Category: Thinks

SaaStr: “The 10 Most Important Learnings From a16z’s Latest State of Markets: Horizontal B2B Trades at 2.7x Revenue, New Startups Are Growing 500%+, and 55% of Unicorns Have Under 2 Years of Runway.”

a16z: “Consumer AI usage is wide, but not very deep…for almost everyone. While nearly half of U.S. consumers now report using AI, only 25% are engaged with it daily. And as of August, just 4.5% of U.S. consumers had an active paid personal subscription to ChatGPT, Gemini, or Claude…However, AI’s consumer power users are extremely engaged — and highly monetizable. The top 10% of spenders account for roughly half of all observed spending. These high spenders disproportionately buy coding, productivity, and creative tools.”

WSJ: “The Year of the Political Outsider…Incumbents around the globe are losing favor with voters riled by unsolvable issues.”

James Marriott: “The most stark piece of data is the American Time Use survey of almost a quarter of a million Americans, which found that over the last two decades leisure reading fell by 40 percent.”

Neelkanth Mishra: “The drivers of growth are largely in India’s control. A large part of the demand that we are catering to is domestic. And the external demand, the biggest variable, is services, where the growth momentum is still quite strong. Having said that, the risks are coming from higher energy prices and the struggle to attract more capital. So, even financing 1 per cent of the gross domestic product worth of the current account deficit is becoming challenging because global capital markets are disrupted. What we need to do, therefore, is keep the energy flowing, continue with reforms so that the inflow of capital can be sustained from new sources, and make ourselves more attractive in what is getting to be a tougher beauty parade. If we can sustain our growth rates, I think the inflow of risk capital is also likely to persist. At the same time, in this period of significant turbulence, we have to prioritise macroeconomic (macro) stability. A steady growth rate compounded over many years is far better than trying to maximise growth in any given year.”

NYTimes: “A quick Economics 101 refresh: Growth — and the increases in incomes and living standards that come with it — is a function of just two things, a larger labor force and how much each worker produces. More workers increase the overall size of the pie. But individuals don’t get a larger share unless each worker produces more. Greater national income, particularly if it comes from higher productivity, gives the nation more legroom to redistribute wealth fairly among citizens and reduce inequality. It also provides resources to hold down the national debt and fund public services such as health care and education. On their own, workers generally can’t become more productive. Investment helps, such as by replacing old manufacturing equipment. Training can benefit, too. But the key driver of efficiency is technological innovation. Think robots in factories, self-checkout in a drugstore or automated lawn mowers.”

Brian Potter explains Vision-Language-Action Models. “While a large language model takes text as an input and gives text as an output, a VLA takes text, images, and robot information as an input and gives a series of robot actions as an output. It does this using many of the same components that made LLMs so successful, specifically attention and the transformer architecture. It’s not amazingly obvious if VLA’s will remain the primary robotic AI paradigm, but as of now they’re a very common method for controlling robots.”

WSJ on “On”: “The company that began in Switzerland with a pair of scissors and some garden hose is now conquering America. If you don’t already own a pair or three, you have definitely seen them. The shoes are known for their clean design, lightweight feel and unmistakable soles, including some with those hollow pods that resemble a toothy smile. They help soften each landing, which means the very thing that makes Ons recognizable is also what makes them so comfortable.”

WSJ: “A silent electronic battle is raging at the heart of the Russia-Ukraine war. Both sides are attacking and defending communication signals on an unprecedented scale, innovating at pace, while the U.S. and other Western countries are working to catch up. The Pentagon is investing heavily in new capabilities. “This is a new era of spectrum warfare,” said Lt. Col. Matt Sherburne, the U.S. Army’s chief for electronic warfare capabilities development. Electronic warfare isn’t new, but it fell out of focus in the West after the end of the Cold War. Then came Russia’s war on Ukraine, where swarms of drones have become the dominant weapon on the front lines. Both sides are now engaged in a constant game of cat and mouse, jumping around the frequency spectrum to jam and avoid each other, and they are increasingly using AI to do that.”

FT: “The telephone makes superficial contact easy. That can be the beginning of a more meaningful human relationship, or it can be a replacement for it. Large language models offer the same temptation when it comes to gathering knowledge. A superficial survey becomes effortless. Do we let it catalyse deeper research and understanding? Or replace it? The choice is ours.”

FT: “China is building data centres at rapid speed across its energy-rich hinterland, as Beijing seeks to turn abundant electricity and cheap land into an advantage in the global race to develop AI. The push is creating vast new computing hubs such as Ulanqab in Inner Mongolia, where 89 data centres are built or planned. The AI infrastructure build-out is among the fastest and biggest anywhere in the world. Beijing has outlined AI progress as a national strategic priority, as a means to bolster economic productivity, entrench domestic political influence and secure China’s global influence by ensuring more of the world becomes reliant on its technology.”

Devina Mehra: “Test whether your story about a stock, sector or market makes sense, or are you discarding facts that don’t fit in. Even if the story currently makes sense, will it do so forever? Challenge the assumptions and think of alternative endings. Do this and you may be able to avoid the ‘storification bias’ a bit.”

Mack Paul on his book Research Triangle Park: “MP: The book details the interrelationship among state and local government, universities, and business to produce an ecosystem conducive to research and development (R&D). The RTP strategy prioritized luring companies to North Carolina to undertake R&D, believing that would spur new industries and create jobs. In this sense, the RTP strategy mirrored previous efforts to bring textile and furniture companies to North Carolina. Of course, textiles and furniture eventually declined. I am curious whether R&D faces a similar fate. R&D is not limited to a particular industry, although it’s certainly essential to the tech industry. We’ve seen ups and downs in activity around R&D over the years. Will it eventually decline, or is there something about R&D that insulates it from the same forces that impact industries based on production. The book discusses the distinction between hard tech and soft tech—manufacturing versus creative content. My sense is that the answer is bound up in that distinction, but we did not flesh it out in this book.”

NYTimes: “In his new book, “Sleeping Beauties: Why Good Ideas Go Dormant and How They Wake Up,” [David] Byrne draws on his reading to deliver profound insights about creativity. Throughout our history of information scarcity, the currency of progress was discovery. Now, in the age of information abundance, inspiration and innovation may depend more on rediscovery. Byrne guides us on a captivating quest to understand how we come to remember forgotten ideas, inventions and technologies. With striking dexterity and range, he explores how advocates reawakened Bach’s compositions and Vermeer’s paintings after one and two centuries of slumber; how technological advances rekindled theories of continental drift and black holes; how accidents resuscitated hand-washing and genetics; and how war renewed interest in the submarine and “The Great Gatsby.” He traces the term “sleeping beauties” to a 2004 study of scientific papers that were dormant before being revived.”

FT: “Many tech leaders believe software that can take actions on your behalf, removing the drudgery of routine work and personal tasks, represents a huge opportunity. Exactly what form this will take, however, is less clear. The rewards for finding the right formula are likely to be huge: personal agents could become a new entry point to the digital world and turn their makers into the next gatekeepers. So far these include Muse from Meta and dots from OpenAI, while Microsoft has revamped its Copilot assistant to act more autonomously, Apple added agentic features to its Siri assistant in its latest iOSG and Google launched a new work assistant in Gemini. Several start-ups are also angling for attention, most notably Instinct, which launched an invitation-only test over the summer.” WSJ: “The difference between a regular chatbot and an agent is that users typically ask agents to carry out tasks, like organizing their calendar or booking an appointment. The agent then breaks this prompt down into a series of steps, which are for convenience often not spelled out to the human it’s working for. Agents typically know to ask permission before doing major things like making purchases or deleting files, but other than that, they can have free rein. If a user’s prompt is unclear, or if there are many ways to execute the requested action, mayhem might ensue.”

Cato: “Over the past 25 years, Madrid has quietly carried out a free-market transformation. While much of Europe moved toward expanding regulation and growing state control, Spain’s capital region chose a different path—lowering taxes, cutting red tape, opening markets, and partnering with the private sector to improve public services. Today, Madrid’s economy, roughly the size of Portugal’s, stands as one of Europe’s most dynamic regional success stories. The results are measurable: faster growth than the national average, stronger job creation, rising incomes, and increased foreign investment. But Madrid’s story is not only about economic policy. Nearly half of its residents were born in another Spanish region or abroad. Rather than relying on identity-based politics, Madrid has fostered integration through opportunity and equal treatment under the law.”

WSJ: “As more companies add AI to their daily workflows, they will also need strategies to manage their use, such as training employees on which model to use for a particular task. Otherwise, they could end up with IT bills that come out of a black box.”

Ingar Haaland: “With math essentially being delegated to OpenAI, here’s what I predict for economics and the social sciences more generally: The top tier of research will just become better and it will be normal human-led research where AI is used for scale (e.g. conducting qualitative interviews with relevant populations, running behavioral interventions in the field, analyzing massive text data, etc.) Field experiments will rise in value and generally making “connections” with firms and being able to run stuff (potentially testing AI pipelines) will be in high demand. Research with administrative data will become even more valuable, but the benefits might be concentrated among some prolific authors who are allowed by the government agencies to use local models to analyze the data at scale. PhD students can probably forget about it. Economic history will have a very exciting boom and there will be an enormous race and big rewards for digitizing archives as a source to both identify new research strategies and document new descriptive facts.” [via Tyler Cowen]

Hugo Bowne-Anderson and Eleanor Berger: “The defining combination is autonomous and asynchronous execution. Autonomous: the agent chooses and carries out steps towards an agreed goal. It can inspect what happened, adjust its approach and continue within its authority. Asynchronous: that work can proceed without your ongoing attention. You can leave and return to its result. A scheduled job can run without you. A coding agent can choose and revise an approach during an interactive session. A factory combines those capabilities, with enough context, tools and feedback to carry out the delegated work. Software factories have been hypothesized for decades. Now that we have agents, we’re actually able to operate them. The principles apply to software work traditionally carried out by teams of humans, from bespoke tools for teams and individuals to large and complex software systems.”

Kelsey Piper: “The labs intend to automate the AI development process with AI, which will mean faster and faster progress and necessarily less and less human oversight. If all goes according to schedule, humans will play a rapidly shrinking role in AI development over the next two years, as both OpenAI and Anthropic try to put as much R&D as possible into the hands of their increasingly capable (and increasingly incomprehensible) AI.” [via Arnold Kling]

Arvind Narayanan: “Intelligence isn’t Power.”

Mint: “From first sale to ₹100 crore: The new playbooks building consumer brands in India”

Bloomberg: “A group of former employees from some of the fastest-growing artificial intelligence firms, including Ramp and Cognition AI Inc., are starting a new company to create AI agents that can help run a business, joining a crowded field of services using the technology to streamline work…“What we are doing is building AI that owns outcomes,” said Moritz Stephan, chief executive officer of Hone and a former chief of staff at AI coding startup Cognition. Hone’s technology is meant to “look at what a team does, its goals and responsibilities, and deploy engines to own these,” while checking in with a human as needed. A growing number of companies are selling AI software to businesses that promise to automate complex work, including in coding and customer service. Hone’s ambition is to use AI to have a broader impact on productivity and business operations.”

ET: “Menlo Ventures, which first backed Anthropic in 2023 when the company was pre-product and pre-revenue, said a slowdown in frontier model development will not significantly affect foundation model companies or the overall market. Partner Matt Murphy, who led Menlo’s investments in Anthropic, called the slowdown “a bit of a red herring,” saying compute, not intelligence, is the real constraint.”

Michael Magoon: “The average American taxpayer moves from roughly the 20th percentile of the income distribution at ages 18–25 to the 75th percentile at ages 55–64, simply as they move through the normal life cycle.” [via Arnold Kling]

Andrew Chen: “We are seeing a new wave of general purpose consumer-friendly agents, and it’s causing many thousands of AI startups to re-evaluate their place in the market. Some questions: Will these new horizontal players take over completely? Will it be winner take all? Can products that focus vertical use cases survive? The new wave of products – Muse, Town, Instinct, Grokbot, etc – come from taking the power and magic of Openclaw/Hermes but wrapping it in polished, secure, consumer-friendly experience. This is only the beginning, and soon every major tech company will throw their hat in the ring and thousands of startups that have been working on vertical agents in X (travel, shopping, family, work, SMB, etc etc) have big existential questions to answer. The simple case against winner-take-all dynamics: These agents are tools, not networks. They inherently provide a solo experience, and although packaged as chat experiences, they don’t have network effects the way that Whatsapp or phone networks do.”

NYTimes: “Microdramas, sometimes called “verticals,” are TV shows for your phone. They are typically described as soap operas for Gen Z: hypersexual, over the top and — most important — short, with episodes rarely exceeding two minutes in length. The format was born around 2018 in China, where its revenue now exceeds that of the conventional film industry. America is lagging behind, but it is putting in the work…The form is on track to take in $1.5 billion in revenue in the United States by the end of the year.”

BCG newsletter on China’s next five years: “Beijing knows where it wants to go. But the journey won’t be simple. The government aims to lift consumption through wage growth and expanded social benefits in health care, childcare, and eldercare, and by focusing on health, wellness, and leisure services for an aging population. It is also moving against industrial overcapacity, though provincial leaders—judged on local GDP growth and job creation—sometimes continue to add factories. And it’s making choices about which high-value industries to double down on, while opening further to foreign investment and deepening trade and investment ties across the Global South.”

Ninan: “Two scenarios are now possible. If the AI investment bubble bursts, there will be a global tsunami, since the US stock market accounts for more than two-thirds of global market capitalisation. If not, and the borrowing binge by governments and companies continues, there will be unprecedented competition for capital, leading to possibly higher inflation and/or restrictions on cross-country capital flows. It makes sense, therefore, to shore up reserves, increase economic security by reducing the budget deficit and government debt, and make the country more attractive to investors. The additional reserves can be used to fight currency volatility, and to fund any capital outflows. They can also be used to boost confidence in the economy’s external solvency at a time when no one knows how exactly the cards will fall.”

WSJ: “[Grip strength is] a measure of the force produced by your forearm and hand muscles. But doctors say it’s a good proxy for overall strength in both your upper and lower body…The median grip strength for a 50-year-old man is about 48 kilograms or 100 pounds, while a woman’s is about 29 kilograms or 64 pounds, studies show.”

Bain: “From 2025 to 2035, AI puts $4.7 trillion in profits at stake—created in new categories, shifted among competitors, or lost entirely by those who move too slowly. That’s more than triple the economic impact of the last great technology disruption—the Internet. And it will happen in half the time.”

NYTimes: “But against these long odds, The Times is still standing. Indeed, not just standing but thriving — heralded as a case study in how to transform a legacy company for a new era. In a period when most news organizations watched their audience and revenue collapse, our company has grown from about one million paying subscribers to more than 13 million. Revenue and profit are also growing steadily, enabling an unprecedented expansion of our newsroom. The stock price — now higher than it was at our pre-internet peak — suggests that investors believe that this old newspaper’s best years are still to come. And, yes, we own our headquarters again. Which raises the question: How on earth did The New York Times, a seemingly ossified institution in a dying industry, survive the most destabilizing period in its history — and emerge stronger?”

Alok Mittal: “Micro, small and medium enterprises (MSMEs) form the backbone of Indian economy, contributing over 30% of gross domestic product (GDP), 45% of exports and employing more than 110 million people. In an increasingly deglobalizing world, they represent India’s core capacity to produce and serve its large, diverse population. Over the past two decades, startups have emerged as a key engine of innovation. Be it breakthroughs in enabling marketplaces, driving efficiency or responding to sovereign technology requirements, startups are increasingly seen as an integral driver of India’s march towards a developed economy.”

FT: “The truth is that most white-collar professional roles do not just involve “sitting down at a computer”, from lawyers attending court hearings to project managers going on site visits. Then there are the meetings (internal and external). A recent study of Norwegian workers found that meetings consumed an average of 12 per cent of work hours and 14 per cent of company wage bills. There will be those who say that most meetings are just performative busywork, but if that is the case, it is strange that the most successful companies seem to do more of them. The authors of the Norwegian study found that planning, problem solving, information sharing and project co-ordination accounted for most meeting activity, that high-paying and high-revenue firms devoted more resources to meetings, and that meeting frequency and intensity were positively related to worker wage growth. They concluded that meetings were like broccoli: “widely disliked, but probably good for us anyway”.”

Ben Thompson: “This is the new prize in technology, and it is the ultimate one: not just a platform like Windows, or an Aggregator like Facebook. What Muse and Copilot are making a bid to be is both: the only interface you need for everything, using a computer on your behalf, and generating whatever UI you need when you want it, and doing stuff without you but for you the rest of the time.”

NYTimes: “Being in constant communication can be nice, but it can also create a lot of pressure — on ourselves and others — and a lot of opportunities to be offended when lapses occur. Research reveals that the frequency of interaction does not determine the quality of a relationship. As one viral social media post put it, “true friendship is allowing a certain level of ghosting to go unmentioned.” In fact, time apart can actually bring us closer together.”

FT: “ASML, arguably the most consequential yet low-profile company in the AI ecosystem, employs nearly half the town (population: 47,902) and another 25,000 besides, doing the no-nonsense work of making and marketing the ever-larger machines that churn out ever-smaller computer chips…The secret of this 42-year-old company’s success, Fouquet tells me in a nondescript meeting room at ASML’s headquarters, has been teamwork. “If I had to summarise it, this has been a lot about extreme co-operation with the ecosystem, which is, first, our customers, but also our supply chain,” he explains, portraying ASML as the epitome of the collaborative European company, in stark contrast to a thrusting American one.”

David Deming: “I believe that in this new third era of undergraduate education, we should lean into our strengths and become radically more communal, enabling other forms of learning in which technology is of little use. Many of the most important skills and habits of mind are best learned in community with other people. The ability to explain a concept so that other people can understand it, to inhabit the mind of another person so you can understand their motivations and work alongside them, to persuade others of your point of view, to negotiate over resources and navigate conflict, to lead and be led, to give and receive criticism and to earn another person’s trust— these skills are central to a college education.” [via Arnold Kling]

WSJ: “Most entrepreneurs dream big. [Alan] Dorfman does the opposite. He spends every waking hour trying to shrink things to their smallest possible size without losing their functionality. He has churned out miniature versions of thousands of products including the Rubik’s Cube, the Operation board game and Hungry Hungry Hippos. In a world of endless entertainment options, somehow there’s growing demand for impossibly tiny toys. Dorfman says Americans’ enduring appetite for novelty has gotten stronger recently as people seek out the low-tech thrills of earlier eras.”

Joe Cruz: “What if we are mistaken in treating writing as the highest cultivation of thought? The evidence that supports this belief is thin at best…I believe colleges can permit the use of A.I. writing tools without undermining the next generation of thinkers. But to get there, we need to embrace a different image of a university education — one in which ideas are expressed in real time and analytical thought unfolds as a debate or an oral defense or a group discussion. It will take courage to demand that students demonstrate active, participatory mastery of ideas. And it will take courage on their part to abandon the safe recursion of revision and narrative refinement. To be sure, there will be challenges of scale, and not every kind of institution will be able to adapt. The benefits, however, may be enormous.”

Mike Cannon-Brookes: “If a frontier model is capable of coming and running the entire business for you, if it’s truly running the whole thing, even assuming four or five years worth of increase, you have a relatively simple business. Most businesses aren’t simple. They are global conglomerations of rules, compliance, laws, staff in different places, human inconsistency — which is the same as saying human creativity — that they’re putting it all together to try to deliver products and services for their customers in their industry: healthcare, university, finance, automotive, or whatever the industry is. They’re trying to compete with some other business. They are still going to have a huge number of humans, I would argue more knowledge workers, more developers, in five to 10 years time than today because the ability to do tasks to compete will go up.”