He sold a dream nobody buys anymore

November 22, 2024
by Björn Jeffery
SvD Näringsliv

This analysis was first published in SvD Näringsliv, in Swedish, on November 22nd, 2024. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.

The suit sold the dream. The excavator was supposed to do the work. But the visions and the reality never pulled in the same direction at Northvolt. What Peter Carlsson leaves behind is not what he had imagined.

It is a bright spring day in April 2018. A large yellow excavator stands at Finnslätten in Västerås. The first sod for Northvolt’s demonstration factory is a scoop of gravel. Standing next to the machine is a tall man wearing clothes more usually associated with a different kind of work: a white shirt and a dark suit. Peter Carlsson, Northvolt’s CEO, is pleased to have arrived here after eighteen months of preparation. In a year’s time, the plan is for Northvolt to begin producing its first batteries. “It’s an enormous opportunity for Sweden,” he says. It would not quite work out as he imagined — not for him, not for Northvolt, and not for Sweden.

The project actually begins much earlier, on the other side of the Atlantic. In 2013 Peter Carlsson was head of the supply chain at Tesla. He could see how the need for batteries would grow dramatically as the automotive industry electrified. The geopolitical tensions between East and West existed even then, though they would intensify considerably over the coming decade. The insight about battery demand was not unique. What was unusual was the approach Carlsson absorbed from his then-boss, Elon Musk. Tesla did not have factories — it had “gigafactories.” Today Tesla is not a car company but an AI and robotics company, if you were to ask Musk. He is a master at projecting vast visions over the more mundane reality of what is actually happening in the business day to day.

Back in Sweden, Peter Carlsson executed a manoeuvre his former boss would have been proud of. He painted the picture of a Sweden that would get its first home-grown gigafactory producing batteries. But doing so required financing in the multi-billion bracket. Carlsson’s vision engaged prospective customers like Volkswagen as well as funds and venture capitalists. Everyone was invited along for the ride. It became Europe’s most richly funded startup.

This is roughly where reality started to chafe. Building battery factories is complex. Building several simultaneously — in different countries, as Northvolt did — is harder still. The factories are hit by serious delays and complications. Meanwhile the company needs continued financing to keep operating. It becomes a juggling act between sorting out something as concrete as machinery on a factory floor, and convincing financiers that this is a journey they need to be part of — delays notwithstanding. At the same time, the global environment deteriorates. Inflation takes hold and interest rates surge. Optimism about the future takes a knock. But the money must come in and the factories must start working. The equation eventually becomes too difficult. On Thursday Northvolt announces it is filing for reorganisation in the United States — a so-called Chapter 11 process. On Friday, Peter Carlsson steps down as CEO.

It is impossible not to think back to the excavator and the pile of gravel from 2018. The suit sold the dream. The excavator was supposed to do the work. But the two never quite pulled together. Northvolt is a project that arouses strong feelings. On one side are critics who see it as a failure and a sign of a “green bubble” — an industrial project that was misconceived from the start and should never have been built. On the other side are those who see the value in the vision. Sweden and Europe need entrepreneurship and major ventures of this kind to manage the green transition and build greater independence from China. Risk is part of the price you pay for achieving something never done before.

The tension between vision and reality has characterised Northvolt from the very beginning — and in this case two things can be true simultaneously. What Peter Carlsson has demonstrated is that it is possible to mobilise industry, society and business to create something Sweden has never seen before. That is an achievement of real significance. At the same time, reality has consistently lagged behind, with years of delays and billions in additional costs as a consequence. The execution has not worked. What remains after the reorganisation will be a different kind of company, with different owners and new leadership. For Peter Carlsson the story ends here, as he steps down as CEO and becomes a senior adviser and board member after many years of hard work. For Northvolt, the next chapter remains to be written — but that will be someone else’s task.


Nvidia’s success is a risk factor for all of us

November 21, 2024
by Björn Jeffery
SvD Näringsliv

This analysis was first published in SvD Näringsliv, in Swedish, on November 21st, 2024. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.

Nvidia doubled its profit and the market responded with an “okay.” The expectations placed on the world’s most valuable company have become a problem — for the entire stock market.

There are two types of companies on the stock market. The first requires a trained equity analyst to understand the business — someone meticulous and detail-oriented who reads every quarterly report looking for the smallest signal about where the stock is headed. The second type is companies like Nvidia. This week it was reported that Elon Musk’s AI company xAI is in the process of raising 66 billion kronor from investors, with the purpose of buying 100,000 new chips from Nvidia. You do not need to be a painstaking detective to pick up the signal in that news.

Despite this, markets were jittery ahead of Nvidia’s latest results. Nvidia — now the world’s most valuable listed company — has somewhat involuntarily become a barometer for something much larger than itself. The prevailing AI boom has benefited both tech companies and ordinary listed companies through optimism about increased revenues and future efficiencies. In recent days Meta launched a new division dedicated solely to selling AI solutions to businesses. At the foundation of all these AI ventures sit Nvidia’s products. When AI gains ground in the world, so does Nvidia right now.

When the chipmaker’s quarterly figures were presented Thursday evening Swedish time, this was more than evident. Revenue grew 94 percent and profit more than doubled. Which other company of this size can report that kind of growth? Or has ever done so? But Nvidia is not like others. When the company’s guidance for Q4 indicated growth of 70 percent, that figure had to be measured against the preceding year’s 265 percent growth rate. The expectations are enormous. The stock thus dipped slightly in after-hours trading — despite the almost extraordinary gains.

Nvidia appears to be priced for a perfect delivery in a perfect market. That is precisely what it has experienced over the past two years. But can it continue? One concern is whether AI will gain ground in the way many seem to hope. The tech giants’ investments in infrastructure have been gigantic so far, but there is anxiety that they may be costing more than they are worth. At the scale Nvidia now operates, it would take only a slight reduction in that long-term commitment — investments pushed a little further into the future — to hit Nvidia directly.

As the largest company on the market, it therefore attracts a disproportionate number of eyes. The whole market watches it to see which way the wind is blowing. The situation is unusual. Under normal circumstances most people would probably have ignored a chipmaker like Nvidia. It listed in 1999 and a safe guess is that the name was unknown to most readers until relatively recently. Now, suddenly, it is bigger than Apple, Microsoft and Google. Looking at the top 100 companies by market cap, Nvidia is as large as the bottom 25 on the list combined — a group that includes names like Goldman Sachs, Inditex and BlackRock. That is the scale we are talking about.

That size means the outside world now knows more about chip deliveries and production issues than it ever wanted to. If Nvidia’s new chip, Blackwell, were delayed or failed to perform as intended, it would be a problem for far more than just Nvidia. The company now appears in all major global index funds, in US equity funds, in technology funds. If you have money in the markets or a pension savings account, you are probably exposed to Nvidia too — likely more than you realise. The company that spent so long making graphics cards for gaming PCs has become a barometer for the stock market itself. That was never its intention. For now the tailwind for Nvidia and the AI revolution it leads is still strong. But should that weaken, it could quickly become a problem for far more people than just them.


The mission: cut 20,000 billion kronor

November 13, 2024
by Björn Jeffery
SvD Näringsliv





Musk’s mission: cut 20,000 billion kronor

Published in Svenska Dagbladet, 2024-11-13. Translated from Swedish.

It is now confirmed — super-entrepreneur Elon Musk is stepping into Washington to clean up among American government agencies. The mission: cut 20,000 billion kronor from the federal budget.

Cast your mind back twelve years and try to explain this to someone if you can. A mildly confused Shiba Inu dog becomes popular on the internet. A meme emerges — an internet joke — about the dog misspelling its own name, and it gets called “doge,” a mangled version of the English word “dog.” Doge then becomes a cryptocurrency called Dogecoin, which at the time of writing has a market value of around 600 billion kronor — based on pure speculation. That is a higher valuation than companies like Volvo or Ericsson.

One of the most successful entrepreneurs of our time, Elon Musk, takes a liking to doge and Dogecoin. He uses the letters as an acronym for a proposed government body — the “Department Of Government Efficiency.” And now — in the early hours of Wednesday morning Swedish time — incoming president Donald Trump has appointed Musk to launch exactly that. What started as a joke is now becoming serious.

At a campaign event before the election, Howard Lutnick — responsible for managing the transition to Trump’s presidency — was on stage with Musk, fired up about how much government spending they could cut: “How much do you think we can rip out of this wasteful $6.5 trillion budget from Harris and Biden?” Musk responded with a broad smile: “I think we can do at least $2 trillion!” That translates to cuts of 20,000 billion kronor per year. At the time it was a hypothetical campaign promise from a billionaire hoping Trump would win. Now it may become reality. Together with former presidential candidate Vivek Ramaswamy, they will shape DOGE — the Department of Government Efficiency — to cut costs from the American state apparatus. The irony of appointing two chiefs and launching a new “department” to find unnecessary spending appears to have escaped them. “This will send shockwaves through the system,” Musk said in a statement.

Taking an axe to organisations is something Musk knows well. After buying the social media service Twitter in 2022 he laid off around 80 percent of the staff within a year. Outside observers predicted the service’s immediate death, not believing it could be run with so few employees. Apparently it could. X — as Twitter was renamed — rolls on today with around 6,000 fewer people on the payroll than when he bought it.

The American federal government is not a company, however. As former Treasury Secretary Larry Summers pointed out in a Fox News interview, only 15 percent of its costs relate to personnel. Firing every single government employee would therefore not be enough to reach 20,000 billion kronor in savings. To get there, he added, you would almost certainly need to cut healthcare programmes Medicare and Medicaid — a long-standing Republican ambition.

As a businessman, Musk has not only been good at cutting costs. He has also secured new revenue streams for his businesses. SpaceX carries out missions on behalf of the federal space agency NASA. Tesla has earned many billions from a type of mandated electric vehicle credits that competing car companies have been forced to pay them. Its car-buying customers have benefited from federal tax breaks introduced to encourage the electrification of the automotive industry. There may be savings potential in those areas too — but a fairly safe guess is that Musk will start looking at the other end of the government machinery first.

The situation is simultaneously somewhat bizarre and extraordinarily unprecedented. That money plays a large role in American politics is hardly news — but Musk has built himself an immense position of power in a very short time. Axios co-founder Jim VandeHei has described him as the most powerful private citizen in American history. He runs several billion-dollar companies in strategically important sectors, owns X as a media platform — and now steps across the threshold from the private to the public sphere. A government mandate to cut costs.

Those who have followed Elon Musk for a long time know that he tends to exaggerate and be optimistic about timelines. At the same time, he has delivered on multiple projects that most considered impossible — and did so simultaneously. Musk is not like everyone else. Now that he has a mandate to shake up government operations, one thing is certain: it will be messy. But it will happen. What the consequences will be remains to be seen.


Trump’s victory gives Musk the full payoff

November 6, 2024
by Björn Jeffery
SvD Näringsliv





Elon Musk bet everything on Trump — and is now more powerful than ever

Published in Svenska Dagbladet, 2024-11-06. Translated from Swedish.

The world’s richest man played an extraordinarily high-stakes political game. Elon Musk bet everything on red — and won. Now comes the payoff.

They were not always close friends, the two most talked-about billionaires in America. In the summer of 2022, Donald Trump’s re-election campaign was already under way, and from a podium he described Elon Musk as a “bullshit artist.” But a lot can happen in two years. Musk has rapidly become Trump’s most powerful ally. Now the payoff is coming.

Elon Musk is an entrepreneur whose many businesses are deeply intertwined with the American state. Tesla earned over 18 billion kronor in 2023 from a state-regulated electric vehicle tax credit. SpaceX has in practice functioned as a privatised arm of NASA. Add the regulatory approvals needed by brain-implant company Neuralink, and the political reach of the platform X. The picture is complex, but the common thread is that all are profit-driven companies that have made Musk enormously wealthy while operating in the grey zone between the public and private sectors — sometimes through subsidies, sometimes through regulation, sometimes through its absence.

Having a political leader who is well-disposed toward you and your businesses is therefore extremely valuable for Musk. But he has clearly managed to succeed without Trump’s help until now. To understand Musk’s newly found and very vocal political engagement, we need to look forward instead. What is on the agenda for his business empire?

Start with Tesla, which accounts for the single largest part of Musk’s wealth. In the late 2000s Tesla were pioneers in the electric vehicle category — a small niche at the time. After questions about whether that kind of car production could ever be economically viable, Musk launched several new models that were well received by the market. The small player became dominant and effectively forced the rest of the car industry onto the same track. Today electric vehicles are no exotic category — everyone makes them. Tesla remains a large and important player, but its dominance has faded. More importantly, the niche is no longer unique. To justify a higher stock market valuation than other car manufacturers, the company must talk about something else. Why else would Tesla carry a P/E ratio of 63 while General Motors trades at around 5? Tesla’s answer is self-driving cars and AI. Musk himself has said the company should be seen as an AI and robotics company. But after selling a “self-driving” add-on to Tesla cars for many years, full autopilot has yet to be released — partly because it is not ready, and partly because Tesla lacks regulatory approval. Self-driving cars do exist on American roads, but only from a handful of companies in a couple of specific areas. Tesla is entirely dependent on how regulations are shaped. Having the president’s ear in that situation is very convenient.

Next is SpaceX. Even with breakthroughs like catching and potentially reusing rocket boosters, enormous resources are required to keep the operation running. Exploring Mars — one of Musk’s stated goals — has no commercial basis at present. Musk wants to run the programme, but the American state needs to foot the bill.

Finally, X. The social network has lost enormous numbers of advertisers since Musk took over, but its power position remains strong. During the election campaign it was mobilised to present an alternative worldview to what was available in mainstream media. Accurate? Sometimes. But its role as a hub for a Republican media machine was cemented. Interest in X will increase with Trump as president.

It is possible that the former Democrat Elon Musk has simply grown tired of the establishment. He has expressed a desire to cut unnecessary bureaucracy through a special assignment for Trump — DOGE, the Department of Government Efficiency. The idea that the state is inefficient is hardly a unique insight; it is a well-worn argument from the political right. Being at the centre of such a process would be a new kind of feather in Musk’s cap. Whatever his actual political views, Musk had many billions of reasons to support a Republican election victory. He played a high-stakes game and managed to enrage the Democrats enormously. A defeat would have been devastating for him and for the companies he is involved in. The billionaire won again. And his power has never been greater.


The market has already picked a winner

October 29, 2024
by Björn Jeffery
SvD Näringsliv





Prediction markets ahead of the US election can influence more than money

Published in Svenska Dagbladet, 2024-10-29. Translated from Swedish.

Want to bet on who will win the US presidential election? With the right prediction you can now make a lot of money. But there are also reasons beyond profit that make speculators want to project a clear winner well ahead of the result.

“People overestimate their knowledge and underestimate the probability that they are wrong.” Nassim Nicholas Taleb, the economist and author whose glass is usually half empty, reviews humanity’s view of itself in his book Fooled by Randomness. Probabilities and truths are tricky. Sometimes they resemble each other, but they are not identical. Now — in the intensifying run-up to the American presidential election — that reminder may be timely. While Americans are placing bets on who will become the next president, the process risks sowing the seeds of distrust about how the election actually played out.

So-called prediction markets have existed for a long time — allowing individuals to speculate on specific questions like whether a certain technology will win a Nobel Prize or whether Taylor Swift will be the most-streamed artist on Spotify this year. It has been a hobby for a small minority trying to foresee the near future. But interest has rarely been higher than now. The reason is a newly legalised and rapidly growing market for wagering on the presidential election result specifically.

In early October the American service Kalshi won its appeal against the CFTC — the Commodity Futures Trading Commission, which regulates these markets. The CFTC had argued that this kind of prediction-making could negatively influence the election, but failed to prove that case sufficiently in court. As a result, Kalshi became the first service permitted to open public, money-backed predictions on who will be the next US president. The brokerage app Robinhood has since also allowed users to bet on the election outcome.

“Prediction” is a generous term in this context. Formally that is what users are supposed to be doing — but since there is a financial interest in the outcome, it is perhaps easier to call it what it actually is: a bet. Even a football pools coupon is a form of prediction, at some level. The difference from ordinary sports betting odds is that the market is entirely open. Kalshi sets no odds of its own; the system functions more like a stock exchange driven by supply and demand. The more people who have bet on a given presidential winner, the lower the payout — because the market has determined that outcome to be more likely. But probability, as noted, is not identical to truth.

The gap between prediction markets and opinion polls is in fact very large. At the time of writing, Donald Trump leads with around 62 percent on Kalshi’s market, while the latest polls show a near-dead heat. At least one of them is very wrong. Which one? And more interestingly, why? Outside the United States, additional prediction markets allow wagering on the election. Polymarket, a crypto-based service, can sidestep American regulations by allowing everyone except Americans to speculate on the result. Foreign users appear to have drawn similar conclusions — Trump leads with around 67 percent on Polymarket.

Some speculators appear extraordinarily confident. A user going by “Fredi9999” recently placed around 150 million kronor on Trump becoming the next president. The underlying crypto technology makes it difficult to know who — or how many people — are behind the position. One can only speculate about the motive. Perhaps it is not primarily about trying to profit from a correct guess, but about trying to influence the outcome. If Trump appears to be the winner in voters’ eyes, his chances of actually winning increase. Rajiv Sethi, economics professor at Barnard College, told the Wall Street Journal: “If I were trying to manipulate a market, this is exactly how I would do it.”

The divergence between prediction markets and polls could also become a card to play in any post-election dispute. Given the conflict that followed the 2020 election between Biden and Trump — which culminated in the storming of the Capitol on 6 January — there is already fertile ground among some for the belief that the election will not go fairly. In that context, being able to point to how different the result was from what the market predicted may be all that is needed to establish scepticism and challenge the outcome.

A single bet of 150 million kronor sounds enormous in a Swedish context. But in total around 20 billion kronor is at stake on Polymarket’s exchange alone. The question is how useful markets like Kalshi and Polymarket actually are at predicting the outcome of the American presidential election. The idea of listening to large groups to forecast the future was popularised in 2004 when journalist James Surowiecki published the much-discussed book The Wisdom of Crowds. It gives examples of how well-balanced groups can reach better and faster decisions than individual experts. But not all groups display wisdom. A stock market bubble is the opposite — everyone follows each other’s behaviour, fairly uncritically, and the majority ends up losing.

Looking at how Kalshi actually works, it is quickly clear that it is not designed to aggregate balanced voices that could outperform an expert. It looks like a betting site. Choose a question — which party wins the House majority? — and based on your answer you immediately see how much money you will win if you are right. By using terms like forecasting and prediction, sites like Kalshi and Polymarket have been attributed a seriousness they may not deserve. It is a slippery slope between a political forecast with a financial payoff and an old-fashioned bet. It is wagering dressed in the trappings of political science — and wagering is, as we know, no exact science.

Taleb makes a similar observation in Fooled by Randomness: “No matter how sophisticated our choices, how good we are at dominating the odds, randomness will have the last word.” Neither polls nor prediction markets can cleanly handle the randomness and unknown factors at play in something as vast as a political election. Is it harmless to let them run? That was the American legal system’s judgement, at least. But perhaps it underestimated the reasons people bet on these questions. You don’t necessarily bet to win the wager in the moment — you bet to influence others. And then, with randomness on your side, to win in the long run.


Did Klarna’s shareholders know what they were doing?

October 24, 2024
by Björn Jeffery
SvD Näringsliv





Did Klarna’s shareholders realise they were voting away their own influence?

Published in Svenska Dagbladet, 2024-10-24. Translated from Swedish.

The extraordinary general meeting at Klarna is over and the major shareholders got what they wanted. Sebastian Siemiatkowski strengthens his grip ahead of the listing. Why did the other shareholders let it happen?

“Minor administrative amendments to the articles of association” is not an agenda item that normally excites shareholders. But in Klarna’s case it concealed something that should have made the company’s shareholders furious — assuming they understood what it meant. There is much to suggest that many of them did not.

SvD has reported on the power struggle that preceded the extraordinary general meeting held in London on Thursday. Among the proposals passed was the ability for the board to remove members found to have breached their duties, and to appoint new members. Klarna — which now has a British parent company — is permitted to do this under British company law. In Sweden it would not have been possible: in a publicly listed Swedish company it is the shareholders, through the general meeting, who appoint at least half of the board members.

Power is now shifting from the general meeting to the board — which strengthens those already represented there, including CEO Sebastian Siemiatkowski and chairman Michael Moritz from venture capital firm Sequoia, while major shareholders such as co-founder Victor Jacobsson will lose out. His board representative, Mikael Walther, was forced to resign as a consequence of Thursday’s meeting.

What is not standard is that shareholders vote to reduce their own influence — at least not if they understand that is what is happening. According to Klarna itself, the move of the parent company to the United Kingdom was made to prepare the company for a listing, on the grounds that British law is internationally recognised and well understood by larger investors. Given that the listing will most likely happen in the United States rather than the UK, there were probably other options for domicile. But restructuring a company ahead of a listing is in itself neither unusual nor controversial — it is practically standard.

What is not standard is that shareholders vote to reduce their own influence. At least not if they understand that is what is happening. The decision to relocate was made in March this year. A share in the Swedish Klarna could be exchanged for a share in the British Klarna. It sounds simple, and shareholders reasonably did not want to stand in the way of the approaching listing — the prospect of liquidity beckoned. But then there was that matter of company law.

What Klarna voted through at its extraordinary meeting on Thursday was correct and legal under British company law. The equivalent manoeuvre in Sweden would not have been possible. So why did shareholders vote to reduce their own control over the company? Did they understand that the move — indirectly — entailed more than just listing preparation? A simple safeguard would have been to incorporate the same articles of association as in Sweden into the British company — to keep the same rules as before. But that safeguard was absent.

Whether this was a deliberate move or an unforeseen consequence can only be speculated about. The situation is now what it is. Power has been consolidated among existing board members, and a new standard articles of association has been established for Klarna — one that will likely be the framework presented to new investors at the listing going forward. Given the expected destination of the listing — the United States — that is unlikely to cause problems. The American stock market is full of companies with governance structures different from what Swedes are used to. Among the larger tech companies it is more the rule than the exception. The clarity it provides has probably helped several of them. Everyone understands who is driving and who decides. Zuckerberg is boss at Meta, and if you don’t like that, you shouldn’t invest in its shares.

But that reasoning only applies to the prospective new shareholders in Klarna — those who may come in after the listing and going forward. For the existing shareholders who have now lost much of their control and influence to the board, the situation is not so clear. They appear to have voted themselves out. The question is whether they understood that is what was happening.


How our grip on truth could crumble

October 22, 2024
by Björn Jeffery
SvD Näringsliv





Behind AI-generated deepfakes lies an even bigger problem

Published in Svenska Dagbladet, 2024-10-22. Translated from Swedish.

New digital tools have opened the door to mass manipulation that could influence the American presidential election. But behind AI-generated deepfakes lies an even bigger problem.

A girl in an orange life jacket sits in a boat. She looks as though she has been crying for a long time — tired, red-eyed. The water around her is brown and unwelcoming, and it is raining hard. She is holding a puppy. Hurricane Helene has just swept through American states including Florida, Georgia and South Carolina. The rescue operation has been difficult and the destruction enormous. The image of the little girl spreads quickly online as a symbol of the suffering Helene has caused. The girl does not exist. The image is AI-generated. Even an untrained eye can sense that something is not quite right — it looks partly animated, with a shimmer that gives it an unnatural gloss. Yet it quickly becomes a weapon in a political debate about whether society has prioritised the relief effort properly.

Fears that AI-generated images and deepfakes would influence American elections have existed for years. With the dramatically accelerating pace of AI development, those concerns have intensified ahead of this year’s presidential election. There are already plenty of examples. A user called “Think for yourself!” posted the fake image of the girl in the life jacket on X with the comment: “I don’t care if it’s AI, it’s still true!!!” The phenomenon raises an interesting question. Is it AI images influencing public opinion that we should worry about — or is there possibly an even bigger problem: voters deliberately allowing themselves to be influenced by an image they know is fake?

The term “deepfake” was coined on the internet forum Reddit in 2017 — a combination of “deep” from “deep learning,” a type of AI technique, and “fake.” On Reddit, this new technology was used to create videos where pornographic content was reworked to include celebrity faces. That genre of content has a tendency to be quick off the mark in major technology shifts. There is also no shortage of “cheapfakes” — poorly executed deepfakes using crude methods like pasting heads onto other bodies. The intent is the same, but the execution low quality.

In 2023 AI technology made a major breakthrough. Services like Midjourney and Stable Diffusion suddenly gave tech enthusiasts powerful tools to create new kinds of images. In March of that year an image appeared of the Pope wearing an incredibly elegant and fashionable white puffer jacket. It went viral immediately, with the Pope praised for his bold fashion choices. The image was, of course, fake. The creator had to issue an apology after what had seemed like a harmless joke spiralled out of control. In another example, Trump posted a series of images appearing to show Taylor Swift fans — so-called Swifties — rallying behind him politically. Also fake. In a Fox Business interview he distanced himself from the images, but in a telling way: “I know nothing about them other than somebody else generated them. I didn’t generate them.”

Trump didn’t create the images. But he spread them. And through that, uncertainty is created about what is true, what is uncertain, and what is entirely false. In a world of deepfakes, the opposite problem also arises: genuine photographs are assumed to be fake — or can at least be dismissed by a political opponent as exactly that. When presidential candidate Kamala Harris landed at Detroit Metropolitan Airport in early August, a large group of supporters with banners was visible beside the plane, enthusiastically cheering her arrival. Trump was not equally enthusiastic. On his own social media platform Truth Social he accused Harris of having manipulated the images: “Has anyone noticed that Kamala cheated at the airport? There was nobody at the plane, but she ‘AI’d’ it and then it showed a massive crowd, but they didn’t exist!” Given the number of people present, plenty of other images from the same moment existed. The crowd was real. But once the seed of doubt is planted, it becomes an argument one can deploy against almost anything. Does a picture make you look bad? Then it’s fake. Does a picture make your opponent look good? Also fake.

Political actors have always used a range of methods to smear opponents and try to win elections. In 1972 the American newspaper Manchester Union received a letter claiming that senator and presidential hopeful Edmund Muskie had used a derogatory term about a large voter group. The letter later turned out to have been written by an employee of the sitting president, Richard Nixon. It triggered a downward spiral for Muskie, who ultimately did not win the presidential nomination after all. A fake letter — a simple but apparently effective method. Another popular technique is robocalling — automated phone calls. In 2008, thousands of residents in North Carolina received a call in which a voice told them they would receive a voter registration form by post, which they should fill out and send back to ensure they could vote in the upcoming primary. The problem was that by the time the calls were made it was already too late to register, and the calls were going to people who were already registered. Confusion ensued, which may have prevented some from voting at all. The campaign was traced to a group called “Women’s Voices Women Vote,” which had connections to Hillary Clinton’s primary campaign.

What the introduction of deepfakes has done is dramatically lower the threshold — and the cost — for creating fake material. What previously required a professional video production team can now be done in a couple of minutes by anyone. The quality is often quite poor, and a new term has emerged to describe the enormous volume of low-quality AI imagery that has appeared: AI slop. Given the pace of AI development, we are months rather than years away from substantially more realistic images and videos of this kind. The companies behind these tools claim to have policies against such use, but enforcement is practically very difficult. And the damage can already be done by the time the source is identified.

The volume of political deepfakes is now so large that they have been documented in a database administered by researchers affiliated with Purdue and Northwestern universities. At the time of writing it contains over 540 examples.

There are two different perspectives on how the deepfake problem will develop. A pessimist would say it will likely get worse quickly. The quality of these services is improving, and in just the past few months AI tools for both audio and video have nearly exploded in capability. With better tools accessible to far more people, it is hard to believe the problem will resolve itself. Relying on human goodwill and good intentions in this context may be naive.

An optimist can note that despite this proliferation of new tools, the problem is still relatively contained. More fact-checkers than before — both news services and social media platforms — are now examining this kind of material. A fake image spreads fast, but it can also be debunked fast. AI development may even assist with that too. When Trump was shot at a political rally in Pennsylvania, an image spread appearing to show smiling Secret Service agents — as if pleased with the outcome. The image turned out to be false and was quickly verified as such by multiple independent sources. The problem is created fast, but the solution follows shortly after.

Taken together, we have a media landscape that may face a larger problem than individual fake images or video clips. Our shared sense of what is true and what is false risks eroding. The quality of the material does not necessarily determine whether someone believes it — they may simply have decided to trust the source, regardless of what it says. It takes only a drop of doubt before what we have collectively accepted as truth begins to crumble. Should that trend continue it will be a major challenge for society. But it is not strictly a problem that arose with AI and deepfakes. If — like the person who posted the girl in the life jacket on X — you have already decided what is true and false in the world, there are few things that can make you change your mind. Even when you know it is fake.


The paedophile alarm concerns every Roblox parent

October 20, 2024
by Björn Jeffery
SvD Näringsliv





Roblox accused of being a haven for paedophiles

Published in Svenska Dagbladet, 2024-10-20. Translated from Swedish.

The gaming platform has been described as a “nightmare landscape of paedophiles.” At the same time, two thirds of all primary school children in Sweden use it. Should parents be concerned about the accusations from Hindenburg Research?

“Escape to Epstein Island,” “Beat up the pregnant” and “Survive the killer” — these are some examples of games on the Roblox platform. In “Beat up the pregnant,” the objective is to kill pregnant women with knives and frying pans in a car park. The company itself describes its goal as to “connect a billion people with civility and optimism.” As you may notice, something does not quite add up.

Last week, short-selling firm Hindenburg Research published a report on Roblox. The platform allows users to create and publish their own games, which has led to a rather different range of content than games typically contain.

Hindenburg — named after the airship that exploded in New Jersey in 1937 — is, as the name suggests, a firm that profits when companies do badly. They short the shares and then publish reports — with obvious self-interest — that cast target companies in a poor light. When Hindenburg Research went after India’s Adani Group, the market value fell by over 120 billion kronor on the first day. Roblox’s share also took a hit after Hindenburg’s publication, though it has since recovered.

The firm’s criticism centres on two main areas. First, that Roblox is opaque about — and overstates — the number of players on its platform. Second, that it is a “nightmare landscape of paedophiles” where children and young people can encounter both inappropriate content and inappropriate people. Roblox allows children under 13 to play, but treats them differently on the platform — though in many cases children and adults play the same games simultaneously.

User numbers are an almost constant topic of debate for gaming and social media companies. What seems like a straightforward calculation rarely is. In 2016 Facebook had to apologise for having incorrectly reported how many people had watched video on its service — and overestimating video views had also distorted the associated advertising, which is Facebook’s primary revenue stream. In Roblox’s case it is a question of definition, and the company itself dismisses the criticism in a statement, saying its definitions are clear and properly disclosed. From the outside it is hard to judge — but making numbers look slightly more favourable than they are is hardly unique to Roblox on the stock market.

The second issue — around paedophiles and other inappropriate behaviour — is more serious. According to Roblox’s latest quarterly report, around 79.5 million users were active on the platform daily. Keeping track of all those users — particularly with chat functions and the ability to create custom games — is to put it mildly a challenge. Hindenburg’s report provides many examples of how obviously inappropriate content has slipped through the existing safety systems, and cites cases of multiple people arrested after attempting to arrange meetings with children on the platform. Games depicting simulated sexual assaults and similar content are numerous.

That sounds terrible — for parents especially. But provocative use of offensive names and actions is something that occurs throughout the digital world. It is not necessarily desirable, but it does not automatically mean that it is happening in reality, or that it was even intended to. Roblox maintains that it has a robust safety system that catches and blocks a great deal.

In an interview with Yahoo Finance, equity analyst Michael Pachter from Wedbush Securities commented on the report: “I don’t question that anything in the report is wrong […] but I’m not sure it’s relevant.” Pachter’s point is that the number of paedophile cases and similar incidents is very low relative to the number of users. He has a point — mathematically. But rarely has the contrast been so stark between what markets and ordinary people care about. A single paedophile is not much in statistical terms — but you don’t need more than one to cause serious harm.

Hindenburg Research are not noble knights concerned about children’s wellbeing. They are opportunists trying to make money by damaging a company that appears to have insufficient control over its operations. But what they are saying — regardless of their motives — is something that concerns every parent whose child plays Roblox. And there are many of us. Two thirds of all Swedish primary school children do.


The tech elite found an enemy — and found Trump

October 15, 2024
by Björn Jeffery
SvD Näringsliv





Lina Khan made Silicon Valley switch political allegiance

Published in Svenska Dagbladet, 2024-10-15. Translated from Swedish.

As long as the money kept flowing in, venture capitalists didn’t need to care about politics. A young lawyer changed that. Now the liberal tech power centre of Silicon Valley is a pillar of the conservative movement that wants to make Trump president.

The audience cheers. The stage is framed by stone, with an engraved reminder about the importance of education. We are in Royce Hall at UCLA in Los Angeles — just south of Bel-Air. Duke Ellington and George Gershwin have performed here. But we won’t find jazz musicians today. Quite the opposite. Seated in a row of grey sofas and armchairs are four Silicon Valley venture capitalists. Before a full house, the suited men are interviewing guests including Google co-founder Sergey Brin, Tesla CEO Elon Musk, and outspoken Uber founder Travis Kalanick.

This is the All-In Summit, an annual gathering born from the popular tech podcast of the same name. The talk — and the bickering — between David Sacks, Jason Calacanis, David Friedberg and Chamath Palihapitiya has become one of the most popular tech podcasts in the United States, with hundreds of thousands of weekly listeners.

The next guest on stage, however, is no tech bro like the rest of them. But he is indicative of a larger shift that has taken place in Silicon Valley ahead of this election. “Welcome vice-presidential candidate JD Vance to the stage!” In the 2024 election campaign, technology has become politics — and vice versa.

JD Vance is admittedly a former venture capitalist himself, but that is not why he is visiting. The most outspoken of the hosts, David Sacks, is a vocal Republican who has increasingly steered the podcast toward political topics. One segment of the conversation with Vance is subtitled “how to practically approach deportations, and who is coming through the southern border” — not an entirely obvious subject for an event full of tech enthusiasts. This election year the range of topics in Silicon Valley has expanded dramatically. A new kind of political power player is emerging. But why is this happening now?

Silicon Valley’s liberal image is well-known. In the Castro district, a giant rainbow flag has flown for many years — a marker of the city’s openness toward the LGBTQ movement. Silicon Valley as a whole — the area stretching from San Francisco south to San Jose — is in many ways a textbook example of a well-functioning melting pot. About 37 percent of the area’s population are immigrants, and three of the biggest companies in the tech industry — Google, Nvidia and Microsoft — all have immigrants as their top executives. The region is a mix of nationalities and worldviews, though the heavy concentration of engineers means the range of professional backgrounds is fairly narrow.

This has contributed to the image of a liberal tech sector — and one that has not cared much about party politics at all. Silicon Valley has been like California writ large: a reliable Democratic win, every election year. Uncontroversial, and therefore rather uninteresting. But the reality has always been more complex.

A 2017 survey revealed more nuance. Among tech company founders, views were liberal on social issues — abortion, immigration, support for the less well-off — but significantly more conservative on questions of regulation. And especially on labour regulation — trade unions and the like. Looking at the last presidential election, Trump received more than a third of California’s votes. In Santa Clara County — the district that is home to the headquarters of Apple and Nvidia, among others — one in four people voted for Trump.

Silicon Valley and the surrounding Bay Area is also an economically very prosperous region. Of around 7.5 million residents, over 300,000 are millionaires in dollar terms. The average price of a house in Atherton — an upscale suburb sandwiched between Menlo Park and Redwood City — is roughly 75 million kronor. There is, simply put, enormous wealth — and there has been for a very long time. And that very wealthy individuals tend to vote more to the right than the left is a pattern easy to identify, even within this otherwise liberal enclave of America.

The tech elite’s newly awakened interest in politics can be traced back to a specific date: March 22, 2021. That was when Joe Biden nominated Lina Khan, a young law professor from Columbia Law School, as the new head of the Federal Trade Commission. Her arrival was a turning point in how tech companies needed to relate to politics. It had been relatively easy and quiet until then. The tech giants had been allowed to become giants, with minimal intervention from regulators. That was about to change.

Khan was already well-known before her appointment. As a 27-year-old doctoral student she had published an academic article in the Yale Law Journal. “Amazon’s Antitrust Paradox” was the title — a nod to a well-known 1970s book on competition law, Robert Bork’s “The Antitrust Paradox.” In her article Khan argued for the opposite of what Bork’s book claimed. Judging competition cases by whether consumer prices went up or not — Bork’s main argument — was not applicable across all contexts, she argued. For tech companies — and Amazon in particular — low prices for consumers could still lead to a long-term erosion of competition. Amazon’s data on consumer behaviour helped it build a better store, to the point that competing with it could become nearly impossible over time. What felt cheap to individual buyers on Amazon could end up being very costly for society as a whole. The article had major impact and set the tone for how the FTC would approach competition cases in an era dominated by a handful of tech giants.

Under Lina Khan’s leadership, the FTC sued Microsoft when it tried to acquire Activision Blizzard, one of the world’s largest video game companies. It sued Amazon over convoluted subscription cancellation flows for Prime. And in autumn 2023 the agency sued Amazon for holding an illegal monopoly in the e-commerce market — the thesis Khan had advanced in her original article. The message is clear: maintaining fair competition in the tech sphere is worth fighting for.

For the first time since today’s tech giants grew large, they now have to reckon with politics. For boards and investors it has become an almost existential question. For Silicon Valley’s venture capitalists, Lina Khan has become a symbol of how politics interferes with and ruins a functioning market. In a blog post, venture capitalist Marc Andreessen wrote that “regulators are blocking and punishing startups from being acquired” — leaving little doubt about which regulator, and which relatively newly appointed head of it, he had in mind.

Investor Peter Thiel was one of the few to take political sides early — he spoke at the Republican National Convention as far back as 2016. This year he has been joined by heavyweights including Marc Andreessen and Ben Horowitz, two of the most influential venture capitalists in the world, both now publicly backing Trump (though Horowitz recently also donated to Kamala Harris). Why are they so invested in ending Biden and the Democrats’ time in power?

Part of the answer lies in venture capital’s fundamental business model. They invest in high-risk small companies, and if some of them grow, they can either be sold or listed. The modest initial investment has by that point multiplied and generates the fund’s entire return. But this model has run into trouble in recent years — roughly since Lina Khan took over at the FTC. The problem: no one is buying companies any more. Meta has not made a single major acquisition in the past two years. Google has made only two minor ones. Compare that to the equivalent period before Khan’s arrival, when each company completed around ten such deals per year.

At the same time, tech IPOs are at their lowest level since the 2008 financial crisis — worse than any year going back to 1980. The tap that watered venture capital has been turned off. The big companies no longer dare make the same kinds of deals as before, for fear of having them blocked by Lina Khan and the FTC. Buying Instagram for a billion dollars, as Facebook did in 2012, is today unthinkable.

As an alternative, many venture capitalists — including Andreessen and Horowitz — have invested billions in cryptocurrencies, a sector that appears headed toward heavier regulation, which is the last thing investors want. Trump has taken a clear stance on this. He spoke at the major bitcoin conference in Nashville earlier this year, discussing among other things using the cryptocurrency to pay down the national debt — exactly how was unclear. But the crowd cheered when Trump promised to fire Gary Gensler, head of the financial regulator SEC, who has become the face of all cryptocurrency regulation. After making the promise, Trump called out: “I didn’t know he was so unpopular!” For that audience, he is. And Trump’s message may be opportunistic — but it is at least clear: under him, there will be no more regulation of this industry.

Back at UCLA, David Sacks sits at the far right of the stage. This summer he hosted a fundraising dinner for Donald Trump’s presidential campaign at his home on what is known as “billionaire’s row” in San Francisco. Tickets started at around three million kronor. Sacks is no newly minted Republican, but his profile and volume are higher than ever before. As the political world steps into the tech world’s territory, figures like David Sacks have decided to make the reverse journey loudly. It is easy to understand why. After twenty years of near-total political free rein to do whatever they wanted, the region’s enormous wealth has only grown. The deals have got bigger, and so have the profits.

Now Washington DC is trying to catch up. And Silicon Valley — which until now has had every incentive in the world to stay quiet and keep its head down — is slowly waking up to a world where it may no longer be able to do exactly as it pleases.


A Nobel Prize for Google — which may now be broken up

October 9, 2024
by Björn Jeffery
SvD Näringsliv





Nobel Prizes for Google — which may now be broken up

Published in Svenska Dagbladet, 2024-10-09. Translated from Swedish.

Two Nobel Prizes linked to AI went to Google this week. But it is ordinary search ads that have paid for the development behind them — and now those ads are under threat.

Most companies would be thrilled to have had even one Nobel Prize recipient among their staff at any point. Google has just received two — in one week. Geoffrey Hinton, joint recipient of the physics prize, has admittedly left Google since last year, but worked there for ten years before that. Demis Hassabis, one of three sharing the chemistry prize, heads Google’s AI unit, DeepMind.

Both appointments are in different ways related to the development of artificial intelligence — an area that has been a core focus since Google’s current CEO Sundar Pichai took over in 2015. Less than ten years later, the world is in the midst of an accelerating AI race, with entirely new competitors like OpenAI’s ChatGPT challenging Google at the very heart of its business: delivering the right information, quickly, to those who search for it.

The Nobel announcements come at a particularly interesting — and somewhat ironic — moment for Google. AI development has so far been primarily a cost for the company. The hope is that it is an investment in the future. But what has paid for it to date is something considerably more mundane and undramatic: search advertising.

The way those ads are sold has begun to be questioned — including in the courts. In September a legal process began in the United States to examine whether Google’s ad sales constitute a form of monopoly. The outcome will be known in a few months. In a separate ruling from August, it was established that Google holds a monopoly in the search market. On Tuesday the US Department of Justice submitted its views on what remedies should be imposed to address that monopoly.

It was not pleasant reading for Google. “Radical and sweeping” was how a Google legal executive described the proposals. The DOJ wants Google to be broken up. If the American Justice Department gets its way, products like the Chrome browser and the Android mobile operating system could be forced to separate from the rest of the business — both to prevent Google’s dominant position in search from being further entrenched, and to make the search giant less competitive in artificial intelligence.

The significance of such a break-up would be hard to overstate. Revenue from search advertising accounts for around 57 percent of Google’s total revenue. Add roughly 10 percent from YouTube and around 9 percent from ads shown on third-party sites. Google may want to position itself as an AI company — but in all material respects it is still an advertising company.

The strong profitability from those ads is what has financed major acquisitions and AI investments. Both Geoffrey Hinton and Demis Hassabis came to Google through companies it acquired. Being able to pick up promising research projects early and let them develop inside Google has been a strategy that worked well for many years. Few other companies would have the means — or even the ambition — to run that kind of project without being able to directly attribute revenue to it.

DeepMind — the division where Demis Hassabis works — was acquired ten years ago for around four billion kronor at today’s exchange rate. Add ten years of salaries and expensive infrastructure on top. Even the T in competitor ChatGPT — transformer — is a technology originally developed by Google.

Something as simple as a text ad in a search result has thus paid for one of the most advanced areas of technological development we have today. What happens if the US Justice Department gets its way and Google is split into smaller pieces — and the company’s strong position in the advertising market weakens? The tap for some future technologies could be turned off. That may sound like a good outcome for the free market. But it is today’s ads that are paying for tomorrow’s Nobel Prizes.