Viswanathan K's Thoughts

Recently I read a couple of articles from two different parts of the world – India and the UK – that recommended randomizing selections in various fields. Our instinctive reaction to such a suggestion is negative, as Tim Harford wrote : “We do not usually draw lots to allocate duties, jobs or privileges.”   I want to state at the outset that both articles are not saying that selections (for whatever field) should be entirely random. Rather, they mean set a minimum criteria or qualification that needs to be met. Then, from amongst the shortlisted entries, select at random – not by ranking them as 1 st , 2 nd and so on.   University grants, says Harford, could be given at random (provided the applications meet the minimum criteria). After all, he argues, sometimes a thorough evaluation of each application (to decide on merit) can end up costing a significant chunk of the grant amount itself! In any case, he says, the assumption that experts know best (and can rank ideas and

We think we know our solar system. Obviously not every single object and rock, but at least the categories of objects in it. Sun, planets, moons, comets, asteroids, asteroid belt, meteorites, we’ve covered everything, right?   Not entirely, as Latif Nasser found out. His discovery started off from a kiddish solar system poster on his 2 yo son’s wall. As per that poster: “Venus had a moon called Zoozve. ” What, he thought? Among the rocky planets (Mercury, Venus, Earth, Mars), wasn’t Earth the only one with a moon?   The NASA site said Venus had no moons. He googled up Zoozve. No results, at least not in English. That could have been the end of it – an error in a kid’s poster. But Nasser didn’t stop. He called up the illustrator of the poster – the man said he didn’t cook it up, that he found Zoozve on some online list of moons.   And then Nasser’s friend at NASA called him back. It wasn’t Zoozve, it was 2002-VE, an actual object near Venus. That fit – the “2” may have

In the bad old days of the Internet, finding what you were looking for was, well, like searching for a needle in a haystack. Clay Shirky called it “filter failure”. And now, the pendulum has swung to the other extreme of “filter success” as Nicholas Carr calls it. (Think of the effect of Google search, or the recos of Amazon and Netflix). But that’s just created a new problem, writes Oliver Burkemann - the “Too many needles” problem: “(You have) a large pile (or digital equivalent) of books or articles you've been meaning to get around to reading, plus maybe a long queue of podcast episodes to which you'd love to listen, if only you had the time.”   Burkemann says this is a broader problem most people face. It occurs outside the digital/ Internet domain, in pretty much every aspect of life: “If you're blessed with work you love, or a creative passion you're good at, you may often feel torn between multiple projects you're excited to launch. Others are the f

That UPI has been very useful and successful is unquestionable. But valid questions exist about its viability, specifically on the “Who pays for it?” topic. After all, it costs money for PhonePe or GPay or the banks or the government-owned backend system behind it to develop, maintain and keep things running.   When it was launched, most people had assumed a fee would be charged for its usage. Either as a percentage of the transaction, or as a flat amount. But in 2019, writes Rahul Matthan, the government passed a law prohibiting banks and service providers from charging any fee for it.   There were good reasons behind that decision. For high value transactions, history had shown that shops are willing to foot the fee because nobody carried that much cash and the benefit of closing a sale was worth the fee. But for low value transactions, the fee became a significant fraction of the profit, which is why credit cards never caught on in most places in India. “It is to address

Raghu S. Jaitley makes an interesting observation about the upcoming Lok Sabha elections. On the one hand, almost everyone says “it appears like the most pointless LS election in living memory”. The BJP’s victory appears to be a foregone conclusion (Modi talks of winning 400 seats), and the opposition seems to be in disarray with endless arguments over seat-sharing arrangements. “It is either the failure of the political imagination of the opposition leaders or the relative success of this government that we might not have any overarching theme that serves as the connective tissue to the multiple state or local issues.” Which is why Jaitley says: “While I can foresee lots of drama and rhetoric in campaigns at the state level, this will be a ‘no wave’ election, which means it will be impossible to dislodge the incumbent.”   And yet, there are other signs and actions of the BJP that “belie this optimism”. “The eagerness to stitch alliances across states and political spectru

Pranay Kotasthane’s post on the history of how the water situation in Bangalore has evolved was highly informative. “First, half of the city’s water supply comes from Kaveri, which is 90 km away and 350 m lower.” That “350 m lower ” part is very significant. First, it means the water has to be lifted. The electricity bill for lifting comes to ₹3 crores per day. Second: “Before independence, Bengaluru, due to its elevation, relied largely on the rainwater stored in nearby lakes. But as the city population grew, the engineering marvel of pumping water up from the Kaveri was dreamed and realised.” In turn, that led to a decrease in the dependence on lakes. As the lakes fell into disuse, encroachments began to increase. Since the lakes had begun to matter less (at that point), the pushback against encroachment was proportionally low.   Historically, the other half of the city’s water needs were met by groundwater. But, as the city began to grow with tech parks followed by apar

Is writing video games software harder than other forms of software development? In his biography of a handful of video games, Blood, Sweat and Pixels , Jason Schreier answers it with an emphatic Yes.   First, he points out, video games are interactive. Which means they don’t move in a linear fashion, unlike say an animated movie or a movie like Avatar . Secondly, the game consoles (e.g. Xbox, or PlayStation) have newer, faster, better hardware capabilities every year. If you don’t come up with a game that can use those newer capabilities, your game will look tame and old . But that hardware is still under development, which means things are continuously changing: “Making a game is like constructing a building during an earthquake.”   Third, games need other software tools like photo creators, and physics engines (you don’t want to program the laws of physics each time, do you?). These tools keep changing. Or the tools could do so much more with the improved hardware (see t

There was a recent paper that said that economic inequality in India has increased so much that it’s worse today than even during the British Raj – it’s the era of Billionaire Raj, it proclaims. Apparently the first version of this paper was published in 2017; and what we got now was the updated version.   Pranay Kotasthane isn’t convinced by the paper. He reminds us of a question that was asked about the first version – how does one find out the wealth of individuals? The original paper had used “survey data to estimate the incomes of the non-rich” and was “relying on income tax data to estimate the incomes of the rich”. Many had argued that the methodology was wrong on both fronts – (1) plenty of rich people don’t pay taxes (as we know all too well) and (2) who answers surveys honestly, let alone on a topic like how much wealth they have. Weirdly, the second edition of the paper doesn’t address these questions on its methodology, he points out.   Second, he says that while t

Tim Harford wrote a thought provoking piece on the many similarities, lessons and warnings from a form of fake’ness that’s been around for far longer than social media – art forgeries.   He bases this on things said and done by a master art forger named Eric Hebborn. In his autobiography, he declared that a Brueghel painting in the New York Mets Museum was a fake that he’d created. The museum wasn’t happy: “We don’t believe it’s a forgery, and we believe that the story told by Mr. Hebborn in this book is not true.” Was the museum lying because it didn’t want to admit it had been conned? Was the forger lying to grab headlines? How does one know what the truth is?   This problem applies to fake everything today. Is that video of a politician or actor saying or doing something real? Or is it a fake as those folks claim?   There are fakes and then there are deepfakes. “Deepfake AI is a type of artificial intelligence used to create convincing image, audio and video hoaxe

Recently, there were a lot of articles that tore into Foxconn’s factory at Sriperumbudur, which has a predominantly female workforce. Raghu S Jaitley summarizes the view of those articles: “Restrictive working conditions, poor food, low pay, long hours, semi-skilled work, lack of unions - you get the gist.”   Jaitley then makes an interesting comparison to how the call center and BPO industry was described two decades back: “The mental and physical toll on workers doing the night shift, long hours and low pay, the mental stress of faking an accent and getting abused by customers on the other end of the call, the low quality of work with no future prospects, the packing of workers in a cramped van that was used for commuting to the offices and the unhealthy food served in the canteens. And, of course, no unions.” Sound similar to the Foxconn factory? Fast forward, he says, to present day. BPO’s moved up the value chain, the kind of work improved, and career opportunities are

Everyone, left or right, feels the Internet is biased. Too much of what we see on Facebook, WhatsApp, Google search results, news sites etc is either biased or outright fake. The problem is even more dangerous with AI chatbots like ChatGPT. After all, you ask it a specific question and it generates an answer. But the answers it produces depend on the kind of input it learns from. And since the data it learns from is biased, what it learns and creates as answers is biased too.   What could be a solution to the bias problem? (I’ll leave fake out of this blog).   Google tried to address the problem in its AI chatbot named Gemini. By finetuning Gemini to give certain kinds of answers, and not give other kinds of responses. The intention was to “avoid creating or reinforcing unfair bias (e.g. sexism, racism etc )”, wrote Andrew Sullivan. But as Nate Silver pointed out: “Unbiasedness is hard to define .” As if that wasn’t hard enough, “cultural, social and legal norms” differ,

The last chapter in Ed Yong’s book, Immense World , is about the pollution we humans cause and its impact on the creatures around us. My initial though was, god, I hope he isn’t going to end this awesome book on a paternalistic, preachy note about industrial pollution and deforestation…   I needn’t have worried, because Yong does not talk of CO2, global warming, deforestation, poaching and the likes. Wait, so there’s other damage we do? Yes, it involves stuff most of us don’t even realize is a form of pollution too.   The first one is light pollution. Yes, light can be a form of pollution. “It is jarring for us to think of light as a pollutant, but it becomes one when it creeps into times and places where it doesn’t belong.” After all: “Light at night is a uniquely anthropomorphic phenomenon.” Until we lit up the planet, the “daily and seasonal rhythms of bright and dark” were fixed. Physicists knew this problem for long – it prevented them from seeing the universe –

A while back, I wrote about how intelligence and errors go hand in hand, and therefore, AI will make mistakes. This is a scary prospect for many people –we are going to entrust and empower AI in more and more matters, but they’ll inevitably make some mistakes?   Rahul Matthan describes an interesting idea on how we can make AI “safer”. By replicating the practice of the airline industry: “It is safer to sit in a plane 10,000 metres above sea level than in a speeding car anywhere in the world. Unlike every other high-risk sector, the airline industry truly knows how to learn from failure.” Individuals and companies learn from their mistakes. But the aviation industry is unique in this matter: “It (airline industry) has put in place mechanisms that not only ensure that the company involved learns and improves, but that those findings are transmitted across the industry so that everyone benefits.”   Therefore, argues Matthan: “If AI is as dangerous as so many people cla

In the last but one chapter of Immense World , Ed Yong reminds us of something that was easy to forget given the level of detail each chapter had on one particular sense. Animals’ don’t use just one sense, they use a combo. How do mosquitoes identify the individual to suck blood from? By sensing their body heat and the release of carbon dioxide. Scientists tried masking one of the two signals, and the mosquito would still home in on its, er, food source.   It is obvious why animals use multiple senses: “Each sense has pros and cons, and each stimulus is useful in some circumstances and useless in others.” Beyond complementing each other, senses can also “combine”. In humans, it is rare and we call it “synesthesia” (For people with this ability/condition, sounds have colors, words have tastes and so on). “This perceptual blurring is special among humans, but standard to other creatures.” How do we know this? Because the neurons in their brain that receive signals from one s

The West and the countries that are starting to get richer now, including China and India, don’t agree on how to address the problem of global warming. While everyone agrees that it is a problem, the “what” should be done, “how much” should be done by different countries, and how much the West should “pay” the still developing countries to switch to less polluting (but more expensive) technologies are highly contentious topics.   It was in that context that I was reading Rakshit Shetty’s post on China’s steps. While they, like India, argue with the West on all of the above, the Chinese know they can’t continue doing things they’ve always been done either. China today tops on both fronts: (1) largest CO2 and greenhouse gas (GHG) emitter; and also (2) largest capacity for renewables and hydroelectricity. It is also the second-largest for nuclear power and expected to be the leader in biofuel production.   The West can talk about putting a limit on how much CO2 and GHG they will

Magnetoreception refers to the ability to navigate using the earth’s magnetic field for context. A very interesting thought experiment on this topic was to change the magnetic field around such species and see if they changed direction. In practice, that’s easy to do with small animals, but how could one try that with something as big as a whale? Sounds impossible, except nature itself creates such situations, says Ed Yong in Immense World : “The sun periodically throws cosmic tantrums and produces solar storms – streams of radiation and charged particles that affect the Earth’s magnetic field.” Scientists collated 33 years of data on healthy whales that inexplicably stranded themselves onto beaches. They then compared that with the data on solar activity. And bingo! On days with the most intense solar storms, gray whales were four times more likely to beach themselves.   The magnetic field has two components. The first one is inclination , the angle at which the geomagnetic f

The semiconductor chip industry is now at the heart of geopolitics. America decided to squeeze two of its adversaries wrt access to semiconductor chips – Russia across the spectrum, and China only wrt high-end chips. To try and counter that, Russia smuggles chips from other countries; and China has decided to accelerate its in-country chip industry.   As Taiwan gets increasingly in the crosshairs of China, it sets off alarm bells everywhere because the world’s highest-end chip manufacturer is a Taiwanese company named TSMC. Cut off TSMC in the event of a war or siege, and the whole world will be impacted. America has therefore been pressurizing Taiwanese companies to move their chip manufacturing to other countries – US, Japan etc. Most wonder about the cost impact of moving things to a rich country, but here America’s geopolitical interests trump cost considerations.   India sees this as an opportunity to take baby steps in the chip manufacturing industry. Job creation aside, chips a

Electricity. Species like the electric eel use it. But for humans to understand them, they had to first discover and learn about the concept of electricity, explains Ed Yong in Immense World ! In fact, the study of electric fish is intertwined with the study of electricity. They inspired the design of the first synthetic battery. And fuelled the discovery that muscles and nerves run on electrical currents.   The electric organ on the fish’s tail is like a small battery. It creates an electric field around the animal. Current flows from one end of the organ, through the water, to the other end. All living things nearby act as conductors and increase the flow of current, while insulators (like rocks) reduce the flow of current. These changes affect the voltage on different parts of the fish’s skin. To draw an analogy with our sense of vision: “Conductors shine brightly… insulators cast electric shadows.”   Electric fishes use electricity for locating things. But, unlike light,

Everyone agrees on the need to switch to renewable energy sources – away from coal, petroleum etc. For obvious reasons. But, as Rahul Matthan says , any such switch also requires a re-think and overhaul of our power grid.   The major problem with renewable energy, aside from cost? “Renewable energy is notoriously fickle. All it takes for supply to be disrupted is slightly overcast skies or a subtle change in weather that becalms the winds. ” Solving that problem requires effective energy storage solutions to be established. Expecting only the power grid companies to accomplish this won’t work. And, argues Matthan, that is not that the best approach.   What does Matthan have in mind? For example, he says, take electric vehicles. Many of them can store more energy than they need – or maybe, after charging, it didn’t get used. A possible change here is as follows: “We need to… install a bi-directional charging unit so that EVs recharge their batteries whenever electricity

How bats navigate is well known – via sonar (sound waves) and their reflection. Correct, and it involves at least 10 challenges, lists Ed Yong in in Immense World : (1) distance. Sound loses energy quickly, so the bat can only detect objects at 6-9 yards; (2) volume. Bats “scream” at volumes of nearly 138 dB (that’s as loud as a jet engine). It must be deafening for the bat; (3) speed. It needs to send and process sound very quickly. After all, its prey will dodge and move; (4) overlapping. It would be easy to get lost in its own echoes. Which is the first echo? The third echo?; (5) multiple frequencies. Bats sound out at multiple frequencies – lower frequencies help identify large objects; higher frequencies with small objects. There is so much different info coming in via the echoes and it has to be made sense of; (6) adjustments. In the scan mode, a certain combo of frequencies and duration makes sense. Once a prey has been identified, a different combo is needed; (7) clutte