Reality finally catches up with Lars Wingefors

April 22, 2024
by Björn Jeffery
SvD Näringsliv

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

As the Embracer group is dismantled on the stock market, it marks the definitive end of an era of low interest rates and high expectations. That the share surged on the news shows it should have happened sooner.

Lars Wingefors had hoped to get some rest over Easter. He said as much to SvD in March this year, after announcing that subsidiary Gearbox Entertainment would be sold.

Rest, it seems, was not forthcoming. On Monday, Embracer Group announced that the company will be split into three separate parts, each to be listed independently. Embracer in its current form will cease to exist once the transformation is complete.

To understand today’s news, one can look at the underlying promise that Embracer has carried since its inception. It can be simplified to the idea that “we are better together.” By acquiring a long list of gaming companies — covering everything from board games to mobile games — the combined group would be greater than the sum of its parts.

Or perhaps not better. But at least more valuable.

Because unlike conventional consolidation strategies, there were very limited synergies between the different segments within Embracer. Each ran its own race, and Embracer became a thin superstructure designed to hold everything together in the eyes of the market and shareholders. And more than anything else, the company was an acquisition machine that kept adding new businesses to the portfolio.

The strategy rested on two components: low interest rates and multiple arbitrage. When money was cheap, you could borrow enormous amounts of capital to buy new gaming companies. Those acquisitions were then added to Embracer’s consolidated finances — immediately rewarded with the premium valuation of a highly rated stock.

Both of those components have now disappeared. Interest rates have risen, making Embracer’s debt difficult to manage. They have had to sell several companies to service their borrowings. And the era in which they were rewarded with high multiples on the stock market is over. Over three years, the share price has fallen by around 77 percent.

Wingefors summed up the former era when responding to analyst questions: “We were living through the boom years of 2019, 2020, 2021. It was a different world. That world is now over.”

The market appears to have reached that conclusion before Embracer did. The share surged when the news broke, suggesting that confidence in the strategy — in Wingefors’s new world — had been severely limited.

Embracer’s official explanation is that the three separate entities will offer the stock market clarity. The board games business Asmodee and the working-title divisions “Coffee Stain & Friends” (mobile and PC games) and IP-based “Middle-earth Enterprises & Friends” have different profiles — financially and operationally. That is, of course, true. But the same was true when the company first tried to justify why all the businesses belonged together inside Embracer. Gaming is a broad category to consolidate. It is a bit like working in the category of “food” — yes, people eat it. But there is a difference between a street food stall and a slaughterhouse. And there is a difference between board games and mobile games too.

This confusion showed up not least in the criticism the company received for its complex reporting. Presenting adjusted EBITDA figures works best when the market fully understands what has been adjusted for. That has not always been clear at Embracer.

What has been clear is the group’s high level of debt. The sales of Sabre and Gearbox were intended to pay down a portion of it. As part of the restructuring, the board games division Asmodee will receive just over 10 billion kronor in new loans and financing in order to refinance existing debt.

Embracer captured the spirit of its time perfectly. But it was more a financial innovation than anything discernible in the actual businesses. Now that the zero-interest era has passed, the more than 125,000 individual shareholders and around 300 institutional investors have been left with a clean-up job — to bring clarity and order back to the underlying operations.

The solution is to dismantle the very structure that was sold to the market for so many years. Because even though Embracer packaged itself as a gaming company, it was really something quite different: a financial arbitrage play built on cheap debt. And that era — as Wingefors correctly notes — is now over.

Netflix forgives pirates — and raises prices

April 19, 2024
by Björn Jeffery
SvD Näringsliv

This analysis was first published in SvD Näringsliv, in Swedish, on April 19th, 2024. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.

Netflix beat all expectations when it presented its quarterly results. The company is a profit machine — but it increasingly resembles a perfectly ordinary media company. How do you justify the valuation then?

It has been described as one of the largest Swedish drama investments ever. The new adaptation of “Ronja Rövardotter” recently premiered on Netflix. But it was not the Americans who made the big investment in Astrid Lindgren.

The first two lines of the opening credit remind you of who did: “Netflix presents — a Viaplay Group series.” After Viaplay collapsed, Netflix was able to pick up the finished series and claim half the glory. The timing was perfect — for this kind of major production, there will not be many coming from either Netflix or anyone else in the near future. Acquiring a finished work from a rival is therefore ideal.

Netflix’s focus is now on profitability, after many years of what seemed like near-limitless growth. The shift is visible in its financial reporting. Previously, the company talked almost exclusively about subscriber growth. Now it is trying to operate more like an ordinary company — revenues and profitability.

And from next year, Netflix will stop reporting the total number of subscribers per quarter altogether. The figure has started to become misleading, they believe.

In the early days there was no profitability to speak of. But that looks very different now. On Thursday evening, the company reported its results for the first quarter of the year — and Netflix beat all expectations. Profit increased by a full 51 percent compared with the previous year. The company also beat market expectations on both revenues and new subscribers.

In keeping with becoming more like an ordinary company, one of the key drivers of success is something as simple as pricing. How do companies normally increase revenues? They raise prices. Which is exactly what Netflix has done. But to capture more price-sensitive customers, they have also added advertising to the cheapest subscription tier. Forty percent of all new customers now choose that option. The ads create a new revenue stream — still small for now, but growing.

Another much-discussed change has been getting people to stop sharing subscriptions. If you are logged in from many different locations simultaneously, Netflix will now issue a warning and offer the option to purchase a separate account — even allowing you to migrate your own profile so the system remembers what you have watched. They forgive your past piracy — but now it is time to start paying.

Content has also contributed positively. The latest season of the reality series “Love is Blind” was watched 20 million times, while the drama series “Griselda” reached over 66 million viewers. A notable new addition for Netflix is live sports — including a boxing match between Mike Tyson and Jake Paul — and in January they announced a deal with WWE, the American wrestling league.

The shift in content strategy is tied directly to the focus on profitability. The company’s new film chief, Dan Lin, has a mandate to produce cheaper and simpler films than before. A select few prestige projects are still being greenlit to keep the awards conversation alive, but the general ambition — and budget — is substantially lower. Lin began his new role by laying off ten percent of his staff, then reorganising those who remained. The message was clear: this is a new era.

Quality, however, is a relative concept. Expensive productions are not necessarily more popular than cheap ones. Netflix has had considerable success with stand-up specials, for example — Ricky Gervais and Dave Chappelle were both among the most watched content during the quarter. They certainly charge well for their time, but from a production standpoint, stand-up is remarkably cost-efficient.

Despite these strong results, Netflix’s share price fell in after-hours trading. When every metric is pointing upwards, what more can the market want? Part of the problem is that Netflix’s stock has already risen enormously compared to a year ago — up 89 percent — and the company is now approaching the peak valuation it reached at the end of 2021.

The challenge for Netflix is to sustain that valuation — but for different reasons than the last time. Then, it was still low interest rates and cheap money driving growth. Now, Netflix’s CFO Spencer Neumann talks about steadily rising profit margins. That sounds rather like an ordinary media company. So why should it be valued on the stock market like a tech company? That is the question Netflix needs to answer.

Sweden hopelessly behind in the global AI race

April 15, 2024
by Björn Jeffery
SvD Näringsliv

This analysis was first published in SvD Näringsliv, in Swedish, on April 15th, 2024. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.

Countries and regions are racing to stake their position in the explosive AI development. But when it comes to infrastructure, Sweden appears to have already given up.

It looks like a circuit board in a desert — at least when you view the satellite images. Across just over 450 hectares, just north of Phoenix in the US state of Arizona, lies what is set to become the foundation of the next wave of domestic chip manufacturing in the United States.

TSMC — the Taiwanese manufacturing giant — broke ground on the new facility in April 2021 and is targeting production in 2025. The following year, 2022, plans were announced for a second factory. And this week, a third.

In total, TSMC is investing nearly 700 billion kronor in Phoenix. That is a great deal of money — even for the world’s largest producer of chips and semiconductors. And it matters enormously for the city of Phoenix, the state of Arizona, and the United States as a whole. TSMC is receiving around 71 billion kronor in various subsidies — just to complete the latest project alone.

“Chips are the foundation of all artificial intelligence and essential components for the technology underpinning our economy,” commented US Commerce Secretary Gina Raimondo.

The United States is not alone in wanting to stake a position here. Beyond China — one of the dominant players in AI — a feverish build-out is underway across the Middle East to become the region’s hub for artificial intelligence. The focus there is not primarily on chip production, but rather on the data centres in which chips are used. The Financial Times has reported that both Saudi Arabia and the United Arab Emirates have bought thousands of Nvidia’s most sought-after chips to secure access to frontier technology. Both countries have also developed their own language models, similar to those underpinning ChatGPT.

To power the data centres, grid expansion has begun in both countries. Already, Chinese giants such as Tencent and Alibaba, as well as American Amazon, have all committed to operating data centres in Saudi Arabia. The sector is new — and growing rapidly.

Beyond data centres and infrastructure, enormous amounts of investment capital are being deployed. Abu Dhabi’s sovereign fund Mubadala is participating in the establishment of a new company dedicated to investments in AI-related sectors. The fund is targeting over 100 billion dollars — more than 1,000 billion kronor — within a few years. And Saudi Arabia is in talks with US venture capital firm Andreessen Horowitz about investing 40 billion dollars in a joint fund, according to Bloomberg.

In an interview with SvD, the chairman of the Swedish AI Commission, Carl-Henrik Svanberg, said that the US and China have already won the battle for infrastructure. For Sweden, the focus should instead be on building products and services on top of that infrastructure, in Svanberg’s view. He declined to say what the commission he chairs will have achieved before the end of the year, but they will produce a final report by the summer of 2025.

Running parallel to this, the Swedish initiative AI Sweden — a national centre involving participants from both the private and public sectors — has published an AI strategy for the country. It describes a society that makes the most of AI’s potential and builds on areas where Sweden is already strong, including democracy, a developed welfare state, and social stability.

The question is how useful such a strategy really is — and why it was only published now, given that AI Sweden was founded back in 2019. Among the strategy’s core principles, one can read that Sweden should “ensure responsible use,” “encourage collaboration,” and “engage with and learn from the best.” No doubt — but no country has a strategy that advocates impeding collaboration or learning from the mediocre.

In early April, Canada launched a new fund to strengthen domestic AI initiatives. Nearly 19 billion kronor will be invested in this first phase. The money is intended to help Canadians get the most out of AI and strengthen the country’s economy, according to Prime Minister Justin Trudeau.

AI Sweden has a total budget for 2024 of around 150 million kronor. The Swedish AI Commission has an appropriation of 4.8 million kronor. Svanberg tells SvD that Sweden will not become a “tech leader” in the field. The hope is instead to be quick at adopting the technology that others develop.

With the current resources, priorities, and pace, it is an open question whether Sweden will even manage that. Around the world, billions are being invested in factories, data centres, and infrastructure to support this new technological shift. But in Sweden, we will not even see a report with a plausible plan until the summer of 2025. By then, a great deal will already have happened in the global race.

Tesla is now being punished by its own strategy

April 5, 2024
by Björn Jeffery
SvD Näringsliv

This analysis was first published in SvD Näringsliv, in Swedish, on April 5th, 2024. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.

Tesla dominated the EV market for years and was valued like a fast-growing internet company. Now that growth has stalled, a less flattering picture emerges — a fairly ordinary car company.

They were called the “Magnificent Seven.” Seven stocks accounting for 1.4 percent of companies on the S&P 500, but whose performance has outstripped everything else on the market. The seven companies given this epithet are Amazon, Apple, Google (Alphabet), Meta, Microsoft, Nvidia, and Tesla.

But now the seven are moving in different directions. One stands out particularly — in the negative sense. While Nvidia and Meta have risen 85 and 46 percent respectively so far this year, Tesla has seen its market value fall by a third. It has also been overtaken in size by a genuine stock market legend: Warren Buffett’s Berkshire Hathaway. Is old age proving wiser?

Perhaps. But the reasons for Tesla’s difficulties have less to do with novelty than with ordinary competition — and a strategy from an era that may now be over.

One of Tesla’s long-term goals was to achieve an annual growth rate of 50 percent from 2020 onwards. It started strongly, with 87 percent growth in 2021, followed by 40 percent in 2022. The share price multiplied during that period. The electrification of the car industry had gathered pace, and Tesla had a new model — the roomier Model Y — which opened up a new customer segment. Was there nothing that could stop them?

Not if you believed the CEO, Elon Musk. When asked about the 50 percent target — which looked set to be missed — during an analyst call, he said it was not “possible to have a 50 percent annual growth rate forever, because then you would exceed the mass of the universe.” Musk was technically correct, but the target had been set by him just three years earlier. The laws of physics presumably applied then too.

In recent days we have seen further evidence of Tesla’s difficulties. The number of cars delivered in the first quarter fell 8.5 percent compared with the same quarter the previous year. Deliveries also declined versus the last quarter of 2023. That is unusual. Growth has not merely stalled — it is going backwards. Hedge fund manager Per Lekander — who has shorted Tesla — goes so far as to say the company could go bankrupt.

There are several reasons for this. Competition in the EV market has increased substantially. From established players like Volkswagen to the rising Chinese giant BYD, the pressure is intensifying. Tesla’s models, despite some updates, are essentially the same ones they have had for years. The Model 3 launched in 2017 and the Model Y in 2020. In that time, competitors have caught up and offer many new models and alternatives.

Another problem is Tesla’s strategy of owning the entire production chain — all the way from the factory to final delivery to the customer. In a strong upswing, Tesla keeps all the profit, having no dealerships to share it with. But when growth stalls, the strategy can produce the opposite effect. Tesla bears all the costs — at every level. Factories, employees, and showrooms are fixed costs to manage regardless of how many cars they sell.

The final problem is valuation. Tesla’s trajectory between 2020 and 2022 most closely resembled an internet stock with explosive growth. The share was priced as if it were software — rapidly scalable, with an almost infinite market ahead of it. That picture was never quite accurate, but Tesla could perhaps have been more opportunistic while the conditions existed. A capital raise when the stock stood at 300–400 dollars would have delivered low dilution and an enormous war chest to work with. Such a cash pile would be useful now.

But building cars is not just software. And Tesla’s share price has more than halved from its peak. If they cannot find a way to reignite growth, the risk is that they begin to be seen — and valued — like any other car company. Were that to happen, Tesla could face an entirely new kind of challenge: a stock market collapse of a scale we have only seen the beginning of.

Why they’re throwing money at Trump

March 26, 2024
by Björn Jeffery
SvD Näringsliv

This analysis was first published in SvD Näringsliv, in Swedish, on March 26th, 2024. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.

The valuation is astronomical as Donald Trump’s company goes public. The reasons to invest in Truth Social have nothing to do with business fundamentals.

Giving your listed company the same ticker as your own initials is unusual. Doing it a second time should, by any reasonable measure, be unique.

Nothing about it is normal when Donald Trump’s social network, Truth Social, goes public on Tuesday. The company being listed via SPAC is called “Trump Media & Technology Group Corp,” but will trade under the ticker “DJT” — Trump’s own initials.

The prospects look better this time than when Trump took a casino public in 1995 — also under the name “DJT.” That ended with the company running into financial difficulties and eventually filing for bankruptcy protection in 2004.

This time, it is far more than the threat of bankruptcy that hangs over Trump’s new public company. Its figurehead is a presidential candidate facing criminal charges in a string of legal proceedings. He has also been found liable and needs liquidity to pay fines running into the billions.

Despite all this, the SPAC’s shares surged when the merger was confirmed. What are investors hoping for?

It is not anything as conventional as revenues and profitability, at least. DJT — which in all material respects consists solely of the right-leaning social network Truth Social — had revenues of around 36 million kronor and made a loss of over 500 million kronor during the first nine months of 2023. Such numbers make it difficult to justify a valuation of around 53 billion kronor, which is roughly where the share is expected to trade initially.

A simple calculation shows that DJT would then be valued at approximately 1,100 times its annual sales. Comparing that to equivalent figures for other social media companies: Meta trades at 9.5 times, Snap at 4, and the recently listed Reddit at 6.4.

Investing in DJT has nothing to do with business fundamentals. There are, however, two other powerful motivations at play. Either the hope that something irrational will happen, or an expression of ideological conviction.

The irrational dimension finds its clearest expression in Reddit forums like Wall Street Bets — best known as the gathering place for the GameStop hysteria of 2021. There, traders discuss options contracts with extremely high risk. In that context, the quality of the underlying company matters little — what counts is whether you correctly predict how the price will move. A kind of lottery, but on the stock market. A stock like DJT — with Donald Trump as its face — becomes the perfect candidate for a so-called meme stock. Anything can happen.

The ideological dimension is more about investing in Donald Trump, rather than alongside him. Trump has not bought shares in DJT at this valuation; his ownership is a result of the SPAC merger. Buying shares can become a way of expressing support for Trump.

The timing works well for Trump. In the coming days he needs to raise billions to prevent his property assets and other holdings from being seized by the courts. Trump’s agreement with DJT stipulates that he cannot sell any shares during the first six months after the IPO. But the agreement can be renegotiated — no law governs his ownership. Were that to happen, Trump would be able to sell shares and thereby resolve some of his financial difficulties with the justice system.

There is, however, a problem with this plan. To be a seller, you need buyers. Will the market stand ready to buy Trump’s shares at the current valuation, even when the largest shareholder and the company’s namesake is selling? If not, the share price would fall dramatically.

Buying shares from Trump can also be done with other, less noble, motives. In the US, there are laws regulating how much money you can donate to presidential candidates. But buying shares — and making a bad investment — is not illegal. The major Republican donor Jeffrey Yass, for example, is already a significant shareholder in both ByteDance (TikTok’s parent company) and in the SPAC that is now being converted into DJT.

Being close to the man who could become the next President of the United States can be worth a great deal of money. In that context, even a bad stock trade might end up paying off.

The strategy that gave the AI giant a head start

March 19, 2024
by Björn Jeffery
SvD Näringsliv

This analysis was first published in SvD Näringsliv, in Swedish, on March 19th, 2024. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.

A supremely confident Jensen Huang presented Nvidia’s plans to the world. The message was unmistakable: without Nvidia, there is no AI development. But on the horizon, competition is stirring from a troubling direction.

Jensen Huang, CEO of Nvidia, walks onstage to bombastic orchestral music that could have come from a Hollywood blockbuster. “You have come to a developer conference,” he clarifies.

The music, it turns out, was written by AI — but performed by a live symphony orchestra. Huang himself, in his now-familiar leather jacket, seemed to harbour a certain rock star energy. But only so much.

GTC — Nvidia’s own conference — has been held since 2009. Back then it was a small, internal affair. On Monday evening, thousands of participants gathered in San Jose — with many times more following the livestream — to see what the company at the centre of the AI revolution has planned for the future.

Huang is acutely aware of the position he now occupies — and how far they have come since he co-founded the company in 1993. He pauses on 2017, when Nvidia produced a new kind of computer called the DGX-1. “I delivered the first unit by hand to a startup in San Francisco,” he recalls. “They were called OpenAI.”

The message is clear: without Nvidia, no OpenAI, and no AI revolution. Now it is about scaling up further. From the stage, Huang proudly unveils the company’s next major platform — Blackwell — which will make generative AI more powerful than ever. “God, I love Nvidia,” he says spontaneously. It shows. And he has 23 trillion kronor worth of reasons to feel that way — that is Nvidia’s current market capitalisation.

Is there anything that could stop them?

The new-found success has allowed Nvidia to broaden its operations considerably. Not only do they supply the hardware on which AI development runs — they have also expanded into financing the companies that do the developing. Last year, Nvidia invested in over 30 AI-related companies, and a month ago bought a larger position in the listed Arm Holdings. Nvidia had previously attempted to acquire all of Arm, but that deal was blocked in 2022 by competition regulators. At the end of January, the value of these investments stood at over 16 billion kronor.

The investment activity stands out because it creates a kind of financial circular flow. Nvidia invests money in promising AI companies, which then spend a substantial portion of it on buying products and services from Nvidia itself. Access to Nvidia’s chips has been a challenge for the entire industry. Spending newly raised capital on them is entirely rational. What becomes slightly unusual, however, is that the money comes from the same supplier that companies then must buy from. Investments in other companies effectively become direct revenues for Nvidia.

The insights gained into how products are used also help shape the next generation of development — and make inventory management easier, since demand becomes more predictable. Right now it is exploding, but as recently as mid-2022 Nvidia had to write down part of its inventory after misjudging market demand.

Investing in customers also keeps competitors at bay. By financing companies that use Nvidia’s products, Nvidia creates both demand and a head start for those chosen firms.

Competitors do exist, even if the market currently seems to value Nvidia as unbeatable. The share price is up just over 80 percent since the start of 2024, and over 240 percent compared to a year ago. Nvidia leads specifically in what is called the GPU — graphics processing unit — which, broadly speaking, is the type of processor used to train AI models. But established players such as AMD, Intel, and Qualcomm see the explosive growth and are unlikely to let it pass them by forever.

Another competitor that receives less attention is Apple. They do not sell chips to others, but they manufacture their own — which means they can build chips specifically designed to support Apple’s own strategy. Apple is therefore not a competitor to Nvidia in selling chips to external customers, but rather in terms of the AI functionality that is built with them.

Nvidia’s software development platform, CUDA, is today a standard for AI development. But building platforms of this kind is something Apple has been world-class at for years. In the longer run, this could give rise to a new kind of competitive dynamic between giants in the multi-trillion-kronor bracket.

In the short term, Nvidia still stands largely alone at the top. They have outmanoeuvred the existing players in the industry. Who talks about Intel today? And even if players like Apple eventually mount a challenge, it is many years away from becoming reality.

That more competition will come is, however, entirely certain. Large and small — from startups to tech giants — everyone sees the explosive AI development and wants to stake a position. How long can Huang and Nvidia hold on?

This is when TikTok gets banned — or is it?

March 13, 2024
by Björn Jeffery
SvD Näringsliv

This analysis was first published in SvD Näringsliv, in Swedish, on March 13th, 2024. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.

This is about politics more than technology as the US House of Representatives voted to ban TikTok. Much suggests the bill will not become reality.

When American TikTok users opened their app last week, they were not met with the usual lip-syncing to music. Instead, they were greeted with a black screen reading “Stop a TikTok shutdown.” Below it sat a red button with a phone receiver and the instruction: “Call now.”

US lawmakers were flooded with calls from angry TikTok users worried that their entertainment — and in some cases their livelihood — was about to be taken away.

The calls were not enough. After passing through a committee in the US House of Representatives with the remarkable vote of 50-0, the full House voted on Wednesday. There too, both sides of the political spectrum agreed that TikTok could be considered a risk to US national security, and should therefore be sold or banned outright.

The bill is written specifically for TikTok and its Chinese parent company ByteDance. If it passes all stages, ByteDance would have only 165 days to divest TikTok. But there is much to suggest this will take considerably longer — if it ever gets that far.

Scepticism towards TikTok is nothing new. As far back as 2020, then-President Donald Trump said his administration was examining options to ban the app in the US. It ended in a kind of compromise, with TikTok promising to separate data from American users and store it on US soil, with database company Oracle ensuring it was not misused.

Two years ago, in 2022, the migration of American user data began under what TikTok called “Project Texas.” The arrangement was on track to be completed.

But it did not stop there. In 2023, President Joe Biden banned TikTok from government phones, and the state of Montana banned the app entirely. Montana’s decision was subsequently struck down by a federal court that ruled it violated freedom of speech.

Trying to ban TikTok again now is therefore far from straightforward — despite unusually broad political consensus on the issue.

The committee that drafted the bill is called the “Select Committee on the Chinese Communist Party,” which suggests which motives have been strongest in driving it forward. When Republican politician Mike Gallagher was appointed committee chairman in December 2022, he said: “Even in a divided government, we have an opportunity to build a united front against the aggressions of the Chinese Communist Party.”

In an election year, few politicians want to risk being associated with China — and Gallagher may have found exactly that in TikTok. Voting against TikTok becomes a clear way to signal that.

The bill still faces a difficult path. It must first be taken up and voted through the Senate — which is not a given. Senate Majority Leader Chuck Schumer has promised to look at it, but no vote has been scheduled. Several Republicans have also expressed doubts about a ban — including Donald Trump, who now appears to have changed his position on the issue.

If it nonetheless passes the Senate — and President Biden signs it, which he has indicated he would — a legal challenge is almost certain to follow. The Montana case suggests this could be difficult. And if nothing else, such a process would take a long time.

One plausible outcome is therefore some form of compromise. ByteDance will not be able to continue exactly as before, but there are paths forward. A complete ban is unlikely. A sale is conceivable, but given that the app is valued at a minimum of 500 billion kronor, the list of plausible buyers is short. The very largest — Meta, for example — would happily buy it. But such a deal would never be approved on competition grounds.

One variant would be for parent company ByteDance to list TikTok on the US stock market. Ownership would then be distributed across many more shareholders, with new major investors entering the picture. ByteDance might then be permitted to remain as a minority shareholder.

Whatever happens, we are unlikely to get any definitive answer on TikTok’s future before the US election in November.

A new Pew Research study showed that 58 percent of Americans aged 13 to 17 use TikTok every day. 17 percent of them said they do so “almost constantly.” They cannot vote yet — but the app has around 170 million users in the US. Upsetting that many people in an election year is something few politicians are willing to do.

3 books, 3 movies, 3 songs + 1 prediction

March 11, 2024
by Björn Jeffery
Newsletters

👋

I have been writing so much lately. But since I haven’t sent you a newsletter since… August 2021(!?), I can’t blame you for not knowing that. It has also been almost exclusively in Swedish (sorry).

I have also been reading a lot more books and listening to a lot of music. Which brings me to what this newsletter once was intended for and named after – my recommendations.

With that in mind – below are some things that I’d like to recommend!

/Björn

Three books that stayed with me after I finished them

When We Cease to Understand the World by Benjamín Labatut
One of the best and most fascinating books I have ever read. It moves seamlessly between fact and fiction to create something quite unique. A book I will read again.

The Guest by Emma Cline
A novel which follows a young woman through a series of serendipitous encounters. It has a stressful feeling throughout the whole book – a bit like watching The Bear. Stressful, but in a good way. Hard to put down.

Unscripted: The Epic Battle for a Media Empire and the Redstone Family Legacy by James B Stewart and Rachel Abrams
If you like Succession, this is the book for you (and if you don’t, why are you reading this?) Logan Roy and Sandy Furness are both partially based on Sumner Redstone, the patriarch behind CBS and Paramount, and this book captures his escapades. The real story is incredibly juicy.

Three movies that made me think

The Zone of Interest by Jonathan Glazer
Perhaps no one has missed this one, but just in case you have: see it. The setting is a Nazi commandant that lives next to Auschwitz with his family. Life goes on, and on the other side of the adjacent wall – it ends. Also, I’ve never watched a movie with sound design like this.

Aftersun by Charlotte Wells
A father and daughter go on vacation together, seemingly to create the illusion of normality. But real life simmers in the background. A beautiful and somewhat tragic illustration of the juxtaposition between being a parent and an adult.

Past Lives by Celine Song
A sort of modern day “Sliding Doors”, without the clichés. About a potential life that gets lost and then makes a new appearance. The one that got away? Or perhaps the one that could never have been?

Three songs that I’ve been playing over and over

Bzrp Music Sessions, Vol. 58 by Young Miko
The Argentinian producer Bizarrap has a series where he invites artists to collaborate. The best one I’ve seen is with the Puerto Rican rapper Young Miko who has an incredible flow on this one.

Bye by Ariana Grande
The best track from Ariana’s new album that just dropped. Classic pop, but so well-produced.

Deeper Well by Kacey Musgraves
Likely the artist I’ve listened the most to over the last few years. The dark post-divorce album Star-Crossed was a favourite of mine, but she seems to have turned a corner now. A simple, but beautiful melody.

One prediction I have made

Embarrassing AI mistakes could knock out top executive (in Swedish here)

After a series of embarrassing mishaps, I think Sundar Pichai – the CEO of Google – will step down before the end of the year. Being late to the AI game is okay. But being bad at it – even in its infancy – is not.

Originally published on Substack on March 11th, 2024.

Embarrassing AI mistakes could knock out top executive

March 8, 2024
by Björn Jeffery
SvD Näringsliv

This column was first published in SvD Näringsliv, in Swedish, on March 8th, 2024. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.

It is one of the world’s largest and most influential tech companies. Despite this, Google’s AI launches are marked by embarrassing mistakes. Now questions are being raised if it’s time for its leader to step down.

Fire. Described by many as the most important discovery in history.

It’s easy to understand the surprise of legendary journalist Kara Swisher.

The year is 2018 and Swisher is interviewing Google CEO Sundar Pichai on stage.

“We don’t take just a very optimistic view of AI. AI is one of the most important things humanity is working on. It is more profound than, I don’t know, electricity or fire”

Swisher responds with slight skepticism to the tech exec’s comment.

“Fire? Fire is pretty good. But go ahead.”

At the time, Pichai was about two years into his new role as CEO of Google. He had already said that the company would be “AI-first”, but not much attention had been paid to it.

This time he took it a step further. Google took AI very seriously – and would become a leader in the field. AI – more important than electricity and fire.

Fast forward to the fall of 2022. The somewhat oddly named ChatGPT is launched to the outside world by the then rather unknown company OpenAI. A new era of AI starts. But Google is not the one leading it. If anything, they seem to have fallen behind.

What happened to Pichai’s grandiose ambitions?

After a number of unsuccessful launches, there are now many people – both inside and outside Google – who are asking themselves that question.

It started with the chatbot Bard, which was Google’s equivalent of ChatGPT. It was released three months after its competitor, in a move that looked rushed. In its launch video, Bard gave the wrong answer to a question about astronomy. The stock fell by $100 billion in market capitalization when the error was discovered. If Google can’t even show the right answer in its marketing material, can it be trusted at all?

A few weeks ago it happened again.

Google Gemini – the new name for Bard – was rolled out broadly. Now you could create text and images using Google’s latest, and most capable, AI models. But when Gemini was asked to create images of German soldiers from 1943, a user was shown examples of an Asian woman and a dark-skinned man in Nazi uniform.

There were plenty of other mistakes too.

When asked if Hamas was a terrorist organization, Gemini replied that the conflict in Gaza was “complex”. The AI bot also had a hard time deciding whether Elon Musk or Adolf Hitler had done the most damage in the world.

Google was forced to apologize, has paused the launch, and withdrawn some functionality until further notice.

Being first at major technology shifts does not necessarily indicate long-term success. Nokia was way ahead of Apple with mobile phones. Myspace was the biggest in the world before Facebook took over. And even in the search market – although it feels distant today – there was a time when the market leaders were called Altavista and Yahoo instead of Google.

In an interview with the New York Times, Pichai comments on the perceived stress and how he sees it specifically in AI:

“I don’t want it to be just who’s there first, but getting it right is very important to us.”

But in this case, they were neither first nor right. If Pichai thinks quality in AI is so much more important than speed, why are their launches plagued by so many mistakes?

Although right-wing American media saw the Gemini examples as a telling sign of a company that was now too “woke”, there is a lot to suggest that the service simply wasn’t very good – and was poorly tested. Not finished. It is well known that AI services “hallucinate” and create things that do not exist.

If Google wasn’t in a rush to launch, obvious mistakes like these should have been caught before it was released to the public.

All in all, there’s a company leader – for one of the world’s largest and most influential companies – who seems significantly more stressed than he lets on. And there are other clues that suggest this too. More and more reports pop up online from disgruntled employees who think Google has become bureaucratic and slow. Unable to create new, world-leading products. Can Pichai withstand this pressure?

Big tech trends like cryptocurrencies and the metaverse have passed Google by. There was nothing in them that directly affected either their core business or their future ambitions.

AI, however, is of a completely different nature. It is an area that could shake up the somewhat stagnant search market that has been incredibly profitable for Google for the past 20 years. They don’t have to be first, although Pichai would probably like to be. But if they are not the first, they must at least be the best.

Currently, they are neither.

One could understand that Pichai might be a little bit worried after all. If it continues like this, he may have to look for a new job before the end of the year.

This column was first published in SvD Näringsliv, in Swedish, on March 8th, 2024.

Musk’s new path to power reeks of desperation

March 4, 2024
by Björn Jeffery
SvD Näringsliv

This analysis was first published in SvD Näringsliv, in Swedish, on March 4th, 2024. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.

OpenAI is being sued by co-founder Elon Musk. He claims to want to preserve the company’s original non-profit ambitions. But Musk has other reasons to want to slow down what has now become his competitor.

The complexity of the case is revealed immediately in the lawsuit Elon Musk has filed with the California court system. On one side stands “Elon Musk, an individual.” On the other: OpenAI’s leaders Sam Altman and Greg Brockman — and eight different OpenAI entities, all with names along these lines: OpenAI LP, OpenAI LLC, and OpenAI GP LLC.

Does all of this really exist for something that was meant to be a non-profit aimed at doing good for the world through artificial intelligence? Musk doesn’t think so — and is now suing them.

Elon Musk was one of OpenAI’s original founders and among its earliest and largest funders. It is true that the company subsequently restructured to include a more commercial entity. Hence all those interlocking company names.

But this transformation happened back in 2019. So why is Musk only suing now, five years later?

The first — and most charitable — interpretation is that it is only now that OpenAI has truly broken through at scale. AI development is accelerating in step with the enormous investments being made in the field. The risks that some AI researchers have warned about may therefore be becoming more urgent.

OpenAI was originally set up to ensure that AI technology is used with good intent and in a way that benefits humanity. Musk has been clear about his desire to colonise Mars as an alternative home for humanity. If he genuinely fears that AI could be used in the wrong way, this is the moment to try to stop it.

Working against this interpretation, however, is reporting in The New York Times suggesting that as recently as 2018, Musk was encouraging OpenAI to accelerate its development pace in ways that one researcher there considered “reckless.”

The second reason is that Musk now competes directly with OpenAI through his own companies. The publication Semafor has previously reported that Musk attempted to take over OpenAI in 2018 following a dispute with Sam Altman. When that failed, Musk withdrew a large planned donation.

Several years later, Musk now runs multiple companies with AI connections — including Tesla and the newer project xAI. Interestingly, xAI has been registered as a commercial “benefit corporation,” a legal form designed to benefit more than just its shareholders — a kind of hybrid between the non-profit and the commercial. Regardless of how xAI is structured, it is a direct competitor to OpenAI. Slowing down or forcing OpenAI to restructure could meaningfully benefit Musk’s own ventures.

The third reason is a question of power. Access to AI services is growing explosively, but the underlying systems that provide them are few. At present, it is primarily Google, OpenAI together with Microsoft, Anthropic, Meta, and a handful of others that underpin almost all AI services. This means that the concentration of power is very high — resembling the position that the tech giants have occupied on the internet for the past fifteen years.

Musk is not one of those power players today. And it must sting that he helped fund something that now generates both profits and a formidable competitive position for several of his rivals.

Neither Microsoft nor OpenAI accepts Musk’s characterisation of their relationship. They argue it is a commercial partnership, but that Microsoft cannot control what OpenAI does — and that the non-profit element of OpenAI remains intact as the legal entity where the governing board sits.

It is easy to be idealistic about the current development of AI and how one wants it to be used. But the reason OpenAI went to Microsoft for money in the first place was that the operation required enormous resources — as we can see not least from chipmaker Nvidia’s extraordinary rise. Had OpenAI been funded through donations, it would almost certainly not have achieved the results we see today, nor would it have accelerated competitors’ appetite for investment. Billions of dollars have been committed to this development.

Elon Musk was early in recognising the potential of artificial intelligence. It is easy to see that he would have liked a larger hand in shaping it. Now he has another chance to get that — not through technology, but through the courts. It is a move that carries a whiff of desperation. A San Francisco judge will decide whether he succeeds.