court as well as the US patent office have disagreed and upheld Broad’s patent.
The following FAQ breaks the case down to its nuclear components and assesses the verdict’s implications and future courses of action.
What are CRISPR and CRISPR-Cas9?
CRISPR is a natural defensive mechanism that prokaryotes use to protect themselves against viruses. Prokaryotes are smaller, less complex lifeforms and include bacteria and archaea: they are unicellular, the cells lack a membrane as well as membrane-bound organelles. The bigger lifeforms are classified as eukaryotes, which are multicellular, whose cells have a membrane and the cells also contain membrane-bound organelles. Because of these and other differences (see here, p. 8), it wasn’t clear if a CRISPR system built for use in prokaryotes could be adapted by a person of reasonable skill for use in eukaryotes with a reasonable chance of success. This distinction is at the heart of the patents dispute.
CRISPR-Cas9 is essentially a technology that combines CRISPR with a protein called Cas9 to form a molecular tool. This tool can swim to a eukaryote’s DNA, pick out a specific section of the genes and snip it at both ends, removing it from the DNA sequence. Once the section is out, the DNA strand repairs itself to restore the genes. If the original section of genes was faulty (e.g. containing an undesirable mutation), then CRISPR-Cas9 can be used to remove it from the DNA and force the DNA strand to repair itself to a healthier version.
Researchers have already reported that they are close to using this technology to treat a debilitating condition called Duchenne muscular dystrophy.
Of course, there are other gene-editing technologies, including zing-finger nucleases and transcription activator-like effector nucleases, but CRISPR-Cas9 has proved to be more efficient, effective and easier to use. At the same time, a few concerns are starting to emergeabout unintended side-effects.
What was the case timeline?
The UC team, led by Jennifer Doudna, published a paper in August 2012 describing how an RNA-based system called CRISPR-Cas9 could be used to edit DNA in prokaryotic cells. Editing DNA is a lucrative prospect – then and now – because it allows us in theory to modify the fundamental constitution of biological life, curing debilitating illnesses as much as modifying crops. And what the UC team had found, together with Emmanuelle Charpentier, then of the University of Umea in Sweden, was the first tool that could achieve this. After their paper was published, the UC team filed a patent with the US Patent and Trademarks Office (USPTO) for the use of CRISPR in prokaryotes.
While UC’s patent was pending, a team led by Feng Zhang from the Broad Institute, setup by the Massachusetts Institute of Technology and Harvard University, Boston, published a paper in 2013 and then built a CRISPR system in 2014 that could work in eukaryotes. Zhang and co. then filed for an expedited patent that was granted in 2017. At this point, UC complained to the USPTO that the Broad patent infringed on its own – that, effectively, Zhang et al’s work was not patentably distinct from Doudna et al’s work. UC’s own patent for CRISPR use in prokaryotes was granted in early 2018.
In late 2017, the Patent Trial and Appeal Board (PTAB) of the USPTO upheld the Broad patent, effectively stating that the DNA-editing technologies used in prokaryotes and eukaryotes were “patentably distinct”. Specifically, it had ruled that there was no interference-in-fact, i.e. that UC’s general description of use of CRISPR in biological systems could not have anticipated, under reasonable circumstances, Broad’s more specific CRISPR invention for use in eukaryotes. An interference-in-fact check is pegged on a so called ‘obviousness review’. A 1966 SCOTUS case defines four factors using which it can be undertaken:
(1) the scope and content of the prior art; (2) the differences between the claims and the prior art; (3) the level of ordinary skill in the art; and (4) objective considerations of non-obviousness
UC decided to appeal the PTAB’s verdict with the US Court of Appeals of the Federal Circuit (CAFC). The latter came to its decision on September 10, ruling in favour of the USPTO and upholding the Broad patent.
What happens next?
A lot of things. Let’s classify them as financial, academic, legal and administrative.
Financial – Where there’s a patent, there’s money. However, there’s more money for Broad than for UC because almost all application of the CRISPR technology will happen in eukaryotes, a domain that includes humans and plants. And because the Broad patent has been upheld, this effectively means the UC patent can apply only to prokaryotes and not to eukaryotes.
Public attitudes to this affirmation were partly reflected in the share values of three companies intent on commercialising CRISPR tech: Crispr Therapeutics AG (cofounded by Charpentier) and Intellia Therapeutics have licenses with UC and their shares fell by 5.3% and 2.5% respectively; Editas Medicine Inc., which has licenses with Broad, climbed by 6.8%.
A review in 2017 stated that although “CRISPR IP ownership is claimed by at least seven different parties”, the Broad patent could be a “blocking patent” because of its ancestral nature. This is one reason why the Broad Institute has already issued 13 licenses, more than any of the other patent-holders. In all, the review estimated that the American gene-editing industry will be worth $3.5 billion by 2019, with CRISPR propelling biotechnology to the status of “second highest funded sector in the United States”.
Academic – The contest between UC and Broad has only worsened the mutual, and deleterious, embitterment between the institutions. In 2015, Broad launched an acrimonious campaign to turn public opinion in its favour, which included attempts to rewrite the history of DNA-editing research and present Zhang’s achievement in stronger light. The possibly most damaging thing Broad did was to quote UC’s Doudna herself as having expressed frustration and doubt about whether a CRISPR system for use in bacteria could be adapted for use in eukaryotic cells.
These quotes were used in Broad’s filings for the patent dispute, undermining UC’s case. However, scientists have argued that science is almost never free of frustration and that Doudna was also right to express doubt because that’s what any good scientist would do: lead with the uncertainty until something to the contrary could be demonstrated. However, Broad effectively penalised Doudna for being a good scientist – an action that Michael Eisen, a biologist in Doudna’s department at UC, has said is rooted in universities being able to profit from patents created with taxpayer dollars.
Legal – It’s important to recognise what UC has actually lost here. UC appealed the PTAB verdict, bringing it to the CAFC, who in turn ruled that the PTAB had not wronged in its conclusion. The judge did not reevaluate the evidence and did not hear arguments from the two parties; no new evidence was presented. The court only affirmed that UC, in the eyes of the law, did not have grounds to contend the PTAB verdict. A salient portion from the judgment follows, where the judge writes that some parts of the CRISPR/Cas9 system as used in prokaryotes could have been adapted for use in eukaryotes but that that’s besides the point (emphasis added):
UC expended substantial time and effort to convince this court that substantial evidence supports the view it would like us to adopt, namely, that a person of ordinary skill would have had a reasonable expectation of success in implementing the CRISPR-Cas9 system in eukaryotes. There is certainly evidence in the record that could support this position. The prior art contained a number of techniques that had been used for adapting prokaryotic systems for use in eukaryotic cells, obstacles adopting other prokaryotic systems had been overcome, and Dr. Carroll suggested using those techniques to implement CRISPR-Cas9 in eukaryotes. We are, however, an appellate body. We do not reweigh the evidence. It is not our role to ask whether substantial evidence supports fact-findings not made by the Board, but instead whether such evidence supports the findings that were in fact made. Here, we conclude that it does.
Therefore, this is a judgment of the law, not a judgment of the science.
According to Jacob Sherkow, a professor at the New York Law School, UC can either petition the CAFC for a rehearing or appeal to the Supreme Court. Sherkow added that neither strategy is likely to work because he doesn’t think “this case presents any *novel* legal issues” (emphasis in the original). This means UC will likely return to the patent office and attempt to “salvage what they can from their patent application”.
There is also another problem. To quote Chemical and Engineering News,
… it recently became clear that another CRISPR scientist, Virginijus Šikšnys of Vilnius University [Lithuania], filed a patent for CRISPR/Cas9 just weeks before UC Berkeley filed its patent in 2012. While UC Berkeley and Broad were entangled in their dispute, the Šikšnys patent was approved and made public, meaning that USPTO can now hold the Šikšnys patent against UC Berkeley. “That has the potential to sink whatever is left from Berkeley’s patent application,” Sherkow says.
Administrative – This part is confusing. In the US, the USPTO upheld the Broad patent in February 2017. But in Europe, the European Patent Office (EPO) ruled in favour of Doudna and Charpentier in March 2017. So depending on the jurisdiction, companies that want to commercialise CRISPR technology (for eukaryotes) will have to work with UC in Europe and the Broad Institute in the US. At least one company, DowDuPont, which is using CRISPR to engineer corn and soybean crops to be cultivable without pesticides, has purchased licenses with both institutions.
The different judgments arise from one difference in how the EPO and the USPTO evaluate ‘no interference-in-fact’. According to a May 2017 report by Sherkow, “In Europe, one is entitled to a broad patent on a new technique, if it demonstrates an ‘inventive step’ over prior methods, even if there [is] no guarantee that it will work for all of its claimed applications.” In the US, on the other hand, each “claimed application” has to be demonstrated and is separately patentable if one application doesn’t follow obviously from the previous. The EPO decision is open to challenge and Broad is likely to use the opportunity to do so.
By the way, the country with the second-most patents related to CRISPR is China, after the US. Chinese research institutions and industry players have been focusing mostly on knockout mechanisms of CRISPR, which control how undesirable genes in a DNA sequence are removed. To quote at length from the 2017 review,
The Chinese government has been actively involved in gene-editing funding. The National Natural Science Foundation of China (NNSF), invested $3.5 million in over 40 CRISPR projects during 2015. Through the NNSF and the National Basic Research Program, the Chinese government has funded the first use of CRISPR for the modification of human embryos. Additionally, Shenzhen Jinjia Color Printing Group Co., a public company, has pledged $0.5 million to fund Sun Yat-sen University for studying CRISPR in embryos.
What do scientists say?
Scientists’ reactions are still coming in, although no consensus is likely to emerge soon. In the meantime, awards make for a reasonable proxy to determine what scientists think is laudable. On this count, Doudna and Charpentier are clear leaders. Since 2014, Doudna has won 20 awards (excluding one from UNESCO), Charpentier has won dozens and Zhang, nine (although must be noted that Doudna and Charpentier have been scientists for longer than Zhang has). Doudna, Charpentier and Šikšnys were also jointly awarded the 2018 Kavli Prize in Nanoscience.
The Wire
September 11, 2018
Idea of a country
The short excerpt below from Patriots and Partisans by Ram Guha caught my attention because it offers a simple definition of the idea of India (at the risk of oversimplification). One may have encountered it recently in Amartya Sen’s Argumentative Indian as well but that book was too dense for me. As such, this ‘idea’ is hard to come by in the media because, as Guha explains, it is unglamorous and difficult to sell, even as the press is the institution responsible for the viewpoint’s day-to-day distribution and maintenance. As a result, you get partisanship more often than deliberation. It’s easier.
It might be useful to clarify here that such deliberation is not between leftwing and rightwing vantages but between reason and reason. Unreason has no place here, nor does its conflation with partisanship. And it is often the case that the right is aligned with pseudoscience and illogic that it confuses resistance against unreason with resistance against itself. It’s frustrating not because the pedantic distinction is lost in the muddle – who cares – but because the result is often that the leftists co-opt your turf simply because you’re not right enough.
Anyway:
The groups and individuals mentioned in the preceding paragraph are, of course, merely illustrative [Ela Bhatt, Jean Drèze, etc.]. The work that they and others like them undertake is rarely reported in the mainstream media. For, the task of reform, of incremental and evolutionary change, is as unglamorous as it is necessary. It is far easier to speak of a wholesale, structural transformation, to identify one single variable that, if acted upon, will take India up and into the straight high road to superstardom. Among the one-size-fits-all solutions on offer are those promoted by the Naxalites, whose project is to make India into a purer, that is to say more regimented, version of Communist China; by the Sangh Parivar, who assure the Hindus that if they discover their religion they will (again) rule the world; and by the free-market ideologues, who seek to make India into an even more hedonistic version of the United States of America.
Based as it is on dialogue, compromise, reciprocity and accommodation, the idea of India does not appeal to those who seek quick and total solutions to human problems. It thus does not seem to satisfy ideologues of left or right, as well as romantic populists.
Featured image credit: Samuel Zeller/Unsplash.
Criticism of ISRO
The Statesman‘s editorial on India’s human spaceflight programme ends with the following line:
Only after placing the seventh and the last satellite in the NavIC system costing Rs 1,400 crore did ISRO realise the atomic clocks in the satellites had become dysfunctional, rendering the fleet a dud.
This line is wrong.
- ISRO could not have realised some satellites in the constellation were having issues with their atomic clocks any earlier, so the ‘only’ is misplaced
- Only two satellites out of seven were having issues with their clocks, and ISRO has made efforts to replace them.
- The fleet was rendered unusable but that may not be fair to say given it was temporary. The dysfunctional instruments have been replaced and the constellation currently awaits operationalisation.
There may have been other issues with the IRNSS but the last line in the editorial wasn’t it.
Another example: many people are of the impression that Narendra Modi’s announcement on August 15, that India will launch a human into space by 2022, caught ISRO chairman K. Sivan by surprise. This is true – but it was only the announcement that caught Sivan by surprise, not the ambition itself. ISRO has been preparing for human spaceflight for over a decade now. It is certainly not the sort of ambition that can be prepared for and achieved in four years.
However, NewsBytes said:
Commenting on Modi’s announcement, ISRO chief K Sivan had then said, “It came as a big surprise to us.” Yet, the question is, should it have come as a big surprise to ISRO? Logic says no. ISRO, being the agency responsible for all of India’s space missions, should ideally have been consulted before lofty promises were made. But, Modi went ahead with it anyway. Given the lack of notice, ISRO is now engaged in frantic attempts to recruit astronauts, improve existing technology, and develop new technology to meet the deadline set by Modi.
I’m glad more newspersons are writing critically of ISRO. But space is a sector where there’s very little low-hanging fruit that can be plucked and juiced into a political analysis, so there’s a lot more work required to separate a critique of ISRO from chest-thumping and render the former meaningful.
§
Random thought: Facts can be assimilated into a bundle and bundles lend themselves to interpretation. Now, there’s bound to be a correlation between between the facts-to-interpretations (F-I) ratio and the correctness of news coverage. The larger the F-I ratio is, the more likely it is going to be find more small mistakes in multiple news reports (i.e. on the topic of those facts) and big mistakes in a few – i.e. bigger range. On the other hand, the smaller the F-I ratio is, there are likely to be fewer smaller as well as bigger mistakes – i.e. smaller range. Now, by comparing these two ranges across press coverage of a variety of technical subjects where quantitative answers are common (e.g. in physics but not in sociology), and using normalised values of F-I if necessary, would it be possible to elicit the relative strengths and weaknesses of the mainstream media among those subjects?
Featured image credit: Oleg Laptev/Unsplash.
How do you determine the naturalness of homosexuality?
“Homosexual carnal intercourse between two consenting adults” is legal in India now. It wasn’t for lack of reason or scientific data that the item of legislation that rendered sodomy illegal – Section 377 – had been retained for so long. Instead, it was more a question of whether sodomy offended public decency and morality. On September 6, the Supreme Court of India said no consensual sexual act between adults, whether of the same gender or otherwise, could be considered illegal or offensive to public decency or morality.
The US had this moment in 2003, but there, science did play a role. In the landmark case Lawrence v. Texas, the Supreme Court of the US was able to rule that homosexuality was not a sin against nature on the back of a growing body of evidence that homosexuality exists in nature. More broadly, science helped determine the construction of sexuality in human and non-human species and rescued it from the chokehold of religious ideals and the stigma it carried. CJI Dipak Misra and Justice A.M. Khanwilkar may not have laboured through the scientific evidence in their own judgment but the veins of rationalism are evident in their syntax. Consider this excerpt (from the full; emphasis added):
What nature gives is natural. That is called nature within. Thus, that part of the personality of a person has to be respected and not despised or looked down upon. The said inherent nature and the associated natural impulses in that regard are to be accepted. Non-acceptance of it by any societal norm or notion and punishment by law on some obsolete idea and idealism affects the kernel of the identity of an individual. Destruction of individual identity would tantamount to crushing of intrinsic dignity that cumulatively encapsulates the values of privacy, choice, freedom of speech and other expressions. It can be viewed from another angle. An individual in exercise of his choice may feel that he/she should be left alone but no one, and we mean, no one, should impose solitude on him/her.
From another part of the same judgment:
It is submitted on behalf of the petitioners and the intervenors that homosexuality, bisexuality and other sexual orientations are equally natural and reflective of expression of choice and inclination founded on consent of two persons who are eligible in law to express such consent and it is neither a physical nor a mental illness, rather they are natural variations of expression and free thinking process and to make it a criminal offence is offensive of the well established principles pertaining to individual dignity and decisional autonomy inherent in the personality of a person, a great discomfort to gender identity, destruction of the right to privacy which is a pivotal facet of Article 21 of the Constitution, unpalatable to the highly cherished idea of freedom and a trauma to the conception of expression of biological desire which revolves around the pattern of mosaic of true manifestation of identity. That apart, the phrase ‘order of nature’ is limited to the procreative concept that may have been conceived as natural by a systemic conservative approach and such limitations do not really take note of inborn traits or developed orientations or, for that matter, consensual acts which relate to responses to series of free exercise of assertions of one‘s bodily autonomy. … It is urged that the American Psychological Association has opined that sexual orientation is a natural condition and attraction towards the same sex or opposite sex are both naturally equal, the only difference being that the same sex attraction arises in far lesser numbers.
Many of these arguments hinge on what it means to be natural. But what is nature, and what is naturalness*? The Wikipedia article on homosexual behaviour among animals carries an instructive line in this regard, and vis-a-vis the tenet of peccatum contra naturam (Latin for “sin against nature”): “The observation of homosexual behaviour in animals can be seen as both an argument for and against the acceptance of homosexuality in humans.” It’s ‘for’ because if animals do it, then it’s natural; it’s ‘against’ because humans are not meant to be like other animals. It’s a ridiculous position to be in. I find a quote originally about economics to be useful here:
… if background conditions determine, in a way which in principle falls outside a theory, what counts as the events over which the theory ranges, the theory is at the mercy of changes in these conditions which at any moment can undermine the predictive power of the theory.
The philosopher Richard Norman had intended to develop a theory that could predict how much, rather what kind of, resistance certain technologies would meet from certain cultures based on what traditions each technology appeared to offend. He succeeded in that he was able to explain why some cultures struggled, and continue to struggle, with the acceptability of technologies like vitro fertilisation and contraception, and what the latter might have in common with homosexuality. He pegged it on background conditions. Russell Blackford, a philosopher at the University of Newcastle, Australia, summarised Norman’s thesis thus in a 2006 review (emphasis added)
According to Norman’s approach, anything that may threaten a culture’s basic assumptions about how ordinary human life works – especially assumptions about sex and its relationship with conception and birth, the development and rearing of children, the roles of men and women, the processes of ageing and death – is likely be disquieting to at least some people. For example, homosexual practices may seem to threaten a background condition that relates to sex and procreation. If there are recognised choices that include sexual acts with no possibility of pregnancy, then one of the background conditions has been lost.
Blackford writes in another part of the review (emphasis added):
Norman argues that the discomfort that some people feel about IVF and futuristic prospects such as that of biological immortality comes from a sense that important background conditions to choice – relating to procreation and death – are threatened. In this context, a “threat” to the background conditions seems to mean that certain conditions may no longer pertain. A sense that some background conditions are under threat can be expressed as a claim that nature is being interfered with. When such claims are made, nature is being equated with the background conditions recognised within the culture concerned. Norman, however, defends IVF on the basis that incremental changes to our own culture’s background conditions can be absorbed into our thinking.
However, given that the Bharatiya Janata Party has refused to issue a statement on the historic SC verdict, signalling its moral ambiguity (at the very least) on the subject, it seems unlikely that the party’s members – i.e. the country’s ministers – will be open to making incremental changes in their worldview to accommodate the naturalisation of “unnatural” sexual acts, so to speak.
*Not to be confused with the naturalness of particle physics, or maybe it is.
The Wire
September 7, 2018
Featured image credit: gagnonm1993/pixabay.
DotA redux – AI loses
What happened
An artificially intelligent (AI) gaming system built by a company co-led by Elon Musk took on some of the best human players at a game more complex than chess or Go last week – and lost. AI losing is not something you hear about in the media often because it doesn’t lend itself to interesting discussions about how the AI was better. This is partly because the game in question, Defence of the Ancients (DotA) 2, is like little else that AI has attempted to master.
In late June, the AI, named Five, played against human players at DotA 2. Five’s builder is OpenAI, a California-based non-profit organisation helmed by Musk and Sam Altman, and it was created just to play DotA 2. This is a popular video-game in the style of a battle arena in which two teams engage in combat to take down each other’s base.
A team has five players (hence the AI’s name), each of whom can play as a ‘hero’ selected from over 100 options before the game’s start. Each hero in turn has various unique abilities that can be developed through the game to attack, defend, support other players, etc. When the hero kills a (playable or non-playable) character during the game, she earns gold that can be used to purchase items that enhance the hero’s abilities. A game typically lasts around 45 minutes.
In the three June games, Five demolished its semi-pro human opponents. However, there were three factors that enabled this outcome. First, the benchmark games were heavily rigged such that many of the features of regular human gameplay were disabled. This was because the benchmarks were the first that Five was playing against humans. Before that, including when OpenAI had taught it to play DotA 2, Five had only combated itself.
Second, as a result of the limited featureset, the human players couldn’t deploy the strategies they were used to. As one Reddit user put it, “right now, it is sorta like the two teams are trained in two different games – but the match is being played in the game in which the bots have the most experience.” Third, irrespective of the denial of features, Five – by virtue of being a sophisticated computer – sported faster reaction times than its human counterparts, possibly availing an advantage human players can’t ever do so themselves.
So what did the June games benchmark for? Mostly, Five’s preparedness to play against professional human DotA 2 players at The International, the equivalent of the football world cup for the game. These ‘regular’ games had far fewer restrictions, including the removal of some of the major ones from last time, and featured human players considered to be among the best in the world. And this time, Five lost the two games it played.
Five’s first loss was on August 5 against paiN Gaming, a team from Brazil ranked in the world’s top 20 but which had been eliminated early on in The International. Five’s second loss was scripted later in August by an ensemble of Chinese ‘superstar’ players (xiao8, BurNIng, rOtK, Ferrari_430, SanSheng).
According to an OpenAI blog post, Five “maintained a good chance of winning for the first 20-35 minutes of both games.”
These losses don’t present straightforward takeaways. Thanks to the multidimensional nature of DotA, there are numerous ways to analyse how Five played and little victories, as well as little mysteries, to be salvaged from the ruins. Here are four.
Team-play: The five heroes controlled by Five seemed reluctant to split up. The DotA 2 battle arena is very big (in terms of how much time any hero takes to traverse it), so maintaining an advantage requires heroes to split up when necessary, undertake patrols, plant wards (stationary items that act like sensors) and frustrate/deter opponent heroes. However, Five’s heroes tended to stay together even when it wasn’t necessary, opening themselves up to some attacks that the humans exploited. On the flip side, whenever the moment came for the heroes to play as one, Five’s coordination and execution was flawless. It is possible Five thought that unity was the better overall strategy considering its heroes (and opponents) had been selected by an independent group.
Mortal couriers: Five didn’t seem ready for single couriers. In DotA 2, a courier is a diminutive FedEx-like service that goes between a shop and the hero, carrying items that she might need. The benchmark games in June had five invulnerable couriers that couldn’t be killed by opponent heroes. The games played on the sidelines of The International, however, allowed for one very mortal courier. Five’s heroes seemed unprepared to deal with this (virtual) reality because they constantly played as if they would be able to access the items they needed wherever they were. In a normal game, heroes usually abandon the battlefield and fall back to pick up the items they need if they don’t have couriers.
Reaction time: After the benchmark tests, some players (not involved in the games against Five) had expressed concern that Five might be getting ahead by ‘unnaturally’ faster decision-making and computational abilities, mostly in the form of reaction time. The adult human reaction time is between 150 ms and 500 ms depending on the stimulus, physiology and task. According to OpenAI, Five had had a reaction time of 80 ms, since increased to 200 ms to appease these concerns. However, this does not seem to have made Five appear particularly weaker.
Brain freeze: There were multiple instances in both games where Five made decisions that put it in a weaker position even as human observers were able to quickly identify alternative courses of action that could have helped the AI maintain its dominance. A notable kind of these errors was a Five hero using a very strong attack spell against an opponent too weak to deserve it. When a hero uses a stronger spell, the longer she has to wait for it to recharge so she can use it again.
Why you should care
Games like DotA 2 present unique, low-cost opportunities to train AI in multiplayer combat with a colossal number of possible outcomes. However, a bigger gain in the context of Five is the techniques that OpenAI is pioneering and which the company says can be deployed in a variety of real-world situations. Some of them include the objects of interest called Rapid, Gym and competitive self-play, each building on the next to deliver sophisticated AI models.
Rapid is the name of a general-purpose reinforcement learning algorithm developed by OpenAI. It uses a technique called proximal policy optimisation (PPO), which in turn is a form of policy gradient method, to make and reward decisions. Policy gradient methods are used to help machines make decisions in environments in which they don’t have enough data to make the best decisions and when the machines have limited computational power to work with. OpenAI researchers published a preprint paper in July 2017 describing PPO as a gradient method that “outperformed other online policy gradient methods” while being simpler to implement.
The algorithms born as a result are generally classified as agents, and their performance is tested as they play out within an ambient context known generally as the environment. The latter role is played by Gym. That is, Rapid is tested with Gym. To their credit, OpenAI has made the Gym development library freely available to all developers and has qualified it to work with any numerical computation agent (including Google’s widely used TensorFlow). This means any engineer can build an agent and use Gym to test its performance.
When multiple agents made from the same algorithm are tested in a common Gym environment, the result is competitive self-play. This is how Five trained to become better at DotA 2 by playing against itself: two different agents from the same Rapid mould engaged each other, each in pursuit of its own triumph. And to force each version of Five to become better, the researchers only needed to identify the trajectories of success and then add the necessary rewards and sanctions to guide the Fives down those paths.
While DotA 2 may be too complicated to visualise this, consider the following video, where an agent needs to step over a line blocked by another agent, who has the same mandate, in order to win. The first two runs produce the same results, with the agents setting off for their respective lines in an unplanned race. The third run onwards, something else happens…
https://drive.google.com/file/d/0B2EZBre2hp1_RFJpRXUxc3JSWDg/preview
(More examples here.)
Competitive self-play is becoming increasingly popular. The other very-famous AI that trained this way is AlphaGo, the machine that beat the world’s foremost Go champion last year, including one game in which it developed previously unknown strategies to win. Keep in mind this is a game humans have been playing for about 2,500 years and which AlphaGo taught itself in three days. Competitive self-play is especially useful either when engineers want the agent to acquire a skill for which they can’t create rewards in a given environment or when they want the agent to develop new ways to solve an old problem.
So as a consequence of the way it is designed and trained, OpenAI Five is expected to be all the more powerful before The International next year, with some gamers predicting that it will take the crown. But more than disrupting the human kingdom of DotA 2, Five’s education can prepare it for even more complicated tasks. For example, OpenAI has also been developing a robotic hand to be controlled by AI, and the hope is to make it as agile and dexterous as a human hand. Thereon, it becomes a matter of questions: “Five, what is the best way to return a serve?”
OpenAI is one of the two major billion-dollar non-profit organisations in the West focusing on using AI to solve humanity’s problems. The other is the Allen Institute for Artificial Intelligence, founded by former Microsoft bigwig Paul Allen, with the express intention of making “scientific breakthroughs”. However, OpenAI is in a uniquely sticky situation vis a vis its resources. As Miles Brundage, an AI expert recently employed by OpenAI, wrote on his blog in 2015:
With Amazon’s investment and Tesla’s indirect involvement through Musk, it would seem that OpenAI will potentially have access to a lot of the data needed to make deep learning and other AI approaches work well. But these are different sorts of datasets than Google and Facebook have, and may lend themselves to different technical approaches. They also raise the question of proprietary data – how will OpenAI balance its push for openness with this proprietary access? Will it release code but not data? How will its experiments be replicable if they are tailored to particular data streams? How will privacy be addressed? Some of these questions are unique to OpenAI and others aren’t. However, they’re the sorts of questions OpenAI will need to answer.
The Wire
September 3, 2018
Two books
When thoughts turn stale, read a book. I’m reading two at the moment, which is telling.
Patriots and Partisans by historian Ram Guha, an accessible narration of the historical roots of modern Indian national politics.
Flights by Polish writer Olga Tokarczuk, a work of fiction that I noticed because of its beautiful cover and then for its compelling blurb: part musings, part storytelling, on the connections between human anatomy and travelling.
I picked both these books up at the Bangalore airport after noticing an ad for Peps mattresses, and how – per the brand’s claim – it “cures jet lag”. What nonsense.
I fished my phone out of the pocket, took a picture of the banner and began to blog… noticing a couple lines later that it’s just more of the same.
Then I remembered some quote I’d read somewhere by someone famous that you can’t hope to produce good work without reading good writing. Possibly more so for someone who expresses the entirety of his creative potential through words.
So the books. Both of them are gripping – no surprises with Guha and lots of surprises with Tokarczuk. Flights‘ is an intriguing premise, you must admit. Let’s see how they go.
Oh, I also remembered a recent oped in The Hindu by Mini Kapoor, about how reading on the internet has likely diminished literate humans’ capacity for reading large texts and letting complex ideas ferment in the mind instead, as usually happens, of flitting between shorter texts and constantly recycling (allegedly original) opinions.
This isn’t an original argument. I myself have felt that Twitter – as a microblog – has probably made blogging harder.
Nonetheless, here we are. And here we go.
Satire: Ducks in water could increase oxygen content
According to Biplab Deb, the chief minister of Tripura, the oxygen content of water will increase if ducks swim in it. [Satire begins here] This is a sensational new discovery that has drastic implications for Earth’s future. The lawmaker’s thinking suggests it might have something to do with duck-farts.
There are thousands of water bodies around the world where masses of ducks have been swimming for tens of millions of years, and which could now be flush with oxygen. As a result, our planet now appears to be due for major bio-ecological changes as the abundance of oxygen is likely to spur cascading evolutionary effects.
In fact, it is being speculated that, in the aftermath of Deb’s confirmation, the sixth extinction of the Anthropocene epoch might just be halted in its tracks and forced to do a volte face; all it will take is lots of ducks. This might explain why oil companies in Texas are confident that their proposal to have the government erect a $12-billion ‘sea wall’ to protect their coastal facilities against rising water levels will be taken seriously.
At the same time, there also appears to be growing public resentment against scientists, with people wondering whether supposed researchers spending tax dollars might have kept this simple solution away from governments in an effort to maintain their self-importance. Major news publications like OneOp are reporting that this could be an urban naxal conspiracy and that a concerned ministry is expected to conduct raids soon. IndiaIndiaIndia reported that there’s a joke somewhere in here about going quack.
According to Indian Express, it appears Deb had also discovered that the ducks would recycle the oxygen in the water and prevent its molecules from going to waste. Thankfully for the minister as well as for the rest of us, oxygen molecules don’t affect the pH value of water, or we would also be confronted with a major acidity/salinity catastrophe. In all, it’s good news for everyone, including the people who will supply the 50,000 ducklings Deb says he will distribute among Tripura’s fisherfolk.
A senior scientist who didn’t wish to be named expressed surprise at the finding, and said he had applied for a grant to study the molecular chemistry of duck-farts. “I expect to hear back in five years,” he said. The same individual also expressed regret later. “We all had a chance to find this out before but we did not. It’s because we didn’t study the Vedas as thoroughly as we should have. Hopefully we will learn from this mistake. Om.”
Featured image credit: Ryk Naves/Unsplash.
The editor, the writer and the exclusivity
Do all editors share that excoriating sensation that carves through one’s gut when one sees one’s writer writing for another publication?
Of course, “one’s writer” is unjust phrasing because writers don’t belong to editors and are free to write for whoever they please.
However, this isotropic sea of writer-editor relationships is more suitable for the writer than it is for the editor.
Editors build publications – that means responsibilities that don’t look all that beautiful under the light of day.
One such responsibility is to build exclusivity, which is the notion that in order to read writer W, you must read publication P.
If the writer is good, then the editor is implicitly required to encourage the writer to write more, to help them become better, without encouraging diversification.
For writers, however, it is more desirable to diversify, to write for more outlets so that their name is accessible to the public memory through a variety of sources.
Depending on the writer’s choice of outlets, it also means diverse audiences to speak to and diverse opportunities to access.
Most good writers share healthy relationships with the editors they regularly work with, which often leads to long-lasting friendships.
So a friendly editor might encourage a writer to diversify if he has the liberty to think beyond the needs of his publication.
I don’t know how often this is or isn’t the case, but when it is, TFW you see a writer belonging to your vision of exclusivity writing for a competitor is pure pain.
To an outsider, the exclusivity drive may seem like an unfair, even unnecessary, ideal because it acts against the interests of the writer.
Eminently, exclusivity seems opposed to nurturing a community of good writers capable of weathering risk as well as to the democratisation of journalism.
The latter is in that all people have access to all stories because all journalists write for all publications.
In fact, this would be the most efficient way, in all senses of the term, to conduct journalism as a business – as an open, social enterprise.
However, this worldview conveniently forgets that competition exists and extant journalism organisations are founded on market forces, not social good.
Exclusivity exists because competition exists: it is the conservation law born when the competition orders the system along a new symmetry.
There are two forms of exclusivity: of stories and of writers; but where one exists, the other will too, which exposes the distinction for its pedantry.
The point is you, the editor, need both forms in order to orient your organisation towards profitability; without exclusivity, you have no edge.
And of course, competition is as competition does, putting editors at odds with their writers in a proxy battle for the organisation’s conquest for new edges.
The question now is whether writers should be expected to understand all of this and act in a way that protects the editor’s interests to the extent possible.
This is because, if exclusivity is on the table, editors are also likelier to give the writers more leeway in terms of what to write and how.
One might argue that a good editor will always be able to wrangle a good story out of a writer.
However, it is not possible to understate the importance of a long-standing editor-writer relationship and what that can produce.
Perhaps journalism would be more benefited by an arrangement where all editor-writer pairs can work for all publications, instead of just writers.
This is in effect a framing of the ‘story’ as the nuclear unit of journalism instead of the storyteller, an identity that excludes the editor.
Such an arrangement would protect the interests of the editor as well as the writer, and ensure that publications can also produce better stories as a result.
Of course, this would mean more work for the editor but it could also conceivably mean more business.
Featured image credit: Rafaela Biazi/Unsplash.
Metastable 