Google’s answer on AI is logical — but hollow

May 16, 2023
by Björn Jeffery
SvD Näringsliv

This analysis was first published in SvD Näringsliv, in Swedish, on May 16th, 2023. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.

Google is positioning itself as the responsible player in artificial intelligence. It sounds good. But the stance could become complicated as competitors pick up the pace.

There was no rock band performing. It was more like a tech festival — with AI as the theme.

The stage at Shoreline Amphitheatre in Mountain View, California, was unusually colorful when Google CEO Sundar Pichai stepped up.

Google’s annual developer conference, I/O, tends to be a fairly sleepy affair. A raft of internal projects is presented at an event aimed primarily at the developers who build on Google’s products.

For the average user, the presentations can easily feel a bit too technical and inward-looking.

This time, however, there was a theme the outside world was genuinely interested in. AI — artificial intelligence — is on everyone’s lips, and now we would finally learn how Google planned to respond to the threat from products like ChatGPT.

It took only a few minutes into Pichai’s presentation before the key word was uttered: “responsible.”

“With a bold and responsible approach, we are reimagining all of our core products — including Search,” Pichai said.

It has now been seven years since he declared that Google would become an “AI-first company.” But if that was the case, how had they ended up falling behind in this latest boom? The apparent internal answer seems to be precisely that Google has been more “responsible” than the rest. The phrase recurred many times throughout the presentations.

To understand where this framing of responsibility comes from, we need to rewind a little.

On the same stage in Mountain View in 2018, the same presenter — Sundar Pichai — demonstrated how an AI tool could phone a hair salon and book an appointment, apparently without the hairdresser realizing she was speaking with a robot.

The reception was cool. Professor and columnist Zeynep Tufekci described it as an example of Silicon Valley having lost its ethical bearings.

Two years later, Google found itself in controversy again, after firing researcher Timnit Gebru. She had led a team examining the ethics of AI and the potential consequences of its development.

Gebru was also co-author of a research paper that a Google manager objected to. The dispute couldn’t be resolved, and Gebru was let go.

The summer of 2022 brought the next headache. Google employee Blake Lemoine claimed that one of their AI language models, LaMDA, had expressed itself in human-like ways. Had the technology gone too far? Lemoine was subsequently fired.

Seen in this light, the word “responsible” becomes more legible. Google has been at the forefront of these questions, but has also run into difficulties navigating the complicated territory between AI technology and ethics.

Google now wants to communicate that it hasn’t been slow — but rather has been taking responsibility for a more careful approach to progress in the field.

The timing, however, argues against this framing.

At the end of January, Google’s leadership declared a “code red” after ChatGPT’s immediate global success. Suddenly there was urgency — even Google’s co-founders Larry Page and Sergey Brin were brought in to work on the AI strategy.

For a company that had claimed for seven years to prioritize AI above all else, everything now had to happen at once. It looks rather more like competition accelerating the product roadmap than any sudden resolution of ethical questions.

Around this time, AI heavyweight Dr. Geoffrey Hinton chose not only to leave the company but to publicly warn about the pace of AI development.

Positioning around a specific concept is a familiar strategy. Apple’s emphasis on privacy — “privacy” — has not gone unnoticed by anyone.

But Apple’s privacy commitment has genuinely dented competitors. It’s hard to see how Google’s “responsibility” framing could have the same effect.

Responsibility is also a relative concept. When have you taken enough? And does it extend so far that you would sacrifice good business for what others consider to be the right thing? These are the questions Google will face as AI development accelerates.

Being responsible is — and sounds — good. But it’s easier to say than to be. Especially when things are moving fast and the competition is breathing down your neck.

Bluesky is the internet’s latest hype — and social media challengers rarely break through

May 9, 2023
by Björn Jeffery
SvD Näringsliv

This analysis was first published in SvD Näringsliv, in Swedish, on May 9th, 2023. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.

Twitter challenger Bluesky is the internet’s latest hype. But don’t count on it succeeding. There are good reasons why new social media platforms rarely manage to scale.

Anyone who has ever been to a nightclub is familiar with the concept of “the queue.”

The logic is simple and well-tested. You keep the line outside deliberately long to create the impression of popularity. If that many people are waiting to get in, surely it can’t be bad? Until you finally make it inside and discover the place is empty — more guests on the outside than the in.

That, roughly speaking, has been the marketing strategy for most new social media services that have launched. Remember the audio app Clubhouse? That’s exactly what they did. And now it’s happening again: over a million eager users queuing for access to Bluesky, the new Twitter challenger.

In the wake of Elon Musk’s takeover of Twitter, a slew of alternatives to the text-based social network appeared. The technically complex Mastodon, Donald Trump’s Truth Social, and Bluesky — which comes from Twitter co-founder Jack Dorsey himself, launched during his own time as Twitter’s CEO, no less.

A charitable interpretation would see this as an attempt to correct the mistakes Dorsey made with Twitter the first time around. A more critical reading would point out that it’s precisely all the similarities to Twitter that are the problem.

Because Bluesky is — intentionally, of course — very similar to Twitter. It says “reposted” instead of “retweeted,” but an untrained eye would barely be able to tell the two services apart.

Those similarities also mean that the same types of challenges are likely to arise.

The spread of misinformation, harassment, and automated bots have been difficult problems for Twitter to solve for over a decade. Even knowing in advance that these issues will emerge, there are no obvious solutions to them.

Like Clubhouse — which rose like the sun and set like a stone — Bluesky will likely struggle to maintain the atmosphere that exists with around 50,000 users once that number grows tenfold or a hundredfold.

A large part of the challenge in the social media category lies precisely in scale.

Imagine hosting a dinner party for a handful of guests. It’s pleasant but fairly predictable. You know who’s coming and you know, more or less, how they’ll behave. Now compare that to running a restaurant with 300 covers. You can prepare, of course, but things will happen that you didn’t anticipate.

That’s why it’s misleading to look at the temporary success of small social media challengers. Succeeding in the genre with a small number of users is an entirely different kind of challenge.

Scale also attracts actors with intentions other than making new friends or posting cat videos. The state-sponsored disinformation campaigns that have made headlines arise wherever there are large numbers of people to influence.

Is there nothing genuinely new about these challengers? Bluesky and the similar Mastodon both emphasize one aspect that differs from Twitter: control over the platform and ownership of your own content.

Bluesky is decentralized, meaning each user can take their content with them and switch to a different provider. This sidesteps the monolithic — and, since Musk took over, constantly shifting — rulebook that Twitter operates under. If the rules don’t suit you, you can move elsewhere, or start your own version.

Here, however, we encounter another truth about social networks. Users choose the path of least resistance. You can configure extremely specific settings on Facebook that control exactly which friends see which content. But how many people actually do that?

Being able to float freely in a decentralized world is theoretically possible — but the vast majority of users will never even understand why they might want to.

Bluesky does one thing very well. Like the nightclub with the long queue, it creates a feeling of wanting to belong. The demand is so intense that invitations are being sold on eBay. But once you’re inside, you’ll likely discover what always turns out to be true: it looked better from the outside than it actually is.

Nvidia is winning the generative AI gold rush

May 8, 2023
by Björn Jeffery
SvD Näringsliv

This analysis was first published in SvD Näringsliv, in Swedish, on May 8th, 2023. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.

Behind the hype — and the anxiety — surrounding next-generation AI lies a simple question. Who controls the essential hardware? The answer has geopolitical implications.

Should you mine for gold yourself, or sell pickaxes and shovels to those who do?

The question from the gold rush era keeps returning to California. What was once a literal business opportunity in 1848 has become a metaphor for a particular kind of business model — one that enables other companies to succeed.

Now the next gold rush appears to be starting: the explosion of artificial intelligence. And it has focused attention on a specific type of computer chip required to perform the most advanced calculations. Put simply, without the right chip, large-scale AI doesn’t happen.

At the center of this accelerating trend are companies that were previously consigned to being under the hood — invisible, but entirely essential. No company has benefited more from this moment than American chipmaker Nvidia.

Nvidia’s involvement in the AI segment started early. Geoffrey Hinton — often called the “godfather of AI” — built a product back in 2012 that performed advanced image recognition: software that can understand what a photograph depicts. The results were seen as a major breakthrough in AI, and underpin products used by millions of people today.

According to Hinton, it would not have been possible without Nvidia’s chips. Those chips weren’t originally designed for this purpose — they were primarily used for video games. But their architecture turned out to be particularly well suited to this new application as well.

Looking at the most discussed AI product today, ChatGPT, it is believed to have been trained using 10,000 GPU chips from Nvidia. When Sundar Pichai, CEO of Google’s parent company Alphabet, recently presented his quarterly results, he explicitly cited his company’s access to Nvidia chips as a competitive advantage.

The attention has translated into financial results. Nvidia’s share price has risen nearly 100 percent so far this year. Compare that to rivals Intel and Qualcomm, which are up 11 and 8 percent respectively.

The enormous demand for this type of chip is now creating complications.

Early in the pandemic, there was a major shortage of a different kind of chip — semiconductors — which led to delays in everything from cars to consumer electronics. That shortage has since been resolved, partly due to lower demand.

The constraints on AI-grade chips stem from the fact that their manufacture is extremely advanced, and only a handful of actors globally can produce them. Nvidia doesn’t fabricate its own chips; it uses suppliers for that. Among them is Taiwanese company TSMC — also the world’s largest semiconductor manufacturer.

Securing access to the right type of chips is becoming a geopolitical question.

Today, China spends more money importing various types of chips than it does importing oil, according to the book Chip War by Chris Miller, professor of history at Tufts University. China is also attempting to circumvent the American blockade on certain chips by renting access rather than importing them outright.

Both the US and EU have launched initiatives to bring more chip manufacturing to their respective regions. The European Chips Act is one such effort — committing roughly 490 billion kronor to increase Europe’s share of global chip production from 10 to 20 percent. The American equivalent, the CHIPS and Science Act, is similar in ambition and scale.

While manufacturing is concentrated among a handful of global players, chip design — how the chip actually functions — is becoming an increasingly important competitive differentiator. Apple is the clearest example of a company that has shifted strategy, moving away from off-the-shelf chips in favor of custom-built variants engineered for specific functionality. Apple’s new chips are notable in particular for their ability to perform AI calculations on-device.

Amid the explosion of AI products, there are more fundamental questions beneath the usual ones about ethics and existential risk, about who owns AI models and the concentration of power in the field.

Under all of that lies a very simple and practical question: who has access to the right physical chips? Because access to hardware may prove to be a decisive factor in how advanced an AI you can develop. It’s easy to understand why politicians, companies, and entire countries are starting to get nervous.

AI is the word of the quarter at Microsoft and Google

April 26, 2023
by Björn Jeffery
SvD Näringsliv

This analysis was first published in SvD Näringsliv, in Swedish, on April 26th, 2023. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.

While the AI race continues, the core businesses of Microsoft and Google are performing well. That points to billions more in new AI investments to come.

Something unusual happened to Sundar Pichai recently. The CEO of Alphabet — Google’s parent company — appeared on television. Specifically, he was a guest on the CBS interview program 60 Minutes. To talk about AI.

For an ordinary CEO, that might sound like a dream. Interviewed in prime time for the entire American public.

But the CEOs of tech giants are anything but ordinary. They prefer to operate in the shadows and avoid drawing too much attention to themselves and their enormously profitable companies.

The fact that Pichai agreed to appear on television therefore signals that something isn’t quite right.

The issue is that Google has fallen behind and now faces real competition in the hotly contested AI space.

Being anything other than market leader is a position Google is entirely unaccustomed to. After being spoiled with a search market share of over 90 percent, it seems unfamiliar to suddenly be neither the biggest nor the best.

The main competitor in AI, however, is very familiar.

It’s Microsoft — led by CEO Satya Nadella — that has successfully reinvented the company through its billion-dollar investments in OpenAI.

The two tech giants have long competed across several areas, including enterprise software and cloud services. But services like ChatGPT have made the world question the future of the lucrative search market in a way that hasn’t happened before.

It was therefore fitting that both rivals reported their quarterly results at the same time. On Tuesday evening they both surprised with strong results, revealing two companies ready for a continuing and intense contest.

Google beat analysts’ modest expectations on both revenue and profitability. Particularly notable was Google Cloud turning profitable for the first time. Revenue growth was a modest 3 percent year-over-year, however, and it was the third quarter since Google’s 2004 IPO in which revenue had declined.

Microsoft countered with something similar. They also beat expectations, with stalwarts like Office for business growing 14 percent. Their newer cloud services, including Azure, grew 27 percent — lower than previous quarters, but at the top end of their own guidance.

The fact that both companies’ core businesses remain strong argues for even larger investments ahead — particularly in AI. That much was clear when the abbreviation “AI” was mentioned an almost parodically large number of times by both CEOs in their briefings to press and analysts.

Compare this to Meta, for example, where weak growth has forced them to focus on “efficiency” — a euphemism for laying off tens of thousands of employees. Doubling down on the much-discussed metaverse is harder to justify when the core business is struggling.

Neither Google nor Microsoft will have that problem.

On the contrary, Microsoft has more wind in its sails than it has had in a long time. When it also turns out that the company’s existing products haven’t suffered — and in some cases have directly benefited — from the AI focus, continued investment becomes a foregone conclusion.

Microsoft has also managed to demonstrate what a more practical application of AI could look like. GitHub, which Microsoft acquired in 2018, has for example launched an AI product called “Copilot” that can assist software development — essentially an AI assistant that helps you write better code.

For Google, the task is to catch up. They were — and are — one of the world’s leading companies in AI. They are now reorganizing to move faster.

This is Sundar Pichai’s single greatest challenge. His pay last year — roughly 230 million kronor — suggests the expectations are high.

Google cannot afford to lose this battle, and a failure here would likely mean Pichai’s departure. His television appearance last week suggests he is taking the challenge seriously. Tuesday’s quarterly numbers mean he’ll have the resources to act on it. Whether he can deliver remains to be seen.

Alphabet beat expectations in its first-quarter results, with revenue of $69.8 billion against an estimated $68.9 billion. Earnings per share came in at $1.17, versus the Wall Street analyst consensus of $1.07. Microsoft also exceeded expectations in its third-quarter results, posting earnings of $2.45 per share against an anticipated $2.23. Revenue reached $52.9 billion — a 7 percent increase year-over-year, above the $51.0 billion analyst consensus.

The ‘Alecta effect’ is threatening Swedish tech companies

April 21, 2023
by Björn Jeffery
SvD Näringsliv

This analysis was first published in SvD Näringsliv, in Swedish, on April 21st, 2023. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.

After Alecta’s billion-kronor blunder, Sweden’s institutional capital is now focused on one question: how do we avoid the same thing? The answer risks prolonging the tech winter considerably.

A lot can happen in two years.

In 2021, the tech market was booming. The list of companies eyeing a stock market listing was long.

Alecta, together with other institutional investors, was actively investing in what’s known as the “pre-IPO” phase — companies not yet listed but intending to reach the market within 18 to 24 months.

Fast-forward 24 months and the picture couldn’t be more different.

Anxiety, inflation and rising interest rates sent tech stocks tumbling. And Alecta found itself in serious trouble after losing billions in the Silicon Valley Bank collapse.

The criticism of the pension company was fierce, and Alecta’s CEO Magnus Billing was fired.

Even though the investments in question were in American bank stocks, there is now a real risk that the large tech investments in Sweden will dry up entirely — at precisely the moment they are needed most.

As the tech sector has fallen sharply on the stock market, companies at earlier stages have also begun to feel the effects. Valuations have dropped, which can create a trap for companies that haven’t managed or haven’t yet been able to transition to profitability. The industry now talks of significantly longer fundraising processes than before, and tougher terms. For many companies, cash is running low, profitability is a distant prospect, and raising capital at decent valuations is harder than ever.

Meanwhile, the prospect of a stock market listing is more remote than it has been in years. For a tech company to go public in 2023 is widely considered a non-starter. Optimists sense a possible window opening sometime next year. Pessimists aren’t convinced about that timeline either.

In many ways, this is a golden moment for institutional capital. Those investors are sitting on well-stocked war chests and can access tech companies at far more attractive valuations than they’ve been able to achieve in recent years. Competition has cooled, valuations are down, and the need is great. Could the setup be better?

In theory, the moment is perfect. But in practice, the tech sector may have been contaminated by anxiety around Alecta’s banking losses. The association is right there in the name — Silicon Valley Bank. A bank with direct exposure to the world’s most dynamic region for startups and tech innovation.

Tech stocks carry high risk by nature. In a low-interest-rate world, investors went looking for returns wherever they could find them. The macro environment is now almost the opposite: high inflation, rising rates, and a war in Europe. Very little favors a high appetite for risk in this segment.

The question therefore becomes: how much tech-related risk is Swedish institutional capital willing to take on right now? The probability that Alecta would make further large tech investments this year has to be close to zero. Industry news that Kinnevik’s holding, food delivery company Mathem, has lost nearly 90 percent of its valuation since 2021 doesn’t help either.

Given Alecta’s outcome, attention now shifts to players like the AP funds, AMF, and Swedbank Robur. Will they take advantage of the market opportunity? With the risk that a handful of bad bets makes their own leadership team the next Magnus Billing? If one were to hazard a guess, most will sit tight and wait out this wave.

The “Alecta effect” may therefore have effectively shut down institutional tech financing for late-stage companies — at least for the rest of this year. Possibly longer. The irony is not lost: valuations haven’t been this low in over a decade.

A peculiar deal that raises many questions — Yubico’s SPAC listing

April 19, 2023
by Björn Jeffery
SvD Näringsliv

This analysis was first published in SvD Näringsliv, in Swedish, on April 19th, 2023. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.

When Bure’s SPAC vehicle ACQ wants to take security company Yubico public, it raises many questions. The main one: why are the buyers and sellers largely the same people from the same firms?

At Wednesday morning’s press conference, Patrik Tigerschiöld, chairman of ACQ Bure, posed the question that every ACQ shareholder ought to be asking.

“And then perhaps you ask yourself — ACQ has been looking for a company to acquire for two years, and ends up buying a company that Bure already owns a stake in. This fact has been a challenge for me and the entire team.”

Yes, that does sound like a challenge. A communications challenge in particular. How do you spend two years searching for something your own parent company already had in its portfolio?

No such explanation was forthcoming, unfortunately. And it leaves a long list of questions unanswered in the wake of today’s announcement about security company Yubico — founded by Stina Ehrensvärd — which may soon be listed on Sweden’s Nasdaq First North Growth Market.

Investing in a SPAC is a kind of bet more than anything else. The idea is that the SPAC’s management will scour the market for strong listing candidates, evaluate them, and ultimately execute a merger that results in a fast and efficient public debut for a promising company. That’s how it’s meant to work in theory.

In practice, it’s worth noting that there was a considerably cheaper and faster way to get Yubico to market. To understand why, you only need to look at the names on each side of the deal.

Among Yubico’s largest sellers: Bure (via Bure Growth), AMF Tjänstepension, and AMF Fonder.

Among ACQ Growth’s largest owners: Bure, AMF Tjänstepension, and AMF Fonder.

The conflict of interest is obvious, and it has led several board members to recuse themselves from the decisions. Valuation is another problem. How do you price a company when the buyer and seller are largely the same people?

The proposed structure solves a headache for ACQ Bure, whose SPAC format has completely fallen out of favor with investors. In 2021, 791 new SPAC vehicles were created worldwide. In February this year, that number was 10. The SPAC concept is currently dead.

But for SPACs that are already listed, the obligation remains: find a merger partner or return the money to investors. ACQ Bure had until March 2024 to find a suitable company. That problem is now solved.

Even for technically listed SPAC companies, the current market backdrop can’t be ignored. Over the past year, the tech-heavy American Bessemer Cloud Index has fallen 24 percent. In Sweden there are many examples of tech companies down far more — sometimes over 80 percent.

When SvD asked management directly about the tech market environment, they acknowledged it had been poor both last year and this year. But according to Bure CEO Henrik Blomkvist, Yubico is something “special and unique.”

There is indeed something unique about this deal. But it isn’t only the impressive technology that Swedish-American Yubico has developed over its 16 years as a company, or the fact that it has helped set the global standard for cybersecurity — an impressive achievement by any measure.

What may be rather more unique and special here is how majority owners Bure and AMF have handled their responsibilities toward other shareholders and pension savers. It took them two years and significant costs to identify a company they already owned — in order to then list it in the toughest tech market since 2008. That does sound uniquely special, in its own way.

Note: Yubico is one of the nominees for SvD Affärsbragd 2023.

Apple’s savings account could be just the beginning for big banks

April 19, 2023
by Björn Jeffery
SvD Näringsliv

This analysis was first published in SvD Näringsliv, in Swedish, on April 19th, 2023. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.

Apple’s latest product isn’t a gadget — it’s a savings account. But the real threat to traditional banks is what the tech giant might do next. Because Apple, as usual, has several aces up its sleeve.

In a dimly lit, elegant restaurant in Palo Alto, California — at the heart of Silicon Valley — a major Swedish bank’s board and senior leadership are gathered. It’s the early 2010s, and the bank’s executives are in town to be inspired by digital technology and learn from it. The invited guests are local entrepreneurs with some connection to Sweden. I’m one of them.

The conversation turns to challengers in the banking and finance world. Apps are demonstrated, potential disruptors described. The bank’s board is not particularly impressed.

They reference a recent customer satisfaction survey. Customers are happy — therefore competition from new players is not a problem.

That’s the usual script. Among the headlines about “new challengers,” very few in practice actually challenge large and established players. Especially in heavily regulated areas like banking and finance.

Even Klarna — themselves a sort of challenger — recently dismissed the threat from Apple’s relatively new “buy now, pay later” feature. Klarna has many customer offerings and is established across its markets. There’s therefore no reason to worry, you could infer from the interview with CEO Sebastian Siemiatkowski.

When Apple launches a savings account in its mobile app for American customers, Swedish banks are therefore unlikely to be particularly concerned — despite the generous interest rate of 4.15%. They’re used to being able to write off new competitors as ambitious but irrelevant.

A savings account here or there probably doesn’t make much difference.

But it would be a mistake to underestimate what a company like Apple can do when it has decided to take on an industry.

Apple’s financial services are built in partnership with investment bank Goldman Sachs — which has itself had difficulties breaking into the more consumer-facing side of banking.

In January, Goldman Sachs announced that its consumer division had lost around 30 billion kronor since launch in December 2020. Given this, the relative calm among big banks is understandable — even a major bank from a different part of the financial sector appears to find it hard to compete on their home turf.

But Apple is not like other companies. Even setting aside its capacity for innovation, it has an asset that is difficult to compete with: a great deal of money.

A very great deal, in fact.

At the start of the year, Apple held over 210 billion kronor in liquid assets — plus around 350 billion kronor in securities. That’s roughly equal to the combined market capitalization of Handelsbanken, Swedbank and Danske Bank. There are resources to work with.

There is also a strategic ambition for growth, with Apple having identified financial services as one of its chosen areas. iPhone sales are stable but no longer growing the way they once did. Building out an ecosystem of software services to complement the hardware has therefore been CEO Tim Cook’s focus in recent years.

Apple Pay — the ability to pay by tapping your phone, with no physical card needed — is popular with users. It could be seen as something of a Trojan horse in this context. Customers are already paying with their iPhone. Which card handles the underlying transaction becomes, in that way, rather less important.

Apple also has something that the big banks neither have nor can acquire quickly — a brand that people actually like.

That became clear when the company finally opened the doors to its new store in Mumbai, India, drawing enormous crowds.

It’s also visible in surveys. Consulting firm Interbrand publishes an annual ranking of the world’s most valuable brands. Top of the list in 2022? Apple. The first financial name doesn’t appear until position 24, with bank JP Morgan Chase.

If we set aside Apple’s savings account and instead focus on the direction of travel, a picture emerges that should be more troubling for the financial establishment.

Imagine Apple acquiring Goldman Sachs’s struggling consumer division. They already work together. Apple could then integrate banking services and payments directly into the operating systems running across more than 1.8 billion active devices in the market.

They could offer exclusive deals to everyone who has — or buys — a new iPhone. And they do it with a brand that stands far higher in public esteem than any of the big banks.

Last but not least — Apple can afford to do it.

In fact, they can afford to lose money on this for decades. Because they have what no other bank has — entirely different revenue streams.

Core parts of the big banks’ business would, in this scenario, become something of a side benefit for a company like Apple. That is something worth feeling threatened by.

Banning TikTok could become a geopolitical game of cat and mouse

April 17, 2023
by Björn Jeffery
SvD Näringsliv

This analysis was first published in SvD Näringsliv, in Swedish, on April 17th, 2023. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.

While American politicians try to ban TikTok, a similar app — with the same owner — is climbing the charts. The stage is set for a geopolitical game of cat and mouse.

“Mr. Chew, can TikTok access home wireless networks?”

Republican congressman Richard Hudson looked concerned as he questioned TikTok’s CEO, Shou Chew, at the end of March. Together with colleagues, he grilled Chew for five hours in an effort to better understand the popular social media app. The quality of the questions varied considerably, and often seemed more focused on looking good on television than on actually learning anything.

The backdrop to the hearing is the geopolitical tension between the US and China. TikTok first ran into trouble during the Trump presidency, when it came close to being forced into a partnership with American IT giant Oracle — to ensure that data from American users didn’t end up with the Chinese government. President Biden reversed that decision, but has since had to return to something resembling the original plan. After it emerged that TikTok had spied on journalists, the app is now under investigation and faces the prospect of being banned entirely from the American market.

The question to ask is: how much effect can banning a single app actually have?

In the US, an app called Lemon8 is now rapidly gaining popularity. It resembles TikTok in many respects and has received over 650,000 downloads in the past two weeks. The company behind it is called Heliophilia Pte Ltd — a name most people are unlikely to recognize. Its registered address, however, is more familiar: it’s identical to TikTok’s office in Singapore. Both apps share the same majority owner — Chinese conglomerate ByteDance.

As an ordinary mobile user, it can be difficult to know who actually owns the popular apps you use. In Sweden, for example, the video editing app CapCut sits high on the app charts. It is also owned by ByteDance. A few positions lower is FaceApp — the digital face filter app whose parent company is registered in Cyprus, but whose founder, Yaroslav Goncharov, previously sold companies to — and worked at — Yandex, often described as “the Russian Google.” The app was investigated by the FBI a few years ago over concerns about Russian counterintelligence.

Banning a single app may sound like a simple solution, but the situation is more complex than that. In some cases — as above — it requires genuine detective work to figure out who actually owns the apps in question.

China has made things easy for itself in this regard. For many years, it has blocked the largest American internet companies. Facebook, YouTube, Reddit and Pinterest are all on the list of products shut out of the country. China’s strong censorship means many of these sites fall foul of existing laws. It was therefore somewhat ironic when China protested to American authorities about the prospect of TikTok being kicked out of the US. There is no symmetry here.

Successfully regulating this messy situation is no easy task. The poor technological literacy of American lawmakers doesn’t help either. Congressional hearings with tech executives over the years have been near-parodically bad, and have produced no concrete legislation. Focusing a hearing on a single foreign app is one way to simplify the critique. Politicians — both Democrats and Republicans — appear to be saying “we believe China can spy on us via TikTok,” and have through this managed to find common ground. The proposed legislation, however, doesn’t name TikTok specifically — it would enable bans on technology services from any country the US has designated as an adversary. A list that can change over time.

It is now quite possible that TikTok will be forced into a sale or IPO to shed its Chinese ownership. A full ban from the US market is less likely. But even if that happened, the problem wouldn’t be solved. Lose one app and a thousand more appear. Should every popular app undergo a national security review? That’s not sustainable in practice. New services and apps that could pose a risk emerge every day. But that reality doesn’t make for good television hearings.

Aimed for space — crashed on the stock exchange

April 5, 2023
by Björn Jeffery
SvD Näringsliv

This analysis was first published in SvD Näringsliv, in Swedish, on April 5th, 2023. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.

Virgin Orbit is filing for bankruptcy. The story of how the company was listed on the stock market is part of a larger tale about a financial instrument that created enormous expectations — and has now largely collapsed.

Richard Branson said that he started an airline to fund his dream of building a space company. The airline was Virgin Atlantic; the space company, Virgin Galactic.

Between them came Virgin Orbit — a rocket company whose innovation was launching satellites from a converted jumbo jet. The idea was that a plane could carry a rocket to altitude and release it midair, reducing the cost and infrastructure needed for launch.

In January, that mission failed. A rocket launched from a modified Boeing 747 over the Atlantic failed to reach orbit. The company now has around 85 employees left after mass layoffs, and has filed for bankruptcy.

The fact that Virgin Orbit ran out of money is partly explained by the January failure. But the underlying story involves a type of financial structure that became extremely fashionable and is now — just as rapidly — falling out of favor.

Virgin Orbit went public through a so-called SPAC — a Special Purpose Acquisition Company. A SPAC is a listed shell company with no operations of its own; its purpose is to find a real company to merge with, thereby taking that company public without a traditional IPO process. Virgin Orbit merged with a SPAC called NextGen Acquisition Corp in 2021, at a valuation of $3.7 billion.

For some time, SPACs were seen as a smarter and faster path to the stock market. The SPAC raises capital first, then hunts for a target — effectively a blank check with a deadline. High-growth companies that might not yet meet the requirements for a traditional IPO could instead merge with a SPAC and arrive on the market with relatively little scrutiny.

If SPACs sound like a phenomenon from an era of cheap money, there’s a simple reason for that. Back in 2020 — when tech stocks were still sky-high — analyst Byrne Hobart called the entire category a “call option on hype.” In a low-interest-rate environment where everyone was chasing yield, SPACs — even without knowing which company they would eventually merge with — were exactly that: an option on an opportunity that would materialize somewhere down the line. In 2021, 791 SPAC vehicles were listed. In February this year, the equivalent number was 10.

The problem with SPACs, however, isn’t just the growth — it’s that the promised future arrives on a tight schedule. Most SPACs must find an acquisition target within 18 to 24 months of listing. If they fail, the vehicle can be liquidated and investors repaid. So far this year, 70 SPACs have been forced to do exactly that. That list will grow longer as the 24-month windows from 2021 come due.

Among the companies that did manage to find a partner and reach the stock market, the next blow came anyway. In 2022, tech stocks crashed across global markets, and many newly listed SPAC companies fell with them. News site BuzzFeed, for example, has dropped a full 80 percent in a year. It is far from alone. Electric vehicle company Faraday Future is down 93 percent.

In an era of both high inflation and high interest rates, SPACs are unlikely to make a comeback. Neither the structure itself nor the results of those that went through with it show much success to point to. The SPAC chapter can therefore be considered closed for this cycle.

But before it’s fully shut, there’s still some clearing up to do. The hundreds of companies that have already listed still need to either complete a deal or return the money to investors. One SPAC — unwilling to pay its bills — is now being sued by its law firm after the deal fell through. Expect a similar wave of claims from contractors, employees, and investors.

Listing quickly and easily sounded too good to be true. Perhaps it was. And perhaps there was a very simple reason why these companies weren’t publicly listed in the first place. They just weren’t ready for public markets. Not yet — and in some cases, not ever.

There is an urgent problem with AI — and it isn’t what the open letter says

April 4, 2023
by Björn Jeffery
SvD Näringsliv

This analysis was first published in SvD Näringsliv, in Swedish, on April 4th, 2023. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.

When 50,000 researchers and business leaders warn against AI development, the uproar is enormous. But the truly serious problem isn’t the development itself — it’s that a handful of companies hold all the power.

“Humanity swallowed by its machines — body, mind and soul — and civilization poisoned to its encroaching death.”

The illustration showed a person being sucked into their own sausage-making machine.

That was the New York Times in 1921.

The website Pessimists Archive has compiled examples from every decade since, dystopian visions of how machines, robots and computers will replace us and take all our jobs.

Now it’s 2023. This time it’s AI that will obliterate the workforce and destroy the world. Plus ça change.

Few things seem to unsettle a society quite like a technological shift. The sudden availability of new, well-developed AI language models has unleashed an explosion of creativity, entrepreneurship — and anxiety.

When an open letter signed by 50,000 researchers and business figures calls for a thoughtful six-month pause in AI development, it adds more fuel to that anxiety.

The fact that signatories include serial entrepreneur Elon Musk and Apple co-founder Steve Wozniak amplifies it further. They want development of the next generation of language models — what would be called GPT-5 — to pause for six months while society considers the risks.

But a pause in development is unlikely to materialize, regardless of how many signatures are gathered.

The letter references a set of principles drawn up by AI enthusiasts at a conference in Northern California in 2017. Those principles were created as a form of self-regulation within a field that, at the time in particular, almost entirely lacked laws and structure.

Following principles is, as we know, voluntary. And without all AI developers simultaneously making the same voluntary decision — including Russia and China — we won’t have added any thoughtfulness to the process. We’ll most likely have done nothing more than hand a head start to those with arguably worse intentions.

The question of how this should be handled — and potentially regulated — is not uncomplicated. Sweden’s EU Council Presidency is a reminder that nascent legislation is emerging from Brussels to try to bring order to the field. The AI Act and the AI Liability Directive are two relevant initiatives, for instance. Sweden appears to have delegated the question there.

Looking at the EU’s history of regulating technology companies, however, there is reason to be skeptical. It took them over 20 years to create laws preventing monopolistic behavior among the largest tech companies.

When they set out to protect our data, we got GDPR — which for the average person has mostly resulted in an endless stream of cookie consent boxes on every website, and the disappearance of class lists from schools.

At a time when Europe essentially lacks any major tech companies on a par with Apple, Google and Meta, it’s hard to imagine that regulation will do more than worsen the continent’s ability to compete.

The most significant AI development is already happening outside the EU. Laws there therefore risk being a swing at thin air. It would be like Sweden imposing strict rules on viticulture — well-intentioned, but largely meaningless in the bigger picture.

If legislation isn’t sufficient and voluntary principles aren’t followed — what’s left? In these technophobic times, one seemingly radical idea would be: optimism.

Instead of only worrying about the end of the world, let us also consider how we can distribute these superpowers fairly and equitably across the world.

If AI development creates the productivity boom that is widely predicted, this is an excellent opportunity not to recreate the kind of de facto tech monopolies that the Western world and China live with today.

Emily M. Bender, one of the researchers whose paper the open letter references, says her conclusions have been misread. Rather than warning about a hypothetically dangerous AI future, she argues, the real and far greater risk is that too few people will be able to access its benefits. It is about “the concentration of power in the wrong hands,” she writes, among other things.

It is therefore of great value that we discuss the future of AI. But an optimistic and pragmatic version of that discussion would focus on how we ensure that as many people as possible can benefit from the productivity-enhancing capabilities being developed.

Do we want the world’s leading AI development to happen and be controlled by a handful of privately owned companies? As things stand, Microsoft (through its investment in OpenAI), Google and China’s Baidu are among those at the frontier.

They don’t just control how the models are built — they also control what data those models are trained on.

That is the really hard nut to crack.

The most important question, therefore, is not binary — whether we should pursue AI development at all. The question is rather how we ensure it is done in a way that benefits as many people as possible.