Start Up No.2237: China’s internet history vanishes, Kutcher foresees AI movies, Spotify’s AI music problem, and more


The stock market valuation for Nvidia has passed Microsoft to hit $3 trillion – in theory, the total of all the profits it will ever make. CC-licensed photo by Chris Yarzab on Flickr.

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.


There’s another post coming this week at the Social Warming Substack on Friday at 0845 UK time. Free signup.


A selection of 9 links for you. Graphically. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. Observations and links welcome.


Nvidia passes Microsoft to become the world’s most valuable company • The Washington Post

Gerrit De Vynck and Rachel Lerman:

»

Nvidia, the computer chip maker at the center of the artificial intelligence boom, continued its spectacular stock market rise Tuesday, eclipsing Microsoft and Apple to become the world’s most valuable public company.

Nvidia’s shares rose more than 3% Tuesday, giving the company an overall market valuation of $3.34 trillion. Apple and Microsoft, which for years have swapped positions as the world’s most valuable company, were worth slightly less — around $3.29 trillion and $3.31 trillion, respectively, at the end of trading.

Nvidia’s computer chips and software are crucial to training the AI algorithms behind image generators and chatbots like OpenAI’s ChatGPT. As the tech and business worlds throw themselves into the AI boom, demand for the chips has skyrocketed, pushing Nvidia’s revenue up to $26bn in the first quarter of this year, up from just $7.2bn a year ago.

The AI boom has been reshuffling the world’s biggest companies in the past two years. In January, Microsoft surged past Apple to become the world’s most valuable company as investors have poured money into the hot technology.

Nvidia, which was already a large company, has had the “most remarkable growth story,” in tech, said Gene Munster, managing partner at Deepwater Asset Management.

«

What’s notable here is Nvidia’s PER (price earnings ratio – the cost of its share relative to its earnings, ie how many years it would take for the share to earn back its price. (Nvidia does pay dividends.) Nvidia’s PER is 79.7 on its current price, and would get higher as its price rises. Microsoft’s: 38.7. Apple’s: 33.4. Google’s, back at a valuation of “only” 2.2 trillion, has a 27.0 PER.

Also, in theory the stock market valuation is the net present value of all the profits (not revenues!) the company will make in its lifetime. Which implies a lot for Nvidia. And Apple and Microsoft, of course. Google stock looks like a relatively good purchase. (This is not a recommendation.)

It’s going to be quite the spectacle seeing whether Nvidia can justify that lofty valuation. Always good to be selling shovels in a gold rush. But gold rushes end – though, that said, Nvidia has managed to jump from gaming to crypto to, now, the much much bigger goldmine of AI. Can it last?
unique link to this extract


As China’s internet disappears, ‘we lose parts of our collective memory’ • The New York Times

Li Yuan:

»

Chinese people know their country’s internet is different. There is no Google, YouTube, Facebook or Twitter. They use euphemisms online to communicate the things they are not supposed to mention. When their posts and accounts are censored, they accept it with resignation.

They live in a parallel online universe. They know it and even joke about it. Now they are discovering that, beneath a facade bustling with short videos, livestreaming and e-commerce, their internet — and collective online memory — is disappearing in chunks.

A post on WeChat on May 22 that was widely shared reported that nearly all information posted on Chinese news portals, blogs, forums, social media sites between 1995 and 2005 was no longer available. “The Chinese internet is collapsing at an accelerating pace,” the headline said. Predictably, the post itself was soon censored.

“We used to believe that the internet had a memory,” He Jiayan, a blogger who writes about successful businesspeople, wrote in the post. “But we didn’t realize that this memory is like that of a goldfish.”

It’s impossible to determine exactly how much and what content has disappeared. But I did a test. I used China’s top search engine, Baidu, to look up some of the examples cited in Mr. He’s post, focusing on about the same time frame between the mid-1990s and mid-2000s.

I started with Alibaba’s Jack Ma and Tencent’s Pony Ma, two of China’s most successful internet entrepreneurs, both of whom Mr. He had searched for. I also searched for Liu Chuanzhi, known as the godfather of Chinese entrepreneurs: He made headlines when his company, Lenovo, acquired IBM’s personal computer business in 2005.

…I got no results when I searched for Ma Yun, which is Jack Ma’s name in Chinese. I found three entries for Ma Huateng, which is Pony Ma’s name. A search for Liu Chuanzhi turned up seven entries.

«

It’s very 1984, memory hole and all.
unique link to this extract


Ashton Kutcher: AI will soon create full movies, raise bar in Hollywood • Variety

Ethan Shanfeld:

»

Ashton Kutcher looks at OpenAI’s generative video tool, Sora, as the future of filmmaking.

“I have a beta version of it and it’s pretty amazing,” Kutcher said of the platform in a recent conversation with former Google CEO Eric Schmidt at the Berggruen Salon in Los Angeles.

He added, “You can generate any footage that you want. You can create good 10, 15-second videos that look very real. It still makes mistakes. It still doesn’t quite understand physics. … But if you look at the generation of this that existed one year ago as compared to Sora, it’s leaps and bounds. In fact, there’s footage in it that I would say you could easily use in a major motion picture or a television show.”

“Why would you go out and shoot an establishing shot of a house in a television show when you could just create the establishing shot for $100? To go out and shoot it would cost you thousands of dollars,” Kutcher said. “Action scenes of me jumping off of this building, you don’t have to have a stunt person go do it, you could just go do it [with AI].”

Kutcher added that, while playing around with the software, he prompted Sora to create footage of a runner trying to escape a desert sandstorm. “I didn’t have to hire a CGI department to do it,” Kutcher said. “I, in five minutes, rendered a video of an ultramarathoner running across the desert being chased by a sandstorm. And it looks exactly like that.”

«

There isn’t really going to be any stopping this, is there – though at present it’s still, for me, at the level of dreams.
unique link to this extract


China’s nuclear power development is almost 15 years ahead of the US • Fast Company

Reuters:

»

The US is as much as 15 years behind China on developing high-tech nuclear power as Beijing’s state-backed technology approach and extensive financing give it the edge, a report said on Monday.

China has 27 nuclear reactors under construction with average construction timelines of about seven years, far faster than other countries, said the study by Information Technology & Innovation Foundation, a Washington-based nonpartisan research institute.

“China’s rapid deployment of ever-more modern nuclear power plants over time produces significant scale economies and learning-by-doing effects, and this suggests that Chinese enterprises will gain an advantage at incremental innovation in this sector going forward,” the report said.

The US has the world’s largest fleet of nuclear power plants and President Joe Biden’s administration considers the virtually emissions-free electricity source to be critical in curbing climate change.

But after two large plants in Georgia came online in 2023 and 2024 billions of dollars over budget and delayed by years, no US nuclear reactors are being built. A high-tech plant that had been planned to be built at a US lab was cancelled last year.

China’s state-owned banks can offer loans as low as 1.4%, far lower than available in Western economies.

«

Planning laws have become a weird hydra which (to mix metaphors) strangle these big projects at birth. China, meanwhile, has no such concerns and steamrolls over objections. One feels there should be a happier medium.
unique link to this extract


Microsoft postpones Windows Recall after major backlash — will launch Copilot+ PCs without headlining AI feature • Windows Central

Zac Bowden:

»

In an unprecedented move, Microsoft has announced that its big Copilot+ PC initiative that was unveiled last month will launch without its headlining “Windows Recall” AI feature next week on June 18. The feature, which captures snapshots of your screen every few seconds, was revealed to store sensitive user data in an unencrypted state, raising serious concerns among security researchers and experts. 

Last week, Microsoft addressed these concerns by announcing that it would make changes to Windows Recall to ensure the feature handles data securely on device. At that time, the company insisted that Windows Recall would launch alongside Copilot+ PCs on June 18, with an update being made available at launch to address the concerns with Windows Recall.

Now, Microsoft is saying Windows Recall will launch at a later date, beyond the general availability of Copilot+ PCs. This means these new devices will be missing their headlining AI feature at launch, as Windows Recall is now delayed indefinitely. The company says Windows Recall will be added in a future Windows update, but has not given a timeframe for when this will be.

The company does say that Windows Recall will be made available to test in the Windows Insider Program for users with Copilot+ PCs “in the coming weeks.”

«

You do have to wonder very, very hard about how this got greenlit internally. Truly, we all thought Microsoft had long since got past this sort of screwup, which feels like something from the Ballmer days.
unique link to this extract


Spotify has an AI music problem—but bots love it • WIRED

Amanda Hoover:

»

IF A SONG is created by artificial intelligence and listened to by a bot, was it even heard at all? It’s a problem music-streaming companies now face as generative AI is rapidly making it easier for anyone to churn out songs with a few clicks, and then send bots to stream them for cash. 

“It’s a floodgate,” says Tony Rigg, a lecturer in music industry management at the University of Central Lancashire in the UK. He’s talking about the arrival of AI-generated music. And that torrent of new music amplifies the issue of fake listening, giving people a simple way to get streams on low-quality tracks. 

Artificial streaming, or bot listening, isn’t new. Some turn to third-party companies promising to boost streams, which then enlist bot-made accounts to listen to the same playlists on repeat. It’s a problem because streaming companies divide up royalty payments from a limited pool of cash—the more a song plays, the more its creator earns. So, more money for songs listened to by bots can mean that less is sent to those with human fans. Human artists have already been caught up in artificial streaming scandals, but AI is adding a new element. 

The first major test case came last week when Spotify reportedly removed tens of thousands of songs created and uploaded to Spotify by AI music generator Boomy. These made up a small percentage of total Boomy creations, but did include songs that were suspected of being streamed by bots, according to the Financial Times. Spotify did not respond to a request for comment to confirm the removal, but the platform does have policies against fake streaming.

«

This is from a month or so back, but still relevant: how is Spotify going to keep ahead of the tsunami that has previously overwhelmed sci-fi book submission sites, and Amazon’s Kindle bookstore? One has to suspect that it won’t be able to unless it moves to a “trusted sources” model. Which reverts the music business back to what it was for a long time before.
unique link to this extract


How the fridge changed flavour • The New Yorker

Nicola Twilley:

»

In 2010, the open‐data activist Waldo Jaquith decided to make a cheeseburger from scratch, using only agrarian methods. He and his wife had just built a home in the woods of Virginia, where they raised chickens and tended to an extensive vegetable garden. Flush with pride in his self-sufficiency, Jaquith outlined the steps required: bake buns, mince beef, make cheese, harvest lettuce, tomatoes, and onion. Then he realized that he wasn’t nearly committed enough. To really make a cheeseburger from scratch, he would also need to plant, harvest, and grind his own wheat, and raise at least two cows, one for the dairy and another to be slaughtered for the meat.

At this point, Jaquith gave up. The problem wasn’t labor but timing. His tomatoes were in season in late summer, his lettuce ready to harvest in spring and fall. According to the seasonal, pre-refrigeration calendar he was trying to follow, Jaquith would have needed to make his cheese in the springtime, after his dairy cow had given birth: her calf would be slaughtered for the rennet, and the milk intended to feed it repurposed. But the cow that provided his beef wouldn’t be killed until the autumn, when the weather started to get cold. If Jaquith turned the tomatoes into ketchup and aged his cheese in a cellar for six months, until the meat, lettuce, and wheat bun were ready, he could maybe, possibly, make a cheeseburger from scratch. But practically speaking, he concluded, “the cheeseburger couldn’t have existed until nearly a century ago.”

And, in fact, it did not. The cheeseburger is just one of many sensory pleasures made possible by a highly industrialized and refrigerated food system. More obvious ones include the delightful anticipation of pouring a crisp beer at the end of the day, the refreshing clink of ice cubes in a soft drink or a cocktail, and, of course, the joy of licking an ice-cream cone in summer.

«

The fact that we aren’t really self-sufficient any more, and would struggle if we had to be, only emphasises how technology (here, the fridge) has really changed our lives immeasurably over the past millennia.
unique link to this extract


The Substack boom is creating parasocial relationships right in our email inbox • Slate

Sonia Weiser:

»

On one hand, what newsletter creator wouldn’t want their work to resonate with their audience so much that it drives them to reach out? On the other, how does a creator navigate the responsibilities and expectations of a parasocial relationship that formed without their buy-in? Is it now their obligation to not just provide customer service, but be a “good friend”?

“If someone takes the time to cross the digital void, the least I can do is respond,” Ryan Broderick, the creator of Garbage Day, a newsletter devoted to web culture and technology with over 70,000 subscribers, told me. He explains that even if someone writes something mean, the fact that they took the time to write anything at all, and write it privately, is an honor.

Writing an email “isn’t performative,” he continued. “I find the incentives of sending someone an email to be much different than the incentives of tweeting or making a TikTok about someone, because they’re not doing it for any attention other than for your attention.”

Once maligned as teenybopper, kiss-your-Elvis-photo-goodnight foolishness, parasocial relationships have gained credibility as a valid psychological phenomenon. The first time I heard about them was after news of John Mulaney’s divorce hit the internet. On a date soon after, the guy I was with railed against the comedian’s behavior as if he, a social worker in Brooklyn, had a personal stake in the matter.

«

This is not, she suggests, always that good for the newsletter author: the parasocial effect is that people feel that because you’ve been welcome in their inbox for ages, therefore you should welcome them when they rock up.
unique link to this extract


The Microsoft Excel superstars throw down in Vegas • The Verge

David Pierce competed (not really) in the Excel World Championship:

»

At 6PM on the dot, Andrew Grigolyunovich, the founder and CEO of the Financial Modeling World Cup, the organization hosting these championships, takes the modular stage in the ballroom. He loads an unlisted YouTube link, which begins explaining today’s challenge, known as a “case.” It’s a puzzle called “Potions Master,” and it goes roughly like this: You’re training to be a potions master in Excelburg, but you’re terrible at it. You have a number of ingredients, each of which has a certain number of associated points; your goal is to get the most points in each potion before it explodes, which it does based on how much of a white ingredient you’ve added.

The Potions Master case, like so many of the puzzles conquered by these competitive Excelers, is not particularly complicated. This is a flashier, faster, deliberately more arcade-y version of spreadsheeting, more like trying to win 10 simultaneous games of chess on easy mode rather than painstakingly taking on a grandmaster. If you like, you can solve the whole thing manually: figure out when the white number gets too high, count the total points until that spot, then double-check it because it’s a lot of numbers, and eventually answer the first question. That’s my strategy, and I think I get it right. Now there are 119 more, worth a total of 1,500 points, and it’s quickly clear I’m not going to finish in the 30 minutes we’ve been allotted.

While I’m squinting into my 13-inch screen and carefully adding 1s and 3s, the other 26 contestants are whirring through their spreadsheets, using Excel’s built-in formula and data visualization tools to organize and query all that data. Everyone in the room seems to have their own way to chew through the ingredient lists and spends the first few minutes turning a mess of numbers and letters into real, proper capital-d Data. They start answering questions a half-dozen at a time, while I’m still checking my mental math.

«

Wonderful piece of colour writing. Everything’s a game. Even Excel.
unique link to this extract


• Why do social networks drive us a little mad?
• Why does angry content seem to dominate what we see?
• How much of a role do algorithms play in affecting what we see and do online?
• What can we do about it?
• Did Facebook have any inkling of what was coming in Myanmar in 2016?

Read Social Warming, my latest book, and find answers – and more.


Errata, corrigenda and ai no corrida: none notified

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.