Category: Thinks

Arnold Kling: “A lot of advice for prompting seems to assume that you have only one shot at trying a query. I think that this is left over from the days of Google searching, where you had to get the wording right to get what you wanted. But with AI you can start out with a really low-quality prompt, and when the answer is off base you can explain what is wrong with it and continue the conversation. A conversation is an extended give-and-take interaction between you and the model. You may or may not have a specific idea of what you want to get out of the conversation. If you do have a specific objective, then you are trying to accomplish what you would with a query, while taking advantage of the fact that you don’t have to come up with the one best way to word your query. If you don’t have a specific objective, then you and the model can just explore a topic. You ask more open-ended questions until perhaps the conversation leads you to want to home in on something specific.”

WSJ: “In “Click: How to Make What People Want,” Jake Knapp and John Zeratsky examine what makes a product or service resonate with consumers. The authors outline a process for early-stage founders to follow in their quest to create the next revolutionary thing, beginning with what they call the foundation sprint: “drop everything and sprint on the most important challenge until it’s done,” they tell us, starting with “identifying your customer and a real problem you can solve.” Drawing largely on conventional business-school wisdom, Messrs. Knapp and Zeratsky—both former partners at Google Ventures—stick to well-trodden paths, such as finding “your team’s unique advantages” and ensuring product differentiation. These are good reminders for aspiring entrepreneurs. Also familiar will be the heralded 2-by-2 chart, which tidily distills complex decisions into pairs of factors, helping users quickly identify which quadrant they are in.”

Bloomberg: “From Airbnb’s new app icons to Apple’s “Liquid Glass” design, skeuomorphism is back and everything in the digital world has dimensionality again…[Brian] Chesky, an alum of the Rhode Island School of Design, decided to take the home-sharing site in the opposite direction. Feeling Airbnb’s mobile and web experiences had grown too flat, Chesky and team drafted a platform revamp that again embraced skeuomorphism. The term is commonly used to describe decorative computer graphics that ornately mimic physical objects — sacrilegious among Apple apostles who bristle at superfluous veneer. Yet after sharing early mock-ups of Airbnb’s stylized overhaul, Chesky was surprised that Ive especially loved its 3D visuals; he even pushed to imbue the app with more vibrant flourishes and dimensionality. “Jony was a proponent of moving away from flat design, to my shock initially,” Chesky recalls.”

Mint: “Of the 275 initial public offers (IPOs) in India since the covid lockdown, 35% have delivered negative returns  on their issue price. Compared to their listing price, which is the price at which shares start trading, almost half have delivered negative returns. About half underperformed the BSE 500 index’s return in terms of gains on their issue price and 64% in terms of gains on their listing price. Only 36% of IPOs over the past five years have been a worthwhile investment. Surprisingly, qualified institutional placements (QIPs) fared only marginally better. Of the 224 QIPs since the pandemic, only 99 have outperformed the BSE 500 index, giving these professional investors a success rate of 44%.”

WSJ: ““Dots and Lines” is about modern network science. “At their most basic, networks measure whether objects interact,” writes Mr. Bonato, a professor of mathematics at Toronto Metropolitan University and the author of “Limitless Minds: Interviews With Mathematicians.” The objects here—the dots and nodes, or vertices—could represent people or pretty much any entity. Their interactions are represented by lines, edges or links. The pages of Mr. Bonato’s book connect a veritable network of facts, figures and fascinations about networks: The first English use of the word “network,” we learn, was printed in the 1560 Geneva Bible (“a grate like networke of brasse”); in the 17th century the term described vein structures in plants; in the 18th century it was applied to railways, rivers and electrical cables. And then came the World Wide Web. “The internet,” Mr. Bonato reminds us, “was originally referred to by the less catchy internetwork.””

Economist: “India’s space programme…offers a rare example of successful public-private collaboration. In 2020 the government opened the space sector to private firms. Before that, companies could act as suppliers only to the Indian Space Research Organisation. A new agency, IN–SPACe, was created to grant private firms access to state-developed technology and infrastructure. So far, it has approved over 50 firms aiming to launch satellites and offer space-based services. Mr Forbes acknowledges the success of IN–SPACe but sees it as a special case—powered by a capable public agency and with a clear mandate. If India wants to be more than the world’s next assembly line, it needs to bring the same clarity of purpose that launched satellites into orbit to the far messier business of making innovation take off on Earth.”

Sophie Bakalar: “Consumer AI isn’t about chatbots. Not the way we’ve come to know them, anyway. It’s about escaping the tired web of buttons and dropdowns we’ve been clicking through for years. The real opportunity is interface: tools that don’t just respond to commands but anticipate context. It’s not just about better UX. It’s about a new class of interaction altogether. This could mean a home-buying concierge that doesn’t just show you listings, but understands your daily commute, dog-walking routine, and what kind of light you like in the mornings. A personal finance system that syncs with your partner’s calendar and cash flow, so you plan together without talking about money every day. Or a piano coach that listens as you play and adjusts your practice routine in real time, because you finally have a teacher with infinite patience. Chatbots might have kicked off the era of conversational AI, but the revolution is in what comes after: ambient, embedded, invisible tools that behave more like teammates than software…The best consumer AI product of the next five years won’t wow you with its intelligence. It’ll earn your trust. And neveask for your attention.”

Mint: “The AI assistant isn’t your doctor, but it can make you a smarter patient.”

Nicholas Carr: “Unlike carpentry or calculus, learning is not a skill that can be “mastered.” It’s true that the more research you do, the better you’ll get at doing research, and the more papers you write, the better you’ll get at writing papers, but the pedagogical value of a writing assignment doesn’t lie in the tangible product of the work — the paper that gets handed in at the assignment’s end. It lies in the work itself: the critical reading of source materials, the synthesis of evidence and ideas, the formulation of a thesis and an argument, and the expression of thought in a coherent piece of writing. The paper is a proxy that the instructor uses to evaluate the success of the work the student has done — the work of learning. Once graded and returned to the student, the paper can be thrown away. Generative AI enables students to produce the product without doing the work. Rather than reading and making sense of difficult source texts, they can ask a chatbot to gin up simplified summaries. Rather than synthesizing various ideas and perspectives through concerted thinking, they can ask the chatbot for a generic synthesis. And rather than expressing (and refining) their thoughts through the composition of sentences and paragraphs, they can get the bot to spit out a first draft or even a final one. The paper a student hands in no longer provides evidence of the work of learning its creation entailed. It is a substitute for the work.”

SaaStr: “When only enterprise companies could afford sophisticated sales intelligence, it created a competitive moat. The Fortune 500 got smarter about their sales conversations while SMBs winged it with spreadsheets and gut instinct. That asymmetry just collapsed. Now every startup, every mid-market company, every independent consultant has access to enterprise-grade conversation intelligence. The playing field didn’t just level—it tilted toward whoever adopts fastest.”

WSJ: “Revamped navigation systems will plan your ideal route, make hotel reservations and even teach you about the surrounding geography. Car interiors could transform into a mobile movie theater. And lie-flat massaging seats will activate during endless highway stretches. As self-driving cars become more of a possibility, companies are exploring vehicles that break design convention, sans forward-facing seats, steering wheels or dashboards, allowing for enhanced in-car travel experiences.”

Business Standard: “Indian startups are being held back by a lack of adequate domestic investment due to restrictive regulations of the government, according to industry veteran and Aarin Capital Chairman Mohandas Pai, who called for policy reforms and R&D investments to strengthen the ecosystem. Despite India being the world’s third-largest startup hub, Pai cautioned that the country risks falling behind in global innovation unless these challenges are addressed. “We have 165,000 registered startups, 22,000 are funded. They created $600 billion in value. We got 121 unicorns, maybe 250-300 soonicorns. “The biggest issue for startups is the lack of adequate capital. For example, China invested $835 billion in startups and ventures between 2014 and 2024, US invested $2.32 trillion. We just put in $160 billion, out of which possibly 80 per cent came from overseas. So local capital is not coming in,” Pai said.”

FT: “One reason why “knowledge-rich” curriculums have outperformed “skills-based” programmes is that we are terrible at predicting the future. How could it be otherwise? Someone starting compulsory education in Ohio in 1977 couldn’t possibly have been expected to know that by the time they left school, their state would have experienced significant deindustrialisation, the cold war would have ended and personal computers would have started to become affordable for much of middle America. What “skills” will today’s children need in the world of AI? History teaches us that those we once thought would ensure a reliable income forever are no guarantee of any such thing.”

Harry Nelis (Accel Europe): “One of the things that we’ve learnt over the years, is that venture is a people business. And what makes great companies is the entrepreneurs behind them. I think we continue to have a really steady flow of great entrepreneurs emerging. I don’t think much is holding them back. We have a pretty wide and long runway ahead of us. We have a fully functioning venture ecosystem here. We have universities that are educating business leaders and engineers at world-class levels. Great companies can come from anywhere. That’s the other key thing that we’ve realised. We’ve realised that Supercell came out of Finland, which was not a top five market. UiPath came out of Bucharest, Romania. Celonis came out of Munich. So great companies can come from anywhere. And there’s a steady supply of them.”

WSJ: “AI-powered advertising is part of Meta Chief Mark Zuckerberg’s grand vision for his company’s evolution. Advertising makes up the bulk of Meta’s business—it brought in more than 97% of overall revenue in 2024—and funds its multibillion-dollar investments in AI chips and data centers, and for training cutting-edge AI models. Using the ad tools Meta is developing, a brand could present an image of the product it wants to promote along with a budgetary goal, and AI would create the entire ad, including imagery, video and text. The system would then decide which Instagram and Facebook users to target and offer suggestions on budget, people familiar with the matter said.  Meta also plans to enable advertisers to personalize ads using AI, so that users see different versions of the same ad in real time, based on factors such as geolocation, the people said. A person seeing an advertisement for a car in a snowy place, for example, might see the car driving up a mountain, whereas a person seeing an ad for that same car in an urban area would see it driving on a city street.”

Ruchir Sharma: “The dollar is not the world’s strongest currency and has not been for decades. That title goes to the Swiss franc, and the mighty franc has done nothing to undermine Switzerland’s competitiveness. The world’s richest major economy has both a strong currency and a strong manufacturing base. The Swiss franc has been the top-performing currency over the past 50 years, 25 years, 10 years and five years. It is near the top even over the past year when some of the more beleaguered currencies have staged a comeback against the dollar. Nothing can compare for durable strength.”

Brian Chesky: “Let’s zoom out and ask how we think the future’s going to look. There’s this AI maximalist view that there’s going to be like one or two AI models and one or two applications that rule them all and you use this one app and this one model for everything in the world. If you take that to its logical conclusion, you also start to go to this place where almost one company rules everything, and I think there’s numerous problems with the AI maximalist view that it’s one company to rule them all. One problem with it is, I don’t know if everyone wants one company to have total power and primacy, but the other is just one company is not going to build the entire future. This entire future is going to be built by millions of people in thousands or even millions of companies. There’s an alternative view, which is to say that AI can democratize the world. It’s almost like when technology stagnates the world consolidates, and when there’s this campaign explosion of technology that could actually create a lot more startups. I think that’s another alternative. I do think that every company is going to have to be an AI company or risk disintermediation. The models that are being developed we have access to as well. I think there’s a couple of things that are going to play out here. Number one, I think Airbnb will in and of itself be an AI application. We’re hiring really great people. I think we have one of the best software design teams in the world. We have great application layer design, and I think we can broaden and broaden our app. That’s partly what we’re trying to do.”

SaaStr: “The latest data from Jamin Ball’s Clouded Judgement analysis paints a sobering picture of the SaaS landscape. Aggregate quarterly net new ARR added across the cloud software universe dropped to just $1.65 billion in Q1 2025, down from $2.33 billion in Q1 2024. That’s a 29% year-over-year decline. But here’s what makes this particularly concerning: this isn’t just about one bad quarter. The trend has been building for over a year, with growth rates oscillating wildly between spikes and drops. The market experienced a brief resurgence in Q2 2024 with 50% YoY growth, only to crash back down to negative territory by Q1 2025.”

FT: “‘Vibe coding’ is the new DIY..Large language models enable us all to create our own apps, but sometimes you need a professional.”

NYTimes: “A.I. is just as much a challenge to numeracy — our knowledge and ability to use mathematics and reason quantitatively — as it is to literacy…My advice to young students today is to study language and mathematics. When you talk to a chatbot, you’re using everyday language to talk to a mathematical system that, in turn, talks back to you. Technical skills won’t be enough to deal with the unpredictable results in markets, so a broad-based knowledge of math and language will be the only way to adapt. And while jobs might disappear in one sector, we will always need humans who can make sense of A.I.”

Rahul Jacob: “China’s success has wowed the world even as its factory trends catch on globally. It’s ironic that so much policy attention is being paid to manufacturing while this sector’s job generation drops thanks to robot adoption…The most recent figures from the International Federation of Robotics released in November 2024 show that South Korea leads the way with 1,012 robots per 10,000 employees. China is third with 470 robots for every 10,000 workers. An obsession with manufacturing jobs is thus myopic because automation is a principal leitmotif of new factories. “We are now in a tech race over the software and machines that will power manufacturing, more than the manufacturing itself,” Joe Leahy, China bureau chief for the Financial Times and co-author of a recent article on ‘Made in China,’ told me. Electric vehicle plants, he observes, look like “power stations” because there are so few humans on site.”

IEEE Spectrum: “In a development straight out of science fiction, Australian startup Cortical Labs has released what it calls the world’s first code-deployable biological computer. The CL1, which debuted in March, fuses human brain cells on a silicon chip to process information via sub-millisecond electrical feedback loops. Designed as a tool for neuroscience and biotech research, the CL1 offers a new way to study how brain cells process and react to stimuli. Unlike conventional silicon-based systems, the hybrid platform uses live human neurons capable of adapting, learning, and responding to external inputs in real time. “On one view, [the CL1] could be regarded as the first commercially available biomimetic computer, the ultimate in neuromorphic computing that uses real neurons,” says theoretical neuroscientist Karl Friston of University College London. “However, the real gift of this technology is not to computer science. Rather, it’s an enabling technology that allows scientists to perform experiments on a little synthetic brain.””

Demis Hassabis looks ahead: “If everything goes well, then we should be in an era of radical abundance, a kind of golden era. AGI can solve what I call root-node problems in the world—curing terrible diseases, much healthier and longer lifespans, finding new energy sources. If that all happens, then it should be an era of maximum human flourishing, where we travel to the stars and colonize the galaxy. I think that will begin to happen in 2030.”

1000 EconTalk episodes by Russ Roberts.

McKinsey: “Commerce media networks (CMNs) are the latest shift poised for tremendous growth. CMNs are advertising ecosystems that have expanded beyond retail. For industries as varied as financial services, travel, and even healthcare, CMNs integrate ads directly into the customer experience across multiple online and offline channels. Many CMNs use first-party data from transactions, customer behaviors, and loyalty programs to deliver targeted ads, boosting relevance and sales. CMNs are also experiencing extraordinary growth, with a CAGR from 2023 to 2027 expected to be more than 21 percent, outpacing display, connected television, and even search.”

Dwarkesh Patel: “While this makes me bearish on transformative AI in the next few years, it makes me especially bullish on AI over the next decades. When we do solve continuous learning, we’ll see a huge discontinuity in the value of the models. Even if there isn’t a software only singularity (with models rapidly building smarter and smarter successor systems), we might still see something that looks like a broadly deployed intelligence explosion. AIs will be getting broadly deployed through the economy, doing different jobs and learning while doing them in the way humans can. But unlike humans, these models can amalgamate their learnings across all their copies. So one AI is basically learning how to do every single job in the world. An AI that is capable of online learning might functionally become a superintelligence quite rapidly without any further algorithmic progress.”

FT: “The big AI companies are in effect strip-mining all websites for content to train their models and paying back almost nothing in return to sustain the ecosystem. At present, Google accounts for about 90 per cent of the global search market and directs a torrent of traffic — and advertising — to content sites. Even though they may (understandably) complain about the terms, some content creators fear the trade itself may be in jeopardy because far fewer users would visit their sites. AI models aggregate and summarise the web’s content rather than encouraging users to visit the original content source. That may destroy the financial incentives for creating new content…The obvious survival mechanism for creators is to build walled gardens by erecting more paywalls around their content, or to move to closed channels off the open web.”

Ken Jennings: “Trivia, of all things, is a ray of hope in our moment of national crisis. Somehow, it’s still an arena where ideological projects are completely ignored and the thing that matters — the only thing that matters — is the right answer. On “Jeopardy!” the clues are novel and varied and created every night by gifted human writers, never spun or fact-checked by A.I. The canon from which the questions are drawn is unapologetically evidence-based, the product of scholarly and scientific consensus. And yet the show is, I’m told, one of the last great media monoliths, regular viewing for millions of faithful viewers in red and blue states alike. How do we understand the seeming anachronism of “Jeopardy!”? In a dark time, my secret optimism is that our viewers’ love for quiz games is a sign of what can eventually save us: a practical belief in fact and error that is more fundamentally American than the toxic blend of proud ignorance and smarter-than-thou skepticism that’s brought us to this point.”

2025 Hayek Book Prize: “The Manhattan Institute (MI) is pleased to announce that Johan Norberg has won the 21^st annual Hayek Book Prize for his book The Capitalist Manifesto: Why the Global Free Market Will Save the World.”

Jorge Amar on Agentic AI: “First, there clearly needs to be a rationale from the business: customer support, marketing, sales, HR. They would define, “What is the need for an AI capacity?” and decide, “What are the parameters of what this AI capability needs to perform?” Then they would work with their IT or AI function to either develop or procure their agentic capabilities. In many cases, the specificity and complexity of these AI capabilities will require these companies to develop their agent capabilities in-house, because they cannot find them in the market. It’s going to be a hybrid situation. Once that capability exists, you have to onboard and train that agent, which we call “tuning” an agent. Tuning the agent requires a number of things: a good articulation and understanding of the process you are trying to “agentize,” as well as a subject matter expert who really understands the ins and outs.”

NYTimes: “One of the traits most closely associated with creativity is being open to new experiences. Some people don’t naturally lean that way, but as with all aspects of creativity, openness can be developed. In a sense, you’ve been practicing all week: Maybe you tried intentional daydreaming on your drive to work, or drew a hairy belly button on a blank circle. If those activities felt slightly out of your comfort zone, that’s the point. Experts say the easiest way to become more open is to try something new. The “something new” can be tiny, a minor tweak to your routine. Even taking a different route on your daily walk or ordering red snapper for the first time at your regular restaurant counts.”

Kearny: “India’s retail landscape is evolving at an unprecedented pace, shaped by shifting consumer expectations, digital acceleration, and the growing need for convenience. Among the most transformative developments in this space is quick commerce, a disruptive model that is redefining how consumers shop, how brands engage with their audiences, and how supply chains operate to meet the rising demand for instant fulfillment. As quick commerce gains traction, its impact extends far beyond just fast deliveries. It is reshaping the grocery industry by fueling incremental demand, influencing purchasing behaviors, and unlocking new opportunities for businesses and employment. In this report, we delve into these crucial aspects, providing data-driven insights into its value proposition, pricing dynamics, and broader implications for the industry. We also examine how quick commerce is not merely shifting sales between channels but contributing to market expansion and the economy.”

John Arnold on philanthropy: “I think there’re a lot of similarities to private equity or venture capital. This is mostly a money allocation venture. We have some components of being an operating foundation and a lot of internal expertise on areas, and we do a lot of consulting with policymakers internally, but the majority of the effort is in making grants. The question on making grants is very similar to the questions on making investments. It’s what field do I want to be in? What team do I want to back? What’s their theory of change? Do I believe it? What components of their organization do they need help in? And do we need to supplement or require them to supplement with our resources in order to maximize the chance of success?”

Technology Review: “Last year, China saw a boom in foundation models, the do-everything large language models that underpin the AI revolution. This year, the focus has shifted to AI agents—systems that are less about responding to users’ queries and more about autonomously accomplishing things for them. There are now a host of Chinese startups building these general-purpose digital tools, which can answer emails, browse the internet to plan vacations, and even design an interactive website…These emerging AI agents aren’t large language models themselves. Instead, they’re built on top of them, using a workflow-based structure designed to get things done. A lot of these systems also introduce a different way of interacting with AI. Rather than just chatting back and forth with users, they are optimized for managing and executing multistep tasks—booking flights, managing schedules, conducting research—by using external tools and remembering instructions.”

Mark Robichaux: “Our culture is obsessed with urgency, but winning doesn’t always mean being fastest to first. The deeper reward comes from staying in the race long enough to become who you were meant to be.”

FT: “For the first time in almost a generation, governments are starting to face regular resistance from investors when they try to sell long-term debt…“It’s a classic supply-and-demand mismatch problem, but on a global scale,” says Amanda Stitt, a fixed-income specialist at $1.6tn asset manager T Rowe Price. “The era of cheap, long-term funding is over, and now governments are jostling in a crowded room of sellers.” The reticence among some investors has taken 30-year government borrowing costs in countries such as the UK, Japan and the US to or near their highest in decades and moved the question of debt sustainability up the political agenda. In many countries, the mounting cost of servicing debt interest threatens to squeeze government spending in other areas.”

WSJ reviews “Proof”: “Adam Kucharski, a professor of epidemiology at the London School of Hygiene & Tropical Medicine, takes the reader on a fascinating tour of the history of what has counted as proof. Today, for example, we have computerized proofs by exhaustion, in which machines chew through examples so numerous that they could never be checked by humans. The author sketches the development of ever-more-rarefied mathematics, from calculus to the mind-bending work on different kinds of infinity by the Russian-German sage Georg Cantor, who proved that natural integers (1,2,3 . . . ) are somehow not more numerous than even numbers (2,4,6 . . .), even though the former set includes all the elements of the latter set, in addition to the one that contains all odd numbers. My favorite example is the Banach-Tarski paradox, which proves that you can disassemble a single sphere and reconstitute it into two spheres of identical size. Climbing the ladder of proof, we can enter a wild realm where intuitions break down completely.”

Nathan Levine: “[The war against bureaucracy in the US] is the culmination of a once marginalized, now transformative strand of political thought about who really holds power in the modern American system. Namely, that our democracy has been usurped by a permanent ruling class of wholly unaccountable managers and bureaucrats. Anti-managerialism is back. Well positioned to answer decades of frustration with mainstream conservatives’ failure to deliver results, this old idea has become the central principle of the New Right. In fact, much of what is commonly called populist politics can be more accurately described as part of an anti-managerial revolution attempting to roll back the expansion of overbearing bureaucratic control into more and more areas of life.”

Business Standard: “With over 230 companies operating from Sri City, the zone has attracted investments upwards of $5 billion and has facilitated exports exceeding $4 billion. Among the firms that have a presence here are Alstom, which makes rolling stock for Metro trains, consumer durables companies like Blue Star, Panasonic, Daikin, and Havells, fast-moving consumer goods players such as Mondelez India Foods, Tata Foods, Colgate-Palmolive, and Kellogg’s, besides Sundram Fasteners, Thermax, JSW, and so on.” 

FT: “With stablecoins, the promise is that a dollar is a dollar. They are meant to be backed one for one with reserves of equal value. Holders do not receive interest (but the operators often do, to the tune of billions of dollars a year) or any adjustment for inflation. But they do get to shoot something that smells a bit like real money around the cryptosphere with great ease. For years, this has been an intriguing sideshow on the periphery of global finance. Stablecoin operators have shown varying degrees of willingness to spell out exactly what they own…Back in 2021, warnings were emerging about the risk this poses to normal markets. Rating agency Fitch pointed out that if a stablecoin were to fold for any reason, it could be forced to sell all of its holdings — the dollar assets held in reserve — upsetting the underlying markets.”

WSJ: “A radical new kind of desalination technology is finally on the cusp of helping to slake the world’s thirst. The pitch: Put desalination plants on the ocean floor. First proposed in the early 1960s, this deep-sea process would benefit from both the crushing water pressure and relatively pure seawater more than 1,000 feet down. It has been unworkable until now. Only the recent commercialization of enabling innovations—including deep-sea robots from the oil-and-gas industry, and advanced reverse-osmosis filters now standard in terrestrial desalination—make it viable. Water scarcity is projected to become much more acute in the coming decades, owing to more extreme weather patterns, the decimation of the world’s aquifers, saltwater incursion, and growing urban populations. This threatens humanity at a fundamental level—not just because we need water to drink, but because without it there’s no food or manufacturing, and precious little electricity.”