Call centre workers might be some of the first people to be displaced by chatbots, after Klarna found its satisfactory in a trial. CC-licensed photo by ILO Asia-Pacific 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 10 links for you. Your call is important to us. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. Observations and links welcome.

---

Behind Apple’s doomed car project: false starts and wrong turns • The New York Times

Brian Chen and Tripp Mickle:

»

Throughout its existence, the car effort was scrapped and rebooted several times, shedding hundreds of workers along the way. As a result of dueling views among leaders about what an Apple car should be, it began as an electric vehicle that would compete against Tesla and morphed into a self-driving car to rival Google’s Waymo.

By the time of its death — Tuesday, when executives announced internally that the project was being killed and that many members of the team were being reassigned to work on artificial intelligence — Apple had burned more than $10bn on the project and the car had reverted to its beginnings as an electric vehicle with driving-assistance features rivaling Tesla’s, according to a half dozen people who worked on the project over the past decade.

The car project’s demise was a testament to the way Apple has struggled to develop new products in the years since Steve Jobs’s death in 2011. The effort had four different leaders and conducted multiple rounds of layoffs. But it festered and ultimately fizzled in large part because developing the software and algorithms for a car with autonomous driving features proved too difficult.

…Despite having a vote of confidence from Apple’s chief executive, members of the team knew they were working against harsh realities, according to the six employees familiar with the project. If it ever came to market, an Apple car was likely to cost at least $100,000 and still generate razor-thin profit compared with smartphones and earbuds. It would also arrive years after Tesla had dominated the market.

The company held some discussions with Elon Musk about acquiring Tesla, according to two people familiar with the talks. But ultimately, it decided that building its own car made more sense than buying and integrating another business.

«

My feeling is that Jony Ive-thinking infected the project too early: make a thing that offers the fewest affordances possible. Why have a steering wheel or accelerator if the car drives itself? Except the self-driving part isn’t as simple as drawing a keyboard on the LCDs beneath a touch-sensitive surface. It’s orders of magnitude more difficult. And people like having stuff to fiddle with in a car. Some dashboards are basically huge fidget spinners for passenger and driver alike.
unique link to this extract

---

What we lost when Twitter became X • The New Yorker

Sheon Han worked at Twitter for a couple of years, and left with the Musk clearout:

»

Community can be a fuzzy, sentimental notion. But, on Twitter, communities are concrete. The platform’s recommendation algorithm is powered by “SimClusters,” a representation of overlapping communities that, according to the company, “range in size from a few thousand users for individual friend groups, to hundreds of millions of users for news or pop culture,” and are “anchored by a cluster of influential users.”

Pre-Musk Twitter leaned into fostering such communities; the month before the acquisition, an all-hands meeting featured a presentation from the company’s head of global K-pop and K-content partnerships, whose responsibilities involved promoting collaborations between Twitter and key players in the K-pop industry. But if a community can be fostered it can also fade. Every time a high-profile user leaves the platform in response to Musk’s antics, a critical node in the social graph is removed.

I wonder whether Musk understands that to undermine communities is to weaken the principal element that sustains the service. To monitor the health of a social-media platform, you can ask a question you might also ask of an indie-music venue: Is it still cool to hang out there? Since the takeover, for many people, it doesn’t “feel good” to be on Twitter. Friends are leaving, and tweeting feels like shouting into the void.

What does the future hold? It seems likely that users will still come for breaking news, and for expert threads, and for the memes recycled by dedicated joke accounts. Some weirdness will persist—and yet the weirdos will be gone. The platform will have lost its élan. Twitter’s laughably unserious name belied its seriousness. But X, with its overbearing name, may not prosper unless it undertakes the serious work of maintaining a platform on which people want to be.

«

unique link to this extract

---

Gartner predicts search engine volume will drop 25% by 2026, due to AI chatbots and other virtual agents • Gartner

»

By 2026, traditional search engine volume will drop 25%, with search marketing losing market share to AI chatbots and other virtual agents, according to Gartner, Inc.

“Organic and paid search are vital channels for tech marketers seeking to reach awareness and demand generation goals,” said Alan Antin, Vice President Analyst at Gartner. “Generative AI (GenAI) solutions are becoming substitute answer engines, replacing user queries that previously may have been executed in traditional search engines. This will force companies to rethink their marketing channels strategy as GenAI becomes more embedded across all aspects of the enterprise.”

With GenAI driving down the cost of producing content, there is an impact around activities including keyword strategy and website domain authority scoring. Search engine algorithms will further value the quality of content to offset the sheer amount of AI-generated content, as content utility and quality still reigns supreme for success in organic search results.

There will also be a greater emphasis placed on watermarking and other means to authenticate high-value content. Government regulations across the globe are already holding companies accountable as they begin to require the identification of marketing content assets that AI creates. This will likely play a role in how search engines will display such digital content.

«

If we assume that this is correct, then for Google, that’s a near-existential collapse unless it can find some way to replace those searches (and their associated ad revenue) with AI-related ones.
unique link to this extract

---

The leap day is only half of the leap year fun • rachelbythebay

“Rachel”:

»

Only half of the fun of a leap year happens on February 29th.

The rest of it happens in ten months, when a bunch more code finds out that it’s somehow day 366, and promptly flips out. Thus, instead of preparing to party, those people get to spend the day finding out why their device is being stupid all of the sudden.

So, if you got through today unscathed, but are somehow counting days in the year somewhere, you now have about 305 days to make sure you don’t have your own Zune bug buried in your own code.

One more random thought on the topic: some of today’s kids will be around to see what happens in 2100. That one will be all kinds of fun to see who paid attention to their rules and who just guessed based on a clean division by four.

«

(The link to the Zune bug isn’t in the original; it was the first that came up on my search for “zune bug leap year”.)
unique link to this extract

---

The Humane AI Pin worked better than I expected — until it didn’t • The Verge

Allison Johnson got a demo of the Pin at Mobile World Congress:

»

The AI Pin was genuinely impressive at times. There’s a vision feature that will use the camera to scan the scene in front of you when prompted, analyze what’s there, and describe it out loud. I stood in front of a Humane spokesperson as he tried out this feature, and frankly, the pin nailed it. It described Mobile World Congress as “an indoor event or exhibition with people walking around.” Easy enough.

But it also pointed out the name Qualcomm on the signage behind me, and obviously reading the badge around my neck, identified me as “a person wearing a lanyard from the The Verge.” One too many the’s, but pretty impressive when you consider I wasn’t standing all that close to the pin and the lighting was dim.

The gesture navigation was also impressive — more fluid and responsive than I thought it would be. I wasn’t allowed to put the pin on myself, and it’s hard to get into the right spot to project the laser onto your own hand since it’s really a single-user device. I tried. But a couple of Humane employees demoing the product, who obviously had lots of practice with it, navigated the projected menus quickly and easily just by tilting their hands and tapping two fingers together.

But the pin isn’t immune to the thing that gadgets often do: frustrate the hell out of you. Most of the AI is off-device, so there’s a solid few seconds of waiting for responses to your requests and questions — not helped by the convention center’s spotty connectivity. It also shut down on one occasion after briefly flashing a notice that it had overheated and needed to cool off. The employee demoing the pin for me said that this doesn’t happen very often, and that the continued use of the laser for demonstration purposes probably did it. I believe that, but still, this is a device meant to sit next to your chest and go with you into lots of different environments, presumably including warm ones. Not great!

«

unique link to this extract

---

Germany to adopt 2060 target for net-negative greenhouse gas emissions • Clean Energy Wire

Julian Wettengel:

»

The German government is aiming to introduce a 2060 target for net-negative greenhouse gas emissions, as well as intermediate targets for technical carbon sinks, as key elements of its contribution to the Paris climate targets.

By the end of 2024, the ruling coalition wants to agree on a long-term strategy for negative emissions to help deal with residual emissions which are difficult or impossible to avoid. In a document outlining the upcoming strategy, the government says that limiting the global temperature rise to 1.5°C is “increasingly unlikely”, so negative emissions will also be necessary to lower the CO2 concentration in the atmosphere to limit the risks of severe and irreversible consequences for humans and ecosystems.

The strategy will set the targets, evaluate different carbon dioxide removal methods, and analyse economic incentives to help ramp up the necessary technologies. Experts say Germany could become a frontrunner on CO2 removal policy with the strategy.

«

And just think how much sooner you could have done this, Germany, if you hadn’t prematurely shut down your nuclear power stations because one of your political parties worried irrationally that the country would be overwhelmed by a tsunami.
unique link to this extract

---

A most wanted man: fugitive Wirecard COO Jan Marsalek exposed as decade-long GRU spy • The Insider

Roman Dobrokhotov, Christo Grozev, and Michael Weiss:

»

In the city of Lipetsk, 300 miles south of Moscow, stands a yellow chapel. Somewhat out of place next to a modern mirrored-window building, situated on the lip of a roundabout, the 200 year-old Church of Holy Transfiguration caters to the faithful of a large mining town that dates back to the era of Peter the Great. Inside, Father Konstantin Baiazov performs the customary rites and rituals for his flock. Dark and bearded, with a short, military-style buzz cut, the church’s archpriest’s routine is standard – services twice a day. Father Konstantin inherited the job — and the calling — from his own father, a revered Orthodox priest who, as local legend goes, had challenged the authority of the formidable KGB during Soviet times.

Konstantin, the father of three, used to travel abroad. He liked visiting Europe, and was particularly fond of Rome. However, he has not left Russia since September 2020. Since the fifth of that month, Father Baiazov’s official passport, numbered 763391844, has not belonged to a man of God. Rather, it belongs to someone who wears a different kind of white collar, looks a lot like him, and is the most wanted man in Europe.

For more than four years, Jan Marsalek, the former chief operating officer of the disgraced German financial services company Wirecard, has been living in Russia under this assumed identity, a year-long investigation by The Insider, Der Spiegel, ZDF, and Der Standard has uncovered.

…But Marsalek is not only an internationally accused swindler. He is also an agent of the GRU, Russia’s military intelligence service, and he has been for the last decade. More recently, since his defection to Russia, he has also done jobs for the FSB.

«

This is what is known in the news trade as a marmalade-dropper: something that makes you drop your toast in shock. It’s not short but it seems like a classic piece of recruitment, starting with a honeypot.
unique link to this extract

---

The Twitter settings you should change now to block unwanted calls • Forbes

Barry Collins:

»

[Last] week, X (formerly Twitter) announced that audio and video calls are now available to everyone on the service. By default, this means anyone you follow can make an audio or video call to you.

X has automatically turned this on for everyone, there’s no opt-in. Suddenly, the mere act of following a person or brand gives them the right to phone you.

Some users may welcome this new feature, but many will be concerned about the potential for interruptions and unwanted calls. It’s not as if X has a sparkling record with dealing the bot accounts that Elon Musk once promised to eradicate.

If you want to ensure you’re not bothered by junk calls, here are the settings you need to change now.

To access the relevant settings, you’ll need to open the Twitter app on your smartphone. Now you should:

• Tap the envelope icon used to access your direct messages
• Click the settings cog at the top of the screen

You’ll now be presented with a series of options. You can simply block all video and audio calls outright by unchecking the box that says “enable audio and video calling.”

Beneath that are more nuanced options, which let you choose who can call you.

«

How about “nobody”? “Nobody” works for me. (And of course they made it default-on. Ugh.)
unique link to this extract

---

Klarna AI assistant handles two-thirds of customer service chats in its first month • Klarna Media Centre

»

Klarna today announced its AI assistant powered by OpenAI. Now live globally for 1 month, the numbers speak for themselves:

• The AI assistant has had 2.3 million conversations, two-thirds of Klarna’s customer service chats

• It is doing the equivalent work of 700 full-time agents

• It is on par with human agents in regard to customer satisfaction score

• It is more accurate in errand resolution, leading to a 25% drop in repeat inquiries

• Customers now resolve their errands in less than 2 mins compared to 11 mins previously

• It’s available in 23 markets, 24/7 and communicates in more than 35 languages

• It’s estimated to drive $40m in profit improvement to Klarna in 2024

Klarna has also seen massive improvement in communication with local immigrant and expat communities across all our markets thanks to the language support.

«

Seems like there will be a lot of former customer service agents. There are millions with that job title in the US alone. Though I do wonder whether those conversations are truly as satisfying as dealing with humans. Maybe I can get my chatbot to talk to your chatbot and sort all this out? That’s the obvious next stage.
unique link to this extract

---

Wikipedia no longer considers CNET a “generally reliable” source after AI scandal • Futurism

Maggie Harrison Dupré:

»

Remember last year, when we reported that the Red Ventures-owned CNET had been quietly publishing dozens of AI-generated articles that turned out to be filled with errors and plagiarism?

The revelation kicked off a fiery debate about the future of the media in the era of AI — as well as an equally passionate discussion among editors of Wikipedia, who needed to figure out how to treat CNET content going forward.

“CNET, usually regarded as an ordinary tech [reliable source], has started experimentally running AI-generated articles, which are riddled with errors,” a Wikipedia editor named David Gerard wrote to kick off a January 2023 discussion thread in Wikipedia’s Reliable Sources forum, where editors convene to decide whether a given source is trustworthy enough for editors to cite.

“So far the experiment is not going down well, as it shouldn’t,” Gerard continued, warning that “any of these articles that make it into a Wikipedia article need to be removed.”

Gerard’s admonition was posted on January 18, 2023, just a few days after our initial story about CNET’s use of AI. The comment launched a discussion that would ultimately result in CNET’s demotion from its once-strong Wikipedia rating of “generally reliable.”

It was a grim fall that one former Red Ventures employee told us could “put a huge dent in their SEO efforts,” and also a cautionary tale about the wide-ranging reputational effects that publishers should consider before moving into AI-generated content.

«

Wikipedia generates a ton of SEO juice for referred sites, because Wikipedia itself is one of the most linked-to sites on the web. So yes, this is bad for Red Ventures.
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