Start Up No.2005: why Google’s Bard isn’t in the EU, astronomy meets AI, Twitter’s new “CEO” profiled, media’s traffic dream, and more

How did Salvador Dali become the most faked artist? By producing too much “art”, it seems. CC-licensed photo by cea. 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.

A selection of 9 links for you. I’m melting! I’m @charlesarthur on Twitter. On Mastodon: Observations and links welcome.

Google Bard hits over 180 countries and territories—none are in the EU • Ars Technica

Scharon Harding:


On Wednesday, Google detailed the evolution of its Bard conversational AI assistant, including PaLM 2 and expanded availability. The list of 180 supported countries and territories excludes Canada and all of the European Union’s (EU) 27 member states. As the world grapples with how to juggle the explosive growth of generative AI chatbots alongside user privacy, there’s suspicion that the EU’s General Data Protection Regulation (GDPR) is at the center of the omission.

Google’s I/O event this week included flashy announcements around AI developments and expanding Bard access with added Japanese and Korean language support. However, some people quickly noticed that EU countries and The Great White North were not part of the news. This could change, as Google’s support page says the company will “gradually expand to more countries and territories in a way that is consistent with local regulations and our AI principles.”

In the meantime, Google hasn’t explained why it’s not yet bringing Bard to the EU, Canada, or any other excluded geography. However, the EU features more stringent data protection and user privacy policies than Google’s homeland. And the EU’s AI regulatory landscape is on the brink of transformation.

…Italy has rather active privacy regulators and was one of the first countries to restrict access to an AI like [rival OpenAI’s] ChatGPT. When announcing its temporary ban in April, the Italian government said ChatGPT had to comply with measures around “transparency, the right of data subjects—including users and non-users—and the legal basis of the processing for algorithmic training relying on users’ data.”

OpenAI eventually complied with measures like sharing an online form that lets users opt out and delete data from ChatGPT’s training algorithms. OpenAI also checks Italian users’ birth dates upon signup to ensure they’re either 18 or older or have parental permission. Further, OpenAI said it would try to educate users about ChatGPT through a publicity campaign with details like how users can decline to share data.

By not releasing Bard in the EU, Google can avoid jumping through similar hoops OpenAI faced to retain availability in Italy.


unique link to this extract

AI is speeding up astronomical discoveries • Gizmodo

Chris Impey:


Astronomers working on SETI, the Search for Extraterrestrial Intelligence, use radio telescopes to look for signals from distant civilizations. Early on, radio astronomers scanned charts by eye to look for anomalies that couldn’t be explained. More recently, researchers harnessed 150,000 personal computers and 1.8 million citizen scientists to look for artificial radio signals. Now, researchers are using AI to sift through reams of data much more quickly and thoroughly than people can. This has allowed SETI efforts to cover more ground while also greatly reducing the number of false positive signals.

Another example is the search for exoplanets. Astronomers discovered most of the 5,300 known exoplanets by measuring a dip in the amount of light coming from a star when a planet passes in front of it. AI tools can now pick out the signs of an exoplanet with 96% accuracy.

AI has proved itself to be excellent at identifying known objects – like galaxies or exoplanets – that astronomers tell it to look for. But it is also quite powerful at finding objects or phenomena that are theorized but have not yet been discovered in the real world.

Teams have used this approach to detect new exoplanets, learn about the ancestral stars that led to the formation and growth of the Milky Way, and predict the signatures of new types of gravitational waves.

To do this, astronomers first use AI to convert theoretical models into observational signatures – including realistic levels of noise. They then use machine learning to sharpen the ability of AI to detect the predicted phenomena.

Finally, radio astronomers have also been using AI algorithms to sift through signals that don’t correspond to known phenomena. Recently a team from South Africa found a unique object that may be a remnant of the explosive merging of two supermassive black holes. If this proves to be true, the data will allow a new test of general relativity – Albert Einstein’s description of space-time.


unique link to this extract

Why Salvador Dalí is the most faked artist in the world

Mark Dent:


By the 1950s and 1960s, it was clear the demand for Dalí’s work exceeded the supply.

So Dalí, Gala, and others in Dalí’s inner circle devised a solution: prints. Lithographs and etchings took less time to finish than paintings and could be reproduced as limited series.

There were two categories of Dalí prints:
• Fully original: Dalí created the images himself on a printing plate and signed a limited series of prints. Originals sold for up to $3.5k
• Legitimate prints: Some limited-edition lithographs were made by licensed publishers copying a watercolor of Dalí. These were not technically original, although they were marketed as such and approved and signed by Dalí. They could sell for nearly as much as the fully original prints.

Dalí ensured a steady flow of prints by signing his name on thousands of blank sheets of paper before he knew what would be printed on them. (The signature was worth ~$40 on its own.) Members of his inner circle, some of whom exploited Dalí for profit, once told the Wall Street Journal Dalí would sign blank sheets “every two seconds for an hour without stopping.”

The prints, the signatures, and the commercial contracts kept the dollars rolling in. Beyond, the magazine of the St. Regis Hotel, noted Dalí was as much “high finance” as he was “high art.”

But in the 1970s, the artist’s health declined, and he became a recluse for the next decade. Dalí stopped creating prints. He stopped signing his name. And yet, in a stroke of real-life surrealism, the world, and especially the US, was about to see more art attributed to Dalí than ever before.

At the Center Art Galleries in Honolulu, John Proctor’s job was to shadow visitors in the showroom. When they looked at Lincoln in Dalívision, he began his sales pitch, handing them a fact sheet revealing reported increases in value for Dalí’s art and setting them up with a “closer” to convince the visitors to spend as much as $11.5k for the print — enough, at the time, for a down payment on the median US home.

“It was easy to sell art to the tourists,” Proctor told The Honolulu Advertiser in 1980. “Once you tell them they’re going to make money, they get hot for the stuff.”


Guess what happened next. The problem of provenance will never go away.
unique link to this extract

Twitter’s encrypted DMs are deeply inferior to Signal and WhatsApp • WIRED

Andy Greenberg:


the company appears to have stopped short of calling the feature “end-to-end” encrypted, the term that would mean only users on the two ends of conversations can read messages, rather than hackers, government agencies that can eavesdrop on those messages, or even Twitter itself.

“As Elon Musk said, when it comes to Direct Messages, the standard should be, if someone puts a gun to our heads, we still can’t access your messages,” the help desk page reads. “We’re not quite there yet, but we’re working on it.”

In fact, the description of Twitter’s encrypted messaging feature that follows that initial caveat seems almost like a laundry list of the most serious flaws in every existing end-to-end encrypted messaging app, now all combined into one product—along with a few extra flaws that are all its own.

The encryption feature is opt-in, for instance, not turned on by default, a decision for which Facebook Messenger has received criticism. It explicitly doesn’t prevent “man-in-the-middle” attacks that would allow Twitter to invisibly spoof users’ identities and intercept messages, long considered the most serious flaw in Apple’s iMessage encryption. It doesn’t have the “perfect forward secrecy” feature that makes spying on users harder even after a device is temporarily compromised. It doesn’t allow for group messaging or even sending photos or videos. And perhaps most seriously, it currently restricts this subpar encrypted messaging system to only the verified users messaging each other—most of whom must pay $8 a month—vastly limiting the network that might use it.


It’s the latter point which is so strange. Why offer encryption – something which is table stakes (or assumed) in so many other networks – but only for people who pay for it?
unique link to this extract

Google AMP: how Google tried to fix the web by taking it over • The Verge

David Pierce:


Adopting Google’s strange new version of the web resulted in an irresistible flood of traffic for publishers at first: using AMP increased search traffic to one major national magazine’s site by 20%, according to the executive who oversaw the implementation.

But AMP came with huge tradeoffs, most notably around how all those webpages were monetized. AMP made it harder to use ad tech that didn’t come from Google, fraying the relationship between Google and the media so badly that AMP became a key component in an antitrust lawsuit filed just five years after its launch in 2020 by 17 state attorneys general, accusing Google of maintaining an illegal monopoly on the advertising industry. The states argue that Google designed AMP in part to thwart publishers from using alternative ad tools — tools that would have generated more money for publishers and less for Google. Another lawsuit, filed in January 2023 by the US Justice Department, went even further, alleging that Google envisioned AMP as “an effort to push parts of the open web into a Google-controlled walled garden, one where Google could dictate more directly how digital advertising space could be sold.”

Here in 2023, AMP seems to have faded away. Most publishers have started dropping support, and even Google doesn’t seem to care much anymore. The rise of ChatGPT and other AI services pose a much more direct threat to its search business than Facebook Instant Articles and Apple News ever did. But the media industry is still dependent on Google’s firehose of traffic, and as the company searches for its next move, the story of how it ruthlessly used AMP in an attempt to control the very structure and business of the web makes clear exactly how far it will go to preserve its business — and how powerless the web may be to stop it.

AMP succeeded spectacularly. Then it failed. And to anyone looking for a reason not to trust the biggest company on the internet, AMP’s story contains all the evidence you’ll ever need.


Great piece of reporting.
unique link to this extract

Apple M3 chip, Mac specifications and features: CPU, GPU and RAM increase details • Bloomberg

Mark Gurman, in his Power On newsletter:


Apple finally brings Final Cut Pro and Logic to the iPad. After a couple years of development, Apple is bringing two of its core pro apps to the iPad Air and the iPad Pro. The user interface of Final Cut Pro is designed to be touch-first (it works with a trackpad on a Magic Keyboard or similar device) and appears perfect for in-field edits or for high-end content creators. But, of course, it’s not going to replace the full functionality of Final Cut Pro on a Mac. In fact, it’s probably closer to iMovie back in the day, when it was actually functional. 


Yup, fine, sure, except that you can’t round-trip on Final Cut Pro – ie you can’t upload a project from your Mac to the iPad, edit it a bit, and then transfer it back to the Mac. You can only send it from the iPad to the Mac, probably (the ATP folk speculate) due to RAM restrictions. So it’s absolutely not perfect for high-end content creators.

Moving on..


Now, here’s the angle that I am really interested in: how Apple may be adapting Final Cut Pro and Logic to its upcoming mixed-reality headset. I’m told that the headset will have a content-creation focus and that its user interface, which relies on hand and eye control, could be precise enough to handle apps like Final Cut. On top of that, the device is supposed to work with any iPadOS app out of the box. That makes it seem likely that the new apps will run on the headset.


Good grief. Seriously? You think people are going to try to edit video or sound in a headset? This is like the Wall Street analyst who was convinced for years that Apple was going to produce a TV, and was eternally disappointed. More and more I hope Apple doesn’t release a headset, just to see how the amount of copium required.
unique link to this extract

The mother of all photoshoots • Event Photography London

Paul Clarke, who back in 2019 looked at the new set of House of Commons photos of MPs and, as a professional photographer, had a few thoughts:


Did you ever look at a black and white version of your face and think – “oh, that’s so much nicer!”? There are a few reasons why this can happen, but one of them is that black and white is a quick way to create a distance from reality. Given the, er, complex relationship most of us have with our own image, having a bit of room to see ourselves abstracted can often help us accept, or even enjoy, the result. With this portrait set, I think there’s been a deliberate choice to ‘cool down’ the images – shifting the colour palette down to the blue end of the spectrum. It’s what makes these pictures look a little ‘blue’ or ‘cold’ overall. It really helps to give them a distinctive look, but it also helps to make them just a little bit unreal – at least in tone.

They are, however, ruthlessly real in other aspects. They are, as far as I can judge, unretouched. We’re in really interesting territory here in terms of what we mean by the ‘truth’ of a portrait. Whether Cromwell actually used the words “warts and all” to his portrait artist Sir Peter Lely, is unknown. But we all recognise the sentiment. The role of the portrait painter was to convey an artistic impression – very possibly a flattering one – of the subject. But the role of the photographer? Well, within the world of PR photography, not all that different. But in the world of journalism? Very. A little adjustment of colour, brightness and contrast, maybe, but no retouching as such.

So are these photos to be seen as PR work, or journalism? In a sense they fall between the two stools. They are not a “news story” (although they did become one) nor are they an exercise in image management. The project team have come down firmly on the side of the journalists – unairbrushed reality. If the subject has a bit of a sweat on, it’s in. A pimple or a wart? Same. A few flakes of dandruff or a stray hair on their collar? You get the picture. It’s really easy to see the sort of problems a project like this might run into if it were seen to have manipulated the photos to flatter. But it shows the weight of the decisions that were involved.


Paul pointed me to this post of his after the discussion last week about “what is a photo?” In this post he does a little “work” on the photos. His rates are very reasonable, I understand.
unique link to this extract

BuzzFeed, Gawker, and the casualties of the traffic wars • The New Yorker

Nathan Heller reviews Traffic, the new book by Ben Smith (ex-Buzzfeed News); Heller worked at a webzine at that time, and this coda to his review sums up the problem faced by media today aptly:


At the online magazine where I worked, the measure of success in traffic-seeking kept changing. The goal was at first to maximize the number of unique page views by publishing more material.

Then instructions came down that what mattered was not volume but authority (other reliable sites linking to us), and we were instructed to reach out to eminent bloggers to promote our wares. After some months of this, it was decided that, in fact, the most valuable measure of traffic was engagement (how long readers spent reading our articles); our brief was to do work that was longer, better, and nearer the headlines of the day. When that approach, too, generated insufficient revenue, volume was summoned as the solution once again.

The media business has since made at least one more complete turn on this traffic roundabout in the hope of stabilizing its future. (The line is usually that the last model “isn’t how the Web works.”) And the will to traffic is now everywhere: on your phone, in your ears, on your screen.

In dreamy moods, I sometimes fantasize about journalism dropping out of the game—not chasing traffic, not following this year’s wisdom, not offering audiences everything they could possibly want in hastiest form. Imagine producing as little as you could as best you could: it would be there Monday, when the week began, and there Friday, the tree standing after the storm. And imagine the audience’s pleasure at finding it, tall and expansive and waiting for a sunny day. In an age of traffic, such deliberateness could be radical. It could be, I think, the next big thing.


unique link to this extract

Can Linda Yaccorino keep Elon Musk on a tight enough leash to succeed? • Fortune

Kylie Robison:


Sources describe Yaccarino as a “tough,” traditional Italian, “Long Island lady” who can both inspire and terrify the people who work for her. She has an identical twin, who’s a nurse.

Her fearless attitude in the male-dominated ad business is undoubtedly one of her biggest strengths, and could be an important part of her professional relationship with Musk. She’s capable of playing the long game, said a source, describing her rise at NBC: “She came in knowing that she was going to run the whole thing, but she started with cable and she took over broadcast.” Two sources also told Fortune she’s a sharp negotiator, which will help her when it comes to crafting a smart employment contract with Musk.

“She stood up to a lot of misogyny. She stood up to a lot of men,“ the source recalled during her time working with Yaccarino. “There’s nothing demure about her.”

Her values are well documented, too. She’s a devoted Catholic and staunch Republican. When former president Donald Trump was elected to office, Yaccarino attended his inauguration, one source told Fortune. Then, in 2018, Trump appointed Yaccarino to serve a two-year term on the President’s Council on Sport, Fitness and Nutrition, alongside big names like New England Patriots head coach Bill Belichick and Incredible Hulk star Lou Ferrigno, Adweek reported. One source said she joined “just to get near Trump.” 

Some have speculated that Musk chose Yaccarino because their political values aligned. Yet, according to Lou Paskalis, CEO of the marketing consultancy firm AJL Advisory, and a client of Yaccarino’s for 35 years, Yaccarino always kept her political views “fairly private.” That sensibility could provide the needed counterweight to Musk’s tendencies, Paskalis reckoned: ”She’ll probably be able to temper [Musk’s] enthusiasm for extreme commentary, but introduce more balance.”


She kept her politics “fairly private” but attended Trump’s inauguration? And what’s the relevance of her twin being a nurse? (None, it’s just a fact dump.) What’s clear is that she’s been brought in to bring back the advertisers. She won’t be a CEO. Musk will decide what the network looks like; she’ll be in charge of filling in the white spaces where the ads should be.
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 Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

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