Category: Thinks

NYTimes: ““The single strongest predictor of economic mobility across areas is the fraction of higher-income friends that low-income people have,” Chetty told me. “In communities where you have more cross-class interaction, kids do much better.””

Andrej Karpathy: “LLM agent capabilities (Claude & Codex especially) have crossed some kind of threshold of coherence around December 2025 and caused a phase shift in software engineering and closely related. The intelligence part suddenly feels quite a bit ahead of all the rest of it – integrations (tools, knowledge), the necessity for new organizational workflows, processes, diffusion more generally. 2026 is going to be a high energy year as the industry metabolizes the new capability.”

FT: “The global economic order is spiralling out of control: tariffs are increasing, nationalism intensifying, co-operation flagging, institutions decaying. If this were not troubling enough, these same forces are feeding off one another, according to economist Eswar Prasad, creating a vicious doom loop that is accelerating the descent into disorder. Nationalist economic policies, for instance, are weakening international financial institutions, making co-operation less attractive and thereby reinforcing nationalism. Prasad explores this perverse logic in The Doom Loop, explaining why the postwar order is disintegrating and offering a sobering portrait of a world on the brink.”

NYTimes: “For decades crushing debt has spread misery in the world’s poor and lower-income nations. But the menace of unsupportable borrowing that now hangs over the global economy emanates from some of the richest countries. Record or near-record debt in the United States, Britain, France, Italy and Japan threatens to hamstring growth and sow financial instability around the globe. At home, it means countries must make interest payments with money that otherwise could have paid for health care, roads, public housing, technological advances or education. The hunger for more and more loans has also pushed up borrowing costs, gobbling up a bigger share of taxpayer money. It can also push up rates on business, consumer and car loans, as well as mortgages and credit cards; and drive up inflation.”

Rama Bijapurkar: “India’s growth is powered by hard-working people and a rising ‘Middle India’. This aspirational group needs financial enablement and infrastructure support to accelerate mobility and sustain growth.”

Richard Fain: “Every organization has its own culture. The first thing to understand is what you have. That means listening to people. It is important to recognize how important culture is. Everyone says, “Yes, our culture is terrific. This is built into our DNA. This is who we are.” That’s nonsense: You develop a culture with intentionality. You say to yourself, “This really matters. I have to work on it,” not, “I have to take it for granted,” or “I have to make sure it’s there.” I’m working on it every day; I’m measuring the progress. I’m deciding what characteristics I want our people to have, whether I want people to operate a business cheaply or I want to operate a business for total excellence. Whatever the goal, identify it and make sure people understand what you’re trying to achieve.”

NYTimes: “New York City has scads of very large buildings, but not many are as big as the glass-and-steel structure nearing completion on the south side of Kennedy International Airport. The airport’s huge new Terminal One will encompass 2.6 million square feet of passenger check-in zones, security checkpoints, baggage-claim areas, restaurants, duty-free shops and boarding gates. That will make it nearly as big as the Empire State Building, bigger than JPMorgan Chase’s new headquarters on Park Avenue and more than triple the size of the new train hub beneath Grand Central Terminal. It’s so massive that it is supplanting three of the eight terminals that once made up Kennedy: the existing Terminal 1 and the demolished Terminals 2 and 3.”

SaaStr: “I don’t want to be sold to anymore. I want to be enabled. I want to click a link, connect my data, and see value in minutes — not days or weeks. And if you’re asking me to leave something that already works, you better be making my life easier, not harder. If your product can’t do that? You’re not losing to a competitor. You’re losing to a founder with Claude and a free afternoon.”

FT: “Bots are learning how to shop. Sales driven by AI platforms will account for about 1.5 per cent of US retail ecommerce this year, according to a forecast by research company Emarketer, but the potential impact of the technology was the dominant topic of conference conversation. “For years, online shopping has been about keywords, filters, drop-down menus. And scrolling through multiple pages [of search results] until you find what you want,” Google chief executive Sundar Pichai, who was joined on stage by Walmart’s incoming boss John Furner, told his audience at the show. “Now . . . AI can do the hard work.” Its promoters say that so-called agentic AI could be a step forward similar in significance to the start of online shopping in the 1990s, or the advent of smartphones in the 2000s, as autonomous agents cut out the tedium of searching and comparing.”

SaaStr: “Forward Deployed Engineers Are the New CS. Every major AI company has figured this out. Palantir essentially invented the model. Now everyone’s copying it. What Forward Deployed Engineers actually do: work directly with customers to understand their specific processes, build end-to-end workflows and take them to production, handle model training and iteration until it works, solve real-world implementation problems daily. They’re engineers + consultants + AI trainers rolled into one.”

Clay Shirky: “As an academic administrator, I’m paid to worry about students’ use of A.I. to do their critical thinking. Universities have whole frameworks and apparatuses for academic integrity. A.I. has been a meteor strike on those frameworks, for obvious reasons. But as educators, we have to do more than ensure that students learn things; we have to help them become new people, too. From that perspective, emotional offloading worries me more than the cognitive kind, because farming out your social intuitions could hurt young people more than opting out of writing their own history papers. Just as overreliance on calculators can weaken our arithmetic abilities and overreliance on GPS can weaken our sense of direction, overreliance on A.I. may weaken our ability to deal with the give and take of ordinary human interaction.”

NYTimes on how to make friends as an adult: “Tune in to intuition. “When you encounter someone you’ve never met before but feel like you’ve known all your life, you need to act on it,” says Sewell. She looks for an “instant feeling of trust, which is rare.” The New York-based photographer and artist Joshua Woods, 39, a friendly face to many on the fashion and culture circuits, is on the lookout for people whose overall tastes align with his own: “It’s how someone lives their life and the things they’re engaged with,” he says, like a shared interest in a certain writer. Aminatou Sow, 40, the interviewer and co-author of the 2020 book “Big Friendship: How We Keep Each Other Close,” who lives in Brooklyn, agrees. When joining a new group — a running club, for example — “pay attention to who those [like-minded] people are and what you want to know about them.” Even in professional contexts, where there are power dynamics to contend with, it helps to “develop a sensitivity for who sees you as a human,” says Sehgal, noting that those people often make for good friend material.”

WSJ: “In “How Great Ideas Happen,” [George Newman] draws on scientific studies, historical examples and behavioral research to argue that what we call inspiration is better understood as a set of habits and mental practices available to anyone willing to cultivate them. Creativity, in his telling, is more method than miracle. The first myth Mr. Newman challenges is the romantic notion that isolation breeds originality. Retreating to a personal Walden, he suggests, may smother creativity rather than fuel it. Isolating ourselves from colleagues, acquaintances and the wider world severs what sociologists call our weak ties, the people outside of our circle of close family and friends who tend to be the conduits of fresh ideas. As important, Mr. Newman argues, is what happens once those ideas surface: submitting them to the scrutiny of others, whose feedback often sharpens what solitary effort cannot. Mr. Newman also dispels the belief that great ideas are entirely new. In practice, he argues, many innovations grow out of existing ones, often by borrowing or transplanting concepts from one field to another.”

FT: “Universal Commerce Protocol [is] a technology standard to help retailers build their own shopping agents and interact with others, this is part of a growing base of technology that could start to replace human attention — the lifeblood of online advertising — with a growing degree of machine-to-machine interaction. UCP joins a list of other protocols designed to automate online activity. This started a little over a year ago with Anthropic’s Model Context Protocol, which enables AI assistants and agents to tap into data held on other companies’ servers, and has since grown to include standards for agents to interact with other agents (A2A) and to make payments on behalf of users (AP2). If internet users find the services made possible by these technologies a more convenient way to get things done, old forms of online engagement are likely to wither. Advertising is still likely to play an important part, even as machine-to-machine interaction becomes more prevalent. At some level, purchases reflect customer preferences, and influencing that preference will always have value. But how and where that influence happens will change.”

Kate Murphy: “You know it when you feel it, with a co-worker, friend or stranger. The science of interpersonal synchrony explains how ‘clicking’ can be a fast track to intimacy—or drama…Synchrony researcher and psychotherapist Dr. Richard Palumbo advises imagining there is a MUTE button during particularly fraught interactions so you focus less on the words used and more on the other person’s level of arousal and how you might be matching that energy. ”It’s your natural human tendency to sync with someone else,” he says. “What’s not so natural is being aware of it.” Sometimes we need to disconnect to recalibrate and reclaim ourselves. The relationships that endure, however, are the ones where you are in sync more than you are not. Grace is learning to ride the tide.”

Yann LeCunn: “There is a sense in which they have not been overhyped, which is that they are extremely useful to a lot of people, particularly if you write text, do research, or write code. LLMs manipulate language really well. But people have had this illusion, or delusion, that it is a matter of time until we can scale them up to having human-level intelligence, and that is simply false. The truly difficult part is understanding the real world. This is the Moravec Paradox (a phenomenon observed by the computer scientist Hans Moravec in 1988): What’s easy for us, like perception and navigation, is hard for computers, and vice versa. LLMs are limited to the discrete world of text. They can’t truly reason or plan, because they lack a model of the world. They can’t predict the consequences of their actions. This is why we don’t have a domestic robot that is as agile as a house cat, or a truly autonomous car. We are going to have AI systems that have humanlike and human-level intelligence, but they’re not going to be built on LLMs.”

Ethan Mollick: “Software developers write Product Requirements Documents. Film directors hand off shot lists. Architects create design intent documents. The Marines use Five Paragraph Orders (situation, mission, execution, administration, command). Consultants scope engagements with detailed deliverable specs. All of these documents work remarkably well as AI prompts for this new world of agentic work (and the AI can handle many pages of instructions at a time). The reason you can use so many formats to instruct AI is that all of these are really the same thing: attempts to get what’s in one person’s head into someone else’s actions.” Adds Arnold Kling: “His point is that using AI effectively requires the management skill of being able to articulate clearly a project’s goals, context, and constraints. He mentions the skill of knowing what an AI can do. I think this could use more emphasis. Sometimes a simple prompt will work, sometimes a more complex prompt is needed, and sometimes a task is beyond the (current) capability of an AI. Knowing the difference is important.”

Menlo VC: “Our data indicates companies spent $37 billion on generative AI in 2025, up from $11.5 billion in 2024, a 3.2x year-over-year increase. The largest share, $19 billion, went to the user-facing products and software that leverage underlying AI models, aka the application layer. This represents more than 6% of the entire software market, all achieved within three years of ChatGPT’s launch.”

FT: “The great thinkers didn’t just answer questions better. They recast the questions themselves. Tocqueville didn’t ask, “Is democracy desirable?” (the dominant European debate). He asked what it does to character, liberty and thought. The question was the insight. Why did we end up assessing answers instead of questions? Because answers are legible, and questions are not…AI represents a third shift. Its authority is neither visible nor structural, but dynamic and ambient, the autocomplete for life. The priest told you what to think. The library told you where to look. AI generates the thinking and lets you believe it’s your own. Unlike the library, a place you entered and left, AI mediates vast stretches of waking life. The work of discernment, once handled by priests and then by institutions, is now ours completely. Kant’s challenge returns: Sapere aude. Have the courage to use your own understanding.”

WSJ: “Manufacturers say that AI, known for creating instant term papers and pixel-perfect fake videos, is fundamentally changing how new products are created. It is letting companies speed-run a process that can often be a deliberative slog relying on tried-and-true approaches. AI tools have helped Procter & Gamble create new scents for body washes, laundry beads and home fragrances. They have allowed Mars to design a thinner-walled bottle for its Extra brand chewing gum that reduced development time by 40% and saved 246 tons of plastic. And they have assisted 3M in coming up with a sanding disc that optimizes dust collection and grinding performance. John Banovetz, 3M’s chief technology officer, said AI is playing the role of an additional colleague. “When I was in the lab, I might talk to three different experts about something,” he said. “AI would just be the fourth expert I’d talk to.””

Akash Gupta: “Anthropic [is] telling you they stopped competing with OpenAI on chatbots at the end of 2024. Jared Kaplan, their Chief Science Officer, admitted it publicly. They’re building vertical AI infrastructure across five high-margin regulated industries where GPT-4 wrappers can’t compete…They’re becoming the middleware layer that every AI application needs to touch regulated data. That ABCDE is a roadmap for vertical integration into five industries worth trillions.”

Blair Effron: “There is little doubt A.I. will be transformative. And yet, for all the disruption it promises, I am struck by how much will remain unchanged. The most consequential decisions in business have never been about processing information faster or detecting patterns more efficiently. The most salient concerns are questions such as what kind of enterprise a firm should aspire to be, what culture it should embrace, what risks it should tolerate and how its leaders can plan when the path forward is unclear. These are questions of judgment, and judgment cannot be automated — at least not any time soon.”

Mint: “That executives can’t yet pin down how AI will make money seems worrying after so many false starts. Yet, that is precisely what past revolutions looked like at their start. As with electricity, the challenge of AI is not in the tools themselves, but how businesses reshape themselves around them.” [Bloomberg]

Debashis Basu: “History suggests that sustaining export growth of around 13 per cent for a decade requires these conditions: Cheap currency, strong central coordination and disciplined policy execution, a large surplus of labour at low wages, assured access to large and open markets, and a willingness to tolerate overcapacity and frequent failures. India currently possesses none of these in sufficient measure. Instead, it faces headwinds from rising protectionism, aggressive dumping by China, and reforms that are often procedural rather than outcome-oriented.” 

India Dispatch: “Amazon Prime Video has more than three times the subscribers as Netflix in India, according to HSBC. Prime Video, which also comes bundled with Amazon’s e-commerce Prime subscription, has roughly 65 million paying subscribers in India, compared to Netflix’s roughly 20 million, the bank wrote in a note to clients. JioHotstar, the market leader formed as a result of the merger between Disney and Mukesh Ambani-controlled Viacom18, leads the Indian market with over 300 million subscribers.”

WSJ: “Generative AI makes voice interactions with devices more productive—and a lot less annoying…Speaking > typing. Today’s voice-transcription AIs have crossed an accuracy threshold: It’s now more convenient to dictate a message than to type it…Talking = the new touch screen. If you’re driving your car and inspiration strikes, you don’t pull out a laptop and start pounding away…Talking to devices makes those moments of inspiration easier to capture.”

NYTimes: “Ricursive aims to build A.I. systems that can improve the design of these enormously complex chips. If A.I. systems can produce better chips, they argue, the chips will produce better A.I. systems. And then the process would repeat on and on as technology got better and better…“The first phase of the company is just to accelerate chip design,” Dr. Goldie said. “But if we have the ability to design chips very quickly, why not just use that ourselves? Why not build our own chips? Why not train our own models? Why not co-evolve them?””

Ruchir Sharma: “Every tech revolution has inspired fears that innovation will destroy jobs. While those fears have never played out, artificial intelligence is cast as much more disruptive because it has the potential to perform so many tasks the way people do — or better. Is the threat to human labour that different and dire this time? What the current obsession with AI overlooks is that another (counter) force is also advancing rapidly. In the past four decades, the number of countries in which the working age population is shrinking has risen from zero to 55, including most of the major economies. This collapse is accelerating now because families are having even fewer children than expected…There are signs AI is already raising output per worker, which could lower overall demand for human labour. But against a backdrop of rapid population decline, the marvels of AI are more likely to ease the coming labour shortages than trigger mass unemployment.”

WSJ: “The AI era will usher in a new style of warfighting “driven by algorithms, with unmanned systems as the main fighting force and swarm operations as the primary mode of combat,” a group of Chinese military theorists wrote in October 2024. They likened AI’s potential to transform the military to gunpowder, a technology invented in China but more effectively weaponized, many in China believe, by others. Drones, for their part, have emerged as key weapons on the battlefields of Ukraine, where strategies and technology for their use have developed quickly under the pressure of real fighting. Drone swarms can be used as decoys that can force an enemy to burn through munitions, as spies and as devastating weapons that can take out enemy soldiers and tanks in suicide missions.”

Anil Dash on Markdown: “Nearly every bit of the high-tech world, from the most cutting-edge AI systems at the biggest companies, to the casual scraps of code cobbled together by college students, is annotated and described by the same, simple plain text format. Whether you’re trying to give complex instructions to ChatGPT, or you want to be able to exchange a grocery list in Apple Notes or copy someone’s homework in Google Docs, that same format will do the trick.”

The Hindu: “Around the world, mentoring has proven to be a powerful tool for supporting young people through key transitions. Mentoring bridges the space between what systems provide and what young people need at a personal level: someone who listens, understands their context, helps them articulate aspirations, and navigates uncertainty alongside them. Mentoring has particular resonance for India because it responds directly to inequalities in access to opportunity. Our work building India’s mentoring movement through Mentor Together for over 15 years shows that high-quality mentoring significantly improves career decision-making, social intelligence, self-efficacy beliefs, and gender attitudes around work.”

NYTimes: “Proponents of prediction markets argue that the platforms are fundamentally different from gambling companies, offering a valuable new source of information by allowing people to bet on world events. Prediction markets are “the most effective way to aggregate information and the crowd wisdom,” said Tarek Mansour, who co-founded Kalshi in 2018. “People don’t lie when money’s involved. You want to be right about your predictions so you don’t lose money.””

A quote in ET: “The [Indian] quick delivery space is still very much a habit-forming market. Price remains the biggest lever to drive trials and repeat usage, especially for groceries. What we are seeing now is deep-pocketed players using discounts as a customer acquisition strategy and that’s causing pain to the larger incumbents.” Mint: “Brands spanning food, wellness, and personal care say they are now allocating up to half of their digital ad spends to quick-commerce apps such as Blinkit, Swiggy Instamart, and Zepto, as sales velocity and return on ad spends improve sharply…Quick-commerce ad spend today has surged almost 40% to nearly $700 million, compared to about $500 million in the preceding six months, Siddharth Jhawar, country manager at ad-tech company Moloco, estimated.”

Ninan: “The underlying issue is that there hasn’t been enough of a structural change in the [Indian] economy since the launch of reforms in 1990-91, despite per capita incomes multiplying nearly five-fold. Industry accounted for a quarter of GDP then, as it does now. The share of agriculture has declined, with crop yields in many cases well below world standards (necessitating a high level of tariff protection, bedevilling trade negotiations). The service sector has become the largest chunk of the economy, but much of it remains in the unorganised part of the economy. Gig employment is not a substitute for proper jobs. A productivity uptick depends critically on three structural changes that are yet to happen: A substantially bigger manufacturing sector, greater formalisation of the economy, and rapid urbanisation. There is as yet not enough evidence of any of the three. If anything, urbanisation may have slowed down. While many things have improved in recent years, much work remains to be done before the economy can gain significant momentum.”

Business Standard: “In a world dominated by laptops, tablets and smartphones, the simple act of writing by hand is quietly making a scientific comeback. Neuroscience research now shows that handwriting activates more areas of the brain than typing, leading to stronger memory retention, deeper understanding and better learning outcomes. Neurologists explain that the physical act of forming letters forces the brain to slow down, actively process information and encode it more deeply than typing on a keyboard allows.”

FT: “Simply put, the maths is against AI start-ups: they need to deliver gains large enough to justify the work and risk of managing a separate tool. And established software companies will win by integrating innovation rather than fragmenting it.”

Manu Joseph: “Smartphones once turned every bystander into a witness, flooding the world with extraordinary videos and dulling our sense of shock. That era’s gone, now that AI fakes have taken apart a treaty we’ve long had with nature: that seeing is believing.”

Morgan Housel: “I have a theory about nostalgia: It happens because the best survival strategy in an uncertain world is to overworry. When you look back, you forget about all the things you worried about that never came true. So life appears better in the past because in hindsight there wasn’t as much to worry about as you were actually worrying about at the time.”

David Deming: “Unlike simple laptop or internet access, AI enables personalized learning at scale. With traditional web search the inputs are personalized, but the outputs are not. You can type anything you want into a Google search bar, and it will give you a ranked list of webpages containing the information you are seeking. The list depends on your exact query, but the webpages you click through to see look the same to you as they do to everyone else. With generative AI, both the inputs and the outputs are fully personalized. Every time you ask ChatGPT a question, you get a unique response that reflects your conversation history and what the chatbot knows about you. The personalization is what makes AI feel like magic. And yet personalization also creates temptation. Generative AI tools are so flexible, you can ask them anything, and they’ll never tell you to stop messing around and get back to work.”

Adam Kelly: If you look at the bigger picture, I think we’re in a golden age for sport. Through 2025, and in the couple of years before that too, we’ve consistently seen audience records broken: the Euros, the Olympics, NFL games and big boxing. That’s a great context for looking across the industry. We’re seeing metrics, across multiple markets, hit new highs. When I think about what that means for the media landscape, I try to step back and ask what the real driver is. Sport has this unique ability to tap into the highest-value part of the attention economy, and also into the experience economy, with the same core product. Nothing else really does that. It drives additional value through scarcity, but the underlying factors are deeper: it taps into community and passion, and it requires an extra commitment from audiences to make a positive decision to engage, to make that appointment to view. That’s why sport is delivering, and why we’re seeing platforms across the industry competing to play a part in it, competing to gather attention they can convert.”

NYTimes: “Do an attention audit. In general, people underestimate how much they use their phones and how often their minds wander, Dr. Smilek said. But spotting these little detours can blunt their impact and make them easier to defend against, he added. Next time you start a task, keep a tally of every time your attention slips, whether your mind is wandering or there’s an external distraction — including what exactly distracted you. Also try mapping out your attentional rhythm — the natural peaks and valleys of focus — by checking in with yourself every hour and reflecting on how well you have been focusing, suggested Gloria Mark, a professor of informatics at the University of California, Irvine.”