Start Up No.2048: what to learn from AI writing, the impossible heatwaves, superconductor claims go into retreat, and more


Sales of the Oculus Quest VR headset helped Meta’s Reality Labs to $276m in revenue last quarter. But not a profit. CC-licensed photo by Rémy Metalograms 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 about 0845 UK time. Free signup.


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


What AI teaches us about good writing • NOEMA

Laura Hartenberger:

»

Writing that consistently adheres to convention is effective because its predictability makes it easy to read. If you expect to find the main point of a paragraph in its opening, you can read faster than if you had to spend time hunting for it.

But simply abiding by the rules doesn’t make excellent writing — it makes conventional, unremarkable writing, the kind usually found in business reports, policy memos and research articles. In his review of AI-generated novel “Death of an Author,” Dwight Garner describes the prose as having “the crabwise gait of a Wikipedia entry.” Even when a user prompts ChatGPT to include specific grammatical errors or to stray from certain norms, its writing tends to carry a certain flatness. By design, the program relapses to a rhetorical median, its deviations mechanical whereas ours are organic.

That’s not to say that convention flattens prose. In fact, convention lies at the root of much of the best writing — it’s rare to see acclaimed texts that stray dramatically from grammatical and stylistic norms.

Structural convention also underlies much of what we call good writing. Most prize-winning literature innovates within classic story arcs: Aristotle’s three-act structure (beginning, middle and end); Freytag’s five-stage structure (exposition, rising action, climax, falling action and resolution); or a screenwriter’s six categories of dramatic conflict (conflict with self; with others; with society; with nature; with the supernatural; and with the machine).

Indeed, the fact that AI, which is trained to detect and replicate underlying patterns in our writing, can produce such coherent prose is a testament to just how much we rely on convention, both at the sentence and structural level.

«

Just see if you could ever prompt a chatbot to come up with such a perfect phrase as “the crabwise gait of a Wikipedia entry”.
unique link to this extract


Europe and US heatwaves near ‘impossible’ without climate change • BBC News

Georgina Rannard:

»

The heatwaves battering Europe and the US in July would have been “virtually impossible” without human-induced climate change, a scientific study says.

Global warming from burning fossil fuels also made the heatwave affecting parts of China 50 times more likely.

Climate change meant the heatwave in southern Europe was 2.5C hotter, the study finds.

Almost all societies remain unprepared for deadly extreme heat, experts warn.

The study’s authors say its findings highlight the importance of the world adapting to higher temperatures because they are no longer “rare”.

“Heat is among the deadliest types of disaster,” says Julie Arrighi from the Red Cross Red Crescent Climate Centre, and also one of the authors.

Countries must build heat-resistant homes, create “cool centres” for people to find shelter, and find ways to cool cities including planting more trees, she says.

In July, temperature records were broken in parts of China, the southern US and Spain. Millions of people spent days under red alerts for extreme heat.

Experts say extreme heat can be a very serious threat to life, especially among the elderly. According to one study, more than 61,000 people were estimated to have died from heat-related causes during last year’s heatwaves in Europe.

«

Deniers/minimalists retort that more people die from cold. To which the answer is better insulation, not stuffing the atmosphere with carbon.
unique link to this extract


Uncharted territory: do AI girlfriend apps promote unhealthy expectations for human relationships? • The Guardian

Josh Taylor:

»

When you sign up for the Eva AI app, it prompts you to create the “perfect partner”, giving you options like “hot, funny, bold”, “shy, modest, considerate” or “smart, strict, rational”. It will also ask if you want to opt in to sending explicit messages and photos.

“Creating a perfect partner that you control and meets your every need is really frightening,” said Tara Hunter, the acting CEO for Full Stop Australia, which supports victims of domestic or family violence. “Given what we know already that the drivers of gender-based violence are those ingrained cultural beliefs that men can control women, that is really problematic.”

Dr Belinda Barnet, a senior lecturer in media at Swinburne University, said the apps cater to a need, but, as with much AI, it will depend on what rules guide the system and how it is trained.

“It’s completely unknown what the effects are,” Barnet said. “With respect to relationship apps and AI, you can see that it fits a really profound social need [but] I think we need more regulation, particularly around how these systems are trained.”

Having a relationship with an AI whose functions are set at the whim of a company also has its drawbacks. Replika’s parent company Luka Inc faced a backlash from users earlier this year when the company hastily removed erotic roleplay functions, a move which many of the company’s users found akin to gutting the Rep’s personality.

Users on the subreddit compared the change to the grief felt at the death of a friend. The moderator on the subreddit noted users were feeling “anger, grief, anxiety, despair, depression, [and] sadness” at the news.

«

unique link to this extract


Meta’s Reality Labs has lost more than $21 billion since start of 2022 • CNBC

Jonathan Vanian:

»

Meta reported second-quarter earnings on Wednesday and said that its Reality Labs unit, which develops virtual reality and augmented reality technologies needed to power the metaverse, logged a $3.7bn operating loss.

The unit recorded $276m in second-quarter sales, down from the $339m in revenue it brought during the first quarter. Analysts polled by StreetAccount were projecting Reality Labs to record $421m in sales off $3.5bn in operating losses.

Shares of Meta were up about 5% after it reported an 11% pop in revenue as advertising rebounded and the company issued an uplifting sales forecast for the third quarter. It shows that Meta is still very much an ad company with a big cost center.

Last year, Meta’s Reality Labs unit lost a total of $13.7bn while bringing in $2.16bn in revenue, which is driven in part by the company’s sales of Quest-branded VR headsets. Reality Labs lost $3.99bn during the first quarter. That puts its total losses at about $21.3bn since the beginning of last year.

Meta said in its earnings report that it expects operating losses in its Reality Labs unit “to increase meaningfully year-over-year due to our ongoing product development efforts in augmented reality/virtual reality and investments to further scale our ecosystem.”

«

Meanwhile Apple’s going to sell its headset in limited numbers, for $3,500 each, making a profit on each one, and move towards making the Vision division (aha) profitable. Spending so far on the Vision Pro is surely into the billions down the years. But would it be close to $21bn? Facebook/Meta might want to consider how it’s spending its money.

That revenue from Quest VR headsets, though: if all the $276m revenue is from selling $300 headsets, that’s about 0.9m sold. Tolerable? But depends a lot on what “driven in part” means about revenue.
unique link to this extract


Embattled physicist files patent for unprecedented ambient superconductor • Science

Robert Service:

»

Ranga Dias, a physicist at the University of Rochester, has drawn headlines and controversy for his claims of concocting materials that superconduct at room temperature—despite the limitation that they would require extreme pressures to work. His latest creation would be by far his most sensational yet—although he has not sought any attention for it. In a little-noticed patent filing, Dias claims to have made a material that superconducts not only at room temperature, but also ambient pressure.

If true, the discovery would be profound, igniting a host of applications, such as transmission lines that conduct electricity without losses, hyperefficient computer chips, and cheaper levitating trains. “We cannot even imagine how impactful it would be,” says Eva Zurek, a superconductivity theorist at the University at Buffalo. Such a material would also force a major rethink of the physics at play, as current theories cannot account for superconductivity under fully ambient conditions.

But scientists who have been critical of Dias’s data and methods in his previous claims don’t believe his latest work either. “I’m highly skeptical,” says James Hamlin, a superconductivity researcher at the University of Florida, who believes any superconducting behavior reported in the patent filing could be the result of broken electrical contacts in the devices used to characterize the material.

«

All superconductivity, all the time.
unique link to this extract


The latest mega-breakthrough on room-temperature superconductors is probably nonsense

Stuart Ritchie:

»

If we were to invent a superconductor that didn’t require these low temperatures and high pressures, all bets would be off: we’d have ultra-efficient power grids, saving vast amounts of energy; we’d have cheaper medical imaging; we’d have more low-friction, super-fast “maglev” trains that float above the tracks; and we might even be able to develop ultra-fast quantum computers. A room-temperature, ambient-pressure superconductor has long been something of a Holy Grail for scientists.

But this is a scientific field that’s seen many false dawns, and has seen itself descend into serious controversy. Just yesterday, it was reported that the journal Physical Review Letters was planning to retract a paper by a physicist who’d already had a paper on room-temperature superconductors retracted from the journal Nature in 2021. In the Nature case, a crucial analysis was found to be seriously flawed; in the newer case (which wasn’t specifically focused on superconductors), it appears that some data had been falsified or fabricated.

…The i contacted several of the UK’s top superconductor physicists and asked for their opinion on the new studies [from South Korea, linked here yesterday].

Professors Susannah Speller and Chris Grovenor, of the University of Oxford’s Department of Materials, jointly told i that the reports were “interesting, but not wholly convincing”. If the material really was a room-temperature superconductor, they argued, there are features one would expect to see in the data. One thing that’s clear is that there’s a “sharp drop” in resistivity, which is consistent with superconductivity. But we should also observe the material’s magnetisation changing (for the Meissner effect-related reasons discussed above), and its specific heat capacity, too. “Neither of these features,” say professors Speller and Grovenor, “is evident in the data presented.”

“It is too early to say that we have been presented with compelling evidence for superconductivity in these samples,” they argued.

«

Wait for the replication attempts. As the Korean researchers included a second paper on essentially how to make their material, it shouldn’t take long. Derek Lowe in Science, who follows the topic, reckons furnaces will have been busy already making samples of LK-99, as it’s known.
unique link to this extract


The Westminster-Big Tech revolving door keeps spinning • POLITICO

Tom Bristow and Laurie Clarke:

»

Seeking influence in the corridors of power, Big Tech firms can’t get enough of Britain’s leading politicos. From former policy advisers to ex-spin doctors and even the odd journalist, the Big Tech ranks are now dotted with Westminster insiders from across the political spectrum. POLITICO takes you through some of the leading lights.

«

Surprising number of British politicians and spindoctors there who have gone to work for Big Tech. You will surely have heard of one, maybe two, but if you’ve heard of more than five (out of the 22) you’re very clued in.

And guess which company has the most of the 22: Meta, Amazon, Uber, TikTok, Palantir, OnlyFans, Google, Deliveroo or Zoom?
unique link to this extract


Elon Musk takes @x handle from its original user. He got zero dollars for it • Mashable

Matt Binder:

»

Musk now has the @x handle. What happened? Did Musk reach out to [co-founder of event photo company Orange Photography, Gene X.] Hwang? Did Hwang cash in and get a paycheck from the company for the handle as some on social media have speculated?

No, the company just took it from him.

“[I] got an email basically saying they are taking it,” Hwang told Mashable in an email exchange.

Hwang previously told Mashable he was waiting for Twitter (or X now) to contact him so he wasn’t previously in contact with the company. He hadn’t heard from X before that. There was no back-and-forth discussion. The email came from the generic support@twitter.com email address and not from a specific employee within the company.

Musk’s company is within its right to take the username. Barring trademark issues, users don’t have rights to specific handles according to most social media companies’ terms of service. However, with reports that Twitter had been considering launching a service where users could bid on unused, rare handles, some users had thought the company would offer Hwang something.

According to Hwang, the company now formerly known as Twitter did offer “an alternative handle with the history of the @x account” so that his original account, complete with its posts and followers, could live on and continue to be used.

What short, catchy username did Musk’s company change Hwang’s handle to? @x12345678998765.

«

I guess @x12345678998764 was taken. It’s remarkable if Hwang has managed to hang on to that handle for any length of time even without Musk about: hackers adore getting hold of those rare handles on any social media platform and will use all sort of measures to achieve their aim. Or you can buy the whole shebang for $44bn, of course.
unique link to this extract


Elon Musk’s rebranded Twitter cuts ad prices • WSJ

Suzanne Vranica and Patience Haggin:

»

The social network formerly known as Twitter is offering new incentives on certain ad formats in the US and UK and warning brands that they will lose their verified status unless they reach certain spending thresholds, emails sent this week to advertisers and viewed by The Wall Street Journal show.

The company, which makes most of its money from advertising, has struggled to draw new ad commitments under Musk’s ownership in part because brands are concerned about Musk’s approach to management and content moderation. The advertising industry is also in a slump, and several media companies have begun offering brands discounts.

X this week began offering some advertisers reduced pricing on video ads that run alongside a list of trending topics in X’s “Explore” tab, according to emails viewed by the Journal. Such ads give brands 24-hour placement atop the site’s list of trending topics.

It is offering 50% off any new bookings of those ads until July 31, among other discounts. “The goal of these discounts is to help our advertisers gain reach during crucial moments on Twitter such as the Women’s World Cup,” one of the emails read.

X also warned advertisers that beginning Aug. 7, brands’ accounts will lose their verification—a gold check mark that indicates their account truly represents their brand—if they haven’t spent at least $1,000 on ads in the previous 30 days or $6,000 on ads in the previous 180 days, according to the email.

«

TwitterVersion2IsGoingJustGreat.com
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.