Philippa Kaur

Listen to this story. Enjoy more audio and podcasts on More

Listen to this story. Enjoy more audio and podcasts on More

ARMING UNCLE SAM is a great business. America’s latest defence budget earmarks $170bn for procurement and $145bn for research and development (R&D), most of which ends up with the handful of “prime” contractors, which deal directly with the Department of Defence (DoD). So will some of the $44bn in American military aid to Ukraine and some of the extra defence spending by America’s European allies, which account for 5-10% of the primes’ sales. Although those sums do not increase at the same rate as, say, corporate IT spending, leaving less room for spectacular gains, arms manufacturers are also shielded from eye-watering losses by huge, decades-long contracts.Thanks to a big shake-up at the end of the cold war, the industry is also highly concentrated. At a meeting in 1993, dubbed the “last supper”, William Perry, then President Bill Clinton’s deputy defence secretary, told industry bosses that excess capacity was no longer appropriate and that consolidation was in order. As a result, the ranks of the primes have thinned from more than 50 in 1950s America to six. The number of suppliers of satellites has declined from eight to four, of fixed-wing aircraft from eight to three and of tactical missiles from 13 to three.image: The EconomistGuaranteed custom and weak competition have helped American armsmakers’ shares comfortably outperform the broader stockmarket over the past 50 years. A paper published by the DoD in April found that between 2000 and 2019 defence contractors did better than civilian ones in terms of shareholder returns, return on assets and return on equity, among other financial measures. An increasingly unstable world means more money going to the armed forces—and to their suppliers. Total shareholder returns, including dividends, at primes such as General Dynamics, Lockheed Martin and Northrop Grumman rose when Russia invaded Ukraine in February 2022 and when Hamas attacked Israel on October 7th (see chart 1).This cosy oligopoly is now being challenged on two fronts. One is technological. As the tank battles on Ukrainian plains and in Gazan streets show, “metal on the ground” remains important. So do missiles, artillery shells and fighter jets. But both conflicts also illustrate that modern combat relies increasingly on smaller and simpler tactical kit, as well as communications, sensors, software and data. The second challenge is the Pentagon’s efforts to extract greater value for money from the military-industrial complex.Both developments undercut the primes’ big competitive advantages: their ability to build bulky kit and to navigate the mind-boggling procurement process. Cost-effective innovations, such as the Pentagon’s recently announced project “Replicator”, which aims to get swarms of small drones in the air, ASAP, require agile engineering for which the defence giants “are not innately organised”, as Kearney, a consultancy, delicately puts it. If they are to thrive in the new era, they will have to rediscover some of the innovative ways that helped them shape Silicon Valley in the decades after the second world war. So far, though, they are finding this difficult.It is easy to see why the primes (and their investors) like the current setup. The DoD reimburses the primes’ R&D expenses, and adds 10-15% on top of that. This “cost plus” approach spares the companies from funnelling lots of their own capital into risky projects, which offers security but reduces the incentive to deliver things on time and on budget. The project to build the F-35 fighter jet, which has accounted for more than a quarter of Lockheed’s revenues in the past three years, started life in the 1990s. It is running around a decade late and will cost American taxpayers up to $2trn over the lifetime of the aircraft.Once in production, newly developed large kit is sold at a fixed price, often for decades. The B-21 stealth bomber currently in development by Northrop Grumman will cost the Pentagon more than $200bn for 100 planes delivered over 30 years. The Columbia Class nuclear-submarine programme made by a subsidiary of General Dynamics will sail from the early 2030s until at least 2085.Past their primeThe Pentagon’s patience with this time-honoured business model is wearing thin. Last year’s national-defence strategy summed it up succinctly: “too slow and too focused on acquiring systems not designed to address the most critical challenges we now face.” Instead, it wants to “reward rapid experimentation, acquisition and fielding”. This is forcing the primes to think about how they could build fresh functionality atop their existing platforms, by adding new software, modules, payloads and the like, and to create production processes which can be modified to accommodate innovations.As Lockheed Martin’s chief executive, Jim Taiclet, recently acknowledged, soldiers expect seamless integration of sensors, weapons and systems for battle management such as joint all-domain command and control (JADC2), a new concept for sharing data between platforms, services and theatres. Contracting, building and continuously updating such systems will be a struggle for firms that have hitherto produced huge bits of hardware slowly and which, in the words of Steve Grundman of the Atlantic Council, a think-tank, are not “digital natives”.The primes face another problem. The technology the Pentagon has in mind is not inherently military, observes Mikhail Grinberg of Renaissance Strategic Advisors, a consultancy. Most of the defence giants do have civilian divisions—large ones in the case of Boeing, General Dynamics and Raytheon. But the Pentagon’s growing appetite for dual-use technologies means more competition from civilian industry, which is constantly devising new equipment, materials, manufacturing processes and software that could be used for military as well as peaceful ends.In 2020 General Motors won a contract to supply infantry vehicles. The carmaker has now teamed up with the American arm of Rheinmetall, a German weapons firm, in a deal to furnish military trucks. Other challengers are trying to muscle their way into the military-industrial complex, drawn by the DoD’s appetite for more diverse systems. Palantir, founded in 2003 to help avert more attacks like that of September 11th 2001, makes civilian and military software that processes the vast amounts of data that modern life and warfare throws up. Elon Musk’s SpaceX sends payloads, including military ones, into orbit and is being paid by the DoD to provide internet access to Ukrainian forces in their fight against Russian invaders.image: The EconomistBig tech is getting in on the action, too. Amazon, Google and Microsoft have targeted defence and security as promising markets, says Mr Grundman. Military procurement is a rare business large enough to make a difference to the tech titans’ top lines, which are counted in the hundreds of billions of dollars. The trio, along with Oracle, a smaller maker of business software, are already sharing a $9bn cloud-computing contract with the Pentagon. Microsoft also supplies the army with augmented-reality goggles in a deal that could eventually be worth $22bn.The Pentagon’s new approach is also attracting upstart rivals. Anduril, a startup founded in 2017 solely to serve military needs, has developed Lattice, a general-purpose software platform that can be swiftly updated and adapted to solve new problems. The company also makes a short-range drone called the Ghost, which can be operated by a couple of soldiers. Recognising that to win business quickly it needs to be vertically integrated, it has acquired a manufacturer of rocket engines and is developing an underwater autonomous vessel for the Australian navy.The wannabe primes and their financial backers still bemoan the barriers to new entrants. Brian Schimpf, Anduril’s boss, says that when working with the DoD you get “punched in the face every day”. SpaceX and Palantir both had to fight court battles just to be able to contest military contracts. In June Palantir signed an open letter with 11 other companies, including Anduril, and investors, imploring the Pentagon to remove obstacles to smaller contractors. The letter, which drew on proposals from the Atlantic Council, condemned “antiquated methods” that “drastically limited” access to commercial innovation.As the national-security strategy shows, the DoD seems keen to move away from procurement antiquity, for example by shifting more risk onto contractors through fixed-price, rather than cost-plus, development contracts. Such developments are causing palpitations among the primes. Boeing’s recent financial travails are partly a result of catastrophically underbidding in fixed-price contracts for the KC-46 tanker and Air Force One, which ferries around American presidents.In contrast, Anduril has dispensed with the crutch of cost-plus of its own accord, and is investing its own capital to make what it thinks the DoD will need. By clinging on to the old model, the primes may be depriving America of the 21st-century defence industry it needs. ■The Economist‘s A to Z of military terms explains modern warfare in plain English More

Hiring processes can be thought of as a battle between candour and dishonesty. You might imagine this is a simple fight between truth-seeking firms and self-promoting candidates, and to a certain extent it is. But companies themselves are prone to bend reality out of shape in ways that are self-defeating.Start with the obvious culprits: job applicants. The point of a CV or a LinkedIn profile is to massage reality into the most appealing shape possible. Everyone beyond a certain level of experience is a transformational leader personally responsible for generating millions in revenue; the world economy would be about 15 times bigger than it actually is if all such claims were true. The average Briton spends four and a half hours a day watching TV and online video. But the average job candidate uses their spare time only for worthy purposes, like volunteering in soup kitchens or teaching orphans to code.The cover letter is so open in its insincerity (“When I saw the advertisement for this job, I almost fainted with excitement”) that people are starting not to bother with it. At the interview stage one task facing the firm’s recruiters is to winkle out the truth of what a person actually contributed to a project. Those hoary questions about a candidate’s weaknesses and failures are there for a reason; no one will bring them up unprompted. Cognitive and behavioural tests are useful in part because they are harder for applicants to game.But a tendency to stretch the truth infects companies as well as applicants. The typical firm will write a job description that invariably describes the work environment as fast-paced and innovative, and then lays out a set of improbable requirements for the “ideal candidate”, someone who almost by definition does not exist. Sometimes—as when ads demand more years of experience in a programming language than that program has existed for—these requirements include an ability to go back and alter the course of history.Industrialised hiring processes can often reward mindless exaggeration. Services that scan your résumé when you are making an application mark you down if your CV does not match the keywords that appear in the original job advertisement. The message is clear: to get through to the next stage, you have to contort yourself to meet corporate expectations.Substance can matter less to recruiters than form. One software engineer claims to have got a 90%-plus response rate with a spoof CV which showed apparent spells at Microsoft and Instagram but also boasted, among other things, that she had increased team-bonding by organising the company potato-sack race and “spread Herpes STD to 60% of intern team”. References are so prone to inaccuracy that many firms have a policy of not giving them, fearing legal action from defamed candidates or deceived employers.Too few firms offer an accurate account of what a position actually involves. Tracey Franklin, the chief HR officer for Moderna, a fast-growing drugmaker—and an interviewee in this week’s episode of Boss Class, our new podcast—is a fan of “realistic job previews” (RJPs). These are meant to give prospective recruits a genuine sense of the negatives and positives of the job, as well as a clear idea of the company’s corporate culture. One effective tactic is to lay out, in text or video, what a typical day in the role would look like.Such honesty can be its own reward. Longstanding research suggests that RJPs lead to lower turnover and higher employee satisfaction. A paper in 2011 by David Earnest of Towson University and his co-authors concluded that favourable perceptions of the organisation’s honesty are the best explanation for why.The incentives on both sides of the hiring process lean naturally towards glossing reality. If candidates were to give genuinely truthful answers (“I have a habit of making basic but calamitous errors”), many would rule themselves out of jobs. And if firms were to give a warts-and-all description of themselves, many would end up deterring good applicants. But a process designed to uncover the truth about job applicants would run a lot more smoothly if firms were also honest about themselves. ■Read more from Bartleby, our columnist on management and work:Would you rather be a manager or a leader? (Oct 26th)How big is the role of luck in career success? (Oct 19th)Trialling the two-day workweek (Oct 12th)Also: How the Bartleby column got its name More

Good news about America’s economy seems to keep rolling in. In the third quarter, gdp expanded by a barnstorming 4.9% in annualised terms. Heading into earnings season, the month or so each quarter when most firms report their latest results, a stream of upbeat economic figures led stockmarket analysts to hold their profit expectations for the quarter steady, rather than trim them like normal. Many called the end of America’s corporate-earnings recession. Such optimism now looks justified. Following a hat-trick of consecutive year-on-year quarterly profit declines, America Inc’s bottom line is growing again (see chart 1). According to FactSet, a data provider, of the half of big firms in the S&P 500 index that have reported their results, 78% have beaten profit expectations (see chart 2).image: The EconomistYet the mood during the quarterly carnival of conference calls has hardly been celebratory. Plenty of bosses failed to excite investors despite bringing them sound results. The reaction to the performance of big tech was particularly discordant. Alphabet, the parent company of Google, heartily beat profit expectations but saw its share price sink by 10% after investors were underwhelmed by how its cloud-computing division was doing. Meta’s warning on macroeconomic uncertainty meant that the social-media empire’s biggest-ever quarterly revenue figure went unrewarded by markets. The lingering possibility of a recession and anaemic levels of corporate dealmaking overshadowed banks’ profits from lending at higher rates of interest.image: The EconomistWhy the gloom? A boom in the third quarter notwithstanding, the future health of America’s consumer remains bosses’ biggest worry. It is easy to see why. American businesses draw more than a third of their revenues directly from domestic consumers’ pockets, according to Morgan Stanley, a bank. Shoppers have seemed indefatigable; retail sales grew by 0.7% in September, compared with August. Coca-Cola and PepsiCo both raised profit guidance for the rest of the year. But recently their growth has been the result of price rises rather than selling more fizzy drinks and snacks.Other cracks are appearing. According to Bank of America, credit- and debit-card data show a downturn in spending in October, compared with a year ago. Earlier this month Americans with student loans had to resume debt payments after a three-year reprieve. In aggregate, spending is now growing faster than real disposable income, eating into savings. Consumers say they are gloomier about their financial situation—and who can blame them? At the same time, credit-card and car-loan delinquencies have been ticking up (see chart 3).image: The EconomistThat is worrying chief executives. UPS, a delivery firm, said consumers were spending less on goods and more on services, dampening its outlook for profits. Mattel, a toymaker which owns the Barbie brand, among other things, delivered a blockbuster quarter but its outlook for Christmas flopped. Bosses at Alphabet say the tech titan’s data showed customers hunting harder for deals and offers of free shipping for goods. On Tesla’s investor call Elon Musk bemoaned the effect of rising interest rates on consumers’ ability to afford the company’s cars. (Though, as Mr Musk also admitted, some of Tesla’s problems were manufactured: “We dug our own grave with the Cybertruck.”) Since the call, Tesla’s share price has fallen by 15%, wiping more than $100bn off its market value.Companies are also closely watching their costs, especially for labour. Margins were boosted by cooling wage inflation across the economy. Strikes, however, remained a headache in some parts of the economy. By the end of September Hollywood writers had agreed to up pens but many automobile workers’ tools remained resolutely down. On October 25th the United Auto Workers (UAW) union struck a tentative deal with Ford, a Detroit giant, to end industrial action and increase workers’ wages.But General Motors, another carmaker, said that the strike by members of UAW would now cost it $200m per week and withdrew its profit guidance for the year. Detroit’s big carmakers were not the only ones feeling the pressure: Illinois Tool Works, which makes car parts, cut its profit guidance. Even bosses at Delta Air Lines complained that fewer passengers were landing in Motor City.Happenings farther afield were also weighing on bosses’ minds. A common refrain in many earnings calls was sadness at the loss of life in Israel and Gaza. Yet for now at least, conflict in the Middle East is not having large financial effects. A few firms signalled caution—Snap, a social-media firm, said some advertisers in the region paused spending as a result of the war in Gaza. But corporate America as a whole earns only a vanishingly small part of its profits in the Middle East. American bosses who examine the direct risks posed to their business by the war in Gaza are likely to conclude that they are much smaller than the costs of, say, unwinding operations in Russia, let alone the existential worries about America’s relationship with China.By comparison, bosses were silent on a bigger long-term threat to earnings: higher interest rates. During the past year the fortunes of big business have diverged from those of smaller firms, especially ones owned by private-equity funds, as they have been largely immune to the soaring cost of capital. Bank of America reckons that more than three-quarters of debt borrowed by S&P 500 firms is both long-term and fixed-rate, compared with less than half in 2007 when ten-year Treasury bond yields last exceeded 5%. Eventually, however, big businesses’ debt piles will need to be refinanced at a higher rate of interest, which will squeeze profits. The earnings recession might have ended in the third quarter. But plenty of threats still lie ahead. ■ More

IN SOME ways, big business pow-wows are all alike. Talking heads make over-the-top predictions. The world’s problems are packaged into bite-sized quotes. Chief executives vie to use as many words as they can to say as little as possible. So too at the seventh Future Investment Initiative in Saudi Arabia’s capital, Riyadh, on October 24th-26th.Dancers in space-age suits rocked the stage and a young operatic star wowed the audience at the opening session. Large futuristic screens flashed buzzwords du jour—AI, data, sustainability—in an arena fit for e-sports. Corporate and financial bigwigs engaged in a pretend boardroom dialogue at a roundtable in the centre of the conference’s main hall to discuss the state of the world. Drones hovered overhead.But Davos in the desert, as the event is better known, is singular. Unlike at others, including its European namesake, billions of dollars in deals get signed on the sidelines. It is also even harder to get in. This year, for the first time, the organisers charged a fee of as much as $15,000 per person—steep as talkfests go. That, as one financier put it, filtered for “more high-quality people”.The mood among this select crowd was unusually tentative. Techno-optimism about artificial intelligence and medical advances was tempered by an acceptance that the energy transition is perhaps not just around the corner. The metamorphosis of Saudi Arabia from a joyless outpost of orthodox Islam into a bustling business hub—”so dramatic”, gushed Jamie Dimon, boss of JPMorgan Chase, America’s biggest bank—stood in contrast to the war raging in Gaza between Israel and the militant Islamists of Hamas. Business leaders worried how quickly the conflict was escalating. They pondered Saudi Arabia’s role in navigating the tensions. It is the regional giant and the hostilities threaten its ambitious economic blueprint, which seemed like it might involve normalising relations with Israel.Beyond the regional turmoil, inflationary pressures, especially on wages, and huge fiscal deficits in many countries weighed on participants’ minds. Ray Dalio, founder of Bridgewater Associates, the world’s biggest hedge fund, declared that he was pessimistic about the global economy in 2024. The chief executive of Goldman Sachs, David Solomon, pointed to deep uncertainty that has left business bosses feeling on edge.Another Wall Street stalwart, Jane Fraser of Citigroup, summed it up. “It’s hard not to be a little pessimistic,” she said, noting that global risks were increasingly interconnected and that security was becoming one of the biggest concerns. As Ms Fraser put it, businesses will increasingly need “big ears and thick skin”. Not just to prosper, she might have added, but to survive.■To stay on top of the biggest stories in business and technology, sign up to the Bottom Line, our weekly subscriber-only newsletter. More

For foreign investors, India is a puzzle. On the plus side, it is a potentially huge market, recently passing China as the world’s most populous. The IMF predicts that India will be the fastest-growing of the world’s 20 biggest economies this year. By 2028 its GDP is expected to be the third-largest, moving past Japan and Germany. The stockmarket is pricing in heady growth. Over the past five years Indian stocks have beaten those elsewhere in the world, including America’s.The minuses can seem equally formidable. Just 8% of Indian households own a car. Last year the number of individual investors in Indian public markets was a paltry 35m. The smartphone revolution unleashed 850m netizens, but most scroll free apps like WhatsApp (500m users) and YouTube (460m). Blume, a venture-capital (VC) firm, estimates that only 45m Indians are responsible for over half of all online spending. Netflix, the video-streaming giant, which entered India in 2016 and charges Indians less than almost anyone else, has attracted just 6m subscribers.The tension between tomorrow’s promise and today’s reality is reflected in India’s tech scene. Over the past decade giddy projections of spending by hundreds of millions of consumers led investors to pour money into young tech firms. According to Bain, a consultancy, between 2013 and 2021 total annual VC funding ballooned from $3bn to $38.5bn. Now the easy money is running out. In 2022 startups received $25.7bn. In the first half of this year they got a measly $5.5bn.Some of India’s brightest tech stars have fallen to earth. The valuation of Byju’s, an ed-tech darling, has plummeted from $22bn to $5.1bn in less than a year. Oyo, an online hotel aggregator, has delayed its public listing even as investors slashed its value by three-quarters, to $2.7bn. Moneycontrol, an online publication, estimates that since 2022 Indian startups have shed more than 30,000 jobs. Investors now worry that companies in their portfolio will never make money. Heavy losses by Indian “unicorns” (unlisted companies worth $1bn or more) bear this out. According to Tracxn, a data firm, of the 83 that have filed financial results for 2022, 63 are in the red, collectively losing over $8bn. Yet some Indian tech firms manage to prosper. Rather than promise mythical future riches, they are practical and boring, but profitable. Call them camels. Zerodha, a 13-year-old discount brokerage, clocked $830m in revenue and $350m in net profits in 2022. In 2021, the latest year for which data are available, Zoho, a Chennai-based business-software firm founded in the dotcom boom of the late 1990s, made a net $450m on sales of $840m. Info Edge, a collection of online businesses that span hiring, marrying and property-buying, has been largely profitable throughout its 20-year existence. Their success is built on an idea that seems exotic to a generation of Indian founders pampered by indulgent investors: focus on paying customers while keeping a lid on costs.Consider revenue first. Some founders privately grumble that getting the Indian user to pay for anything is hard. But Nithin Kamath, founder of Zerodha, disagrees. He believes that though the wallet size of Indian consumers is small, they are willing to pay for products that offer value. Zerodha charges 200 rupees (around $2.50) to open a new account when most of its competitors do so for nothing. Mr Kamath believes that even this small amount forces the company to ensure that its users find its platform useful enough to pay that extra fee.India’s technology dromedaries are also ruthlessly capital-efficient. Zerodha and Zoho have not raised any money from investors. Info Edge was self-funded for five years before raising a small amount, its only outside financing before going public in 2006. Sanjeev Bikhchandani, who founded Info Edge, advises founders to treat each funding round “as if it is your last”.One way to extend the runway (as VC types call the time before a firm needs fresh funds) is by keeping costs down. Take employee salaries. Richly funded startups throw money at pedigreed developers from top-ranked universities. Zoho enlists graduates of little-known colleges and rigorously trains recruits before bringing them into the fold. The company says that its approach results in a wider talent pool and more loyal employees.Zerodha, meanwhile, in another contrast to profligate unicorns, does not spend any money on advertising, discounts and other freebies to lure customers. It also uses free open-source alternatives to paid software for its technology infrastructure. The company’s tech-support system for its more than 1,000 employees costs just a few hundred dollars a month to run; an external tool would set it back a few million. Despite being a technology-heavy trading platform, it spends just 2% of revenues on software. Keeping overheads low has the added bonus of allowing companies like it to sell their products profitably at bargain prices, reaching many more customers in the price-sensitive subcontinent.Reboot, not copy-paste
The slow, measured approach taken by the camels is the opposite of the Silicon Valley playbook of capturing market share first and worrying about profits later. Karthik Reddy of Blume argues that such a model may be better suited for India, where businesses can take many years to find their feet.One hurdle for companies choosing steady profits over blitzscaling growth remains: the investors themselves. Venture capitalists typically operate on a ten-year clock, bankrolling startups in the first five and cashing out their stakes in the second. This gives investors an incentive to push portfolio firms to pursue growth at all cost. Sridhar Vembu, Zoho’s boss, likens venture capital to steroids—it can boost short-term performance but damage the business in the long run. His may be an extreme view. Still, if investors want big returns on their Indian bets, they are better off backing sturdy camels over sexy unicorns. ■Read more from Schumpeter, our columnist on global business:Are America’s CEOs overpaid? (Oct 17th)Weight-loss drugs are no match for the might of big food (Oct 12th)So long iPhone. Generative AI needs a new device (Oct 5th)Also: If you want to write directly to Schumpeter, email him at [email protected]. And here is an explanation of how the Schumpeter column got its name. More

IT HAS BEEN quite a year for Mark Zuckerberg. The co-founder of Facebook, a social-media Goliath now called Meta, is no stranger to public rebuke. But exactly a year ago even investors appeared to throw in the towel, accusing him of trashing the core business while lavishing money on his pharaonic dreams for the metaverse, a virtual world where he alone appeared to float in a deluded fantasy realm. On the day Meta issued weak third-quarter earnings last year, its share price fell by more than a fifth. Zuck’s name was mud.In the year since it has been rehabilitated. Meta’s core business—engaging 3.1bn people a day on Facebook, Instagram and WhatsApp, and selling advertisers access to their attention—is back to rude health. On October 25th the company reported revenues of $34.1bn in the third quarter, up by 23% year on year. That was the sharpest rise since the digital boom of the covid-19 pandemic. Net profits more than doubled to $11.6bn. Meta’s share price has risen by 250% since last year’s nadir. image: The EconomistIn the media, Mr Zuckerberg gets little credit for his business nous. There is more focus on other stuff: his recent passion for martial arts; the cage fight with Elon Musk that never happened; public haranguings, such as lawsuits filed by dozens of American states on October 24th, alleging that Meta intentionally sought to make users addicted to Facebook and Instagram. And yet, in the space of a few months late last year, he made two transformative business decisions that were remarkable for their humility and agility—all the more so, given that he controls 58% of the firm’s overall voting rights and barely needs to work, let alone listen to shareholders.In response to investor pressure, Mr Zuckerberg performed one of the fastest pivots in tech history. Within a fortnight of the third-quarter rout he slashed Meta’s spending plans, cut costs and fired staff. And in response to OpenAI’s ChatGPT and the blaze of excitement around generative artificial intelligence (gen AI for short) he launched an internal revolution aimed at using the technology to galvanise Meta’s core business. Those manoeuvres reveal a lot about Mr Zuckerberg’s leadership style. They may even end up vindicating his faith in the metaverse.When Mr Zuckerberg realised he had incensed investors, those around him say, he did not panic. He became methodical. As Nick Clegg, a close adviser to Mr Zuckerberg, explains, his boss doesn’t like people around him “shouting and yelling”. He prefers, like an engineer, to break down a problem to its component parts and decide on a course of action. In this case, he understood that his long-term focus was at odds with investors’ short-term horizons. So he decided to “cut his cloth accordingly”. But he kept many of his long-term investment plans intact, emphasising that they mainly concerned AI, not the metaverse. That emphasis looked shrewd weeks later, when ChatGPT burst onto the scene.Meta had spent years building up its AI infrastructure. Rather than creating chatbots, it was looking for ways to use AI to improve engagement and make its ad business more efficient, as well as working on mixed-reality headsets for the metaverse. Its top brass soon realised they had all the ingredients—enough data centres, graphics processing units (GPUs) and boffins—to make the most of gen AI. By February they had worked out what to focus on. By July they had made their Llama 2 large language model available free of charge to developers. In September they announced the first gen-AI-related gadgets, such as smart spectacles. Mr Zuckerberg, for his part, threw himself into the technical nitty-gritty. His competitive instinct awakened. He appears to have been rejuvenated by working on a new technology rather than on the irksome task of cost-cutting.Making Llama open-source helped turn Mr Zuckerberg from Silicon Valley’s villain to its hero. Leigh Marie Braswell of Kleiner Perkins, a venture-capital firm, says startups “really applauded” the move, which helped many develop AI-related businesses. And gen AI may be no less transformative for Meta itself than for Microsoft and Alphabet, owner of Google, whose early bets on proprietary large language models have attracted most of the attention.Start with engagement. Meta is populating its social-media platforms with chatbot avatars which, it hopes, will increase the amount of time people spend on their feeds, and help businesses interact with customers on messaging apps. Some users call them a bit humdrum, probably because the firm is worried about AI’s “hallucinations”. Nonetheless, there is potential. Take Jane Austen, an avatar that emulates the author’s haughty humour. When asked to describe Mr Zuckerberg, she says he is “bright, driven but perhaps a bit too fond of his own ideas”. She describes the metaverse as a “virtual world where people can escape reality and live their best lives. Dear me, how…unromantic.” More compelling in the near term is AI’s potential for advertising. Since Apple restricted Meta’s ability to track user data across third-party apps on iPhones, Mr Zuckerberg’s firm has had to overhaul its advertising business “down to the studs”, says Eric Seufert, an independent analyst. It has done that fairly effectively, he thinks, by using AI to model user behaviour, rather than tracking the behaviour itself. Last year the company rolled out ad technology called Advantage+, which used AI to automate the creation of ad campaigns. Brent Thill of Jefferies, an investment bank, says that advertisers are impressed. J. Crew Factory, a clothing retailer, has told Meta that the features boosted its return on ad spending almost seven-fold.Gen AI could take automation further. This month Meta launched tools that let advertisers instantly doodle with different backgrounds and wording. These are baby steps so far, but Andy Wu of Harvard Business School likens them to the start of a gold rush. He says that by creating gen-AI-infused ad campaigns Meta could benefit from the technology as much as Nvidia, the leading maker of GPUs.Advertisers have their concerns. An ad man at AdWeek NYC, an industry jamboree, described Meta’s AI-assisted campaigns as “black boxes” where it controls all the data. That gives it huge influence over a brand’s identity, which could be tarnished if the AI goes rogue. Others worry about AIs doing untoward things to boost engagements on Meta’s social networks, which could hurt brands by association. Controversies over fake images of the conflict in Gaza on social media illustrate how fraught the terrain remains. Not everyone is convinced by Mr Clegg’s insistence that Meta is prepared for this thanks to years of investment in safety and platform integrity.Some investors, too, remain sceptical. Mark Mahaney of Evercore ISI, another investment bank, reckons that 95% of them would prefer Mr Zuckerberg to spend less on the metaverse. Many are wary of investments in hardware, such as virtual-reality headsets, which tend to generate lower margins than digital products.Still, Mr Zuckerberg has “not resiled at all” from his long-term bet, Mr Clegg says. Some VR enthusiasts see AI as the metaverse’s saviour, helping with the development of crucial hand-tracking technologies and making it cheaper for creators to build three-dimensional worlds. Meta’s Smart spectacles, integrated with its chatbot, MetaAI, and built by Ray-Ban, offer a hint of things to come. They capture what the wearer sees, can live-stream it on social media, and answer questions. Asked for sources on critical thinking in business, the AI replied “The Economist”. Smart, smarmy or scary? Take your pick. ■To stay on top of the biggest stories in business and technology, sign up to the Bottom Line, our weekly subscriber-only newsletter. More