We don’t do business with ants, even though they can do a lot we might find useful. Would super-AIs treat us like ants too? CC-licensed photo by Nicolas Rénac 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. Us, colonial? I’m @charlesarthur on Twitter. On Mastodon: https://newsie.social/@charlesarthur. Observations and links welcome.

---

Instagram connects vast paedophile network • WSJ

Jeff Horwitz and Katherine Blunt:

»

Instagram, the popular social-media site owned by Meta Platforms, helps connect and promote a vast network of accounts openly devoted to the commission and purchase of underage-sex content, according to investigations by The Wall Street Journal and researchers at Stanford University and the University of Massachusetts Amherst.

Pedophiles have long used the internet, but unlike the forums and file-transfer services that cater to people who have interest in illicit content, Instagram doesn’t merely host these activities. Its algorithms promote them. Instagram connects pedophiles and guides them to content sellers via recommendation systems that excel at linking those who share niche interests, the Journal and the academic researchers found.

Though out of sight for most on the platform, the sexualized accounts on Instagram are brazen about their interest. The researchers found that Instagram enabled people to search explicit hashtags such as #pedowhore and #preteensex and connected them to accounts that used the terms to advertise child-sex material for sale. Such accounts often claim to be run by the children themselves and use overtly sexual handles…

…The promotion of underage-sex content violates rules established by Meta as well as federal law.

In response to questions from the Journal, Meta acknowledged problems within its enforcement operations and said it has set up an internal task force to address the issues raised.

«

This is terrible. And it’s worse than the 2019 discovery that paedophiles were using YouTube comments on videos of children to create ad-hoc networks. There, at least, YouTube wasn’t promoting the comments. (The link should be free to read.)
unique link to this extract

---

TV’s streaming model is broken. It’s also not going away • Vulture

Josef Adalian and Lane Brown:

»

Across the town [ie Hollywood], there’s despair and creative destruction and all sorts of countervailing indicators. Certain shows that were enthusiastically green-lit two years ago probably wouldn’t be made now. Yet there are still streamers burning mountains of cash to entertain audiences that already have too much to watch. Netflix has tightened the screws and recovered somewhat, but the inarguable consensus is that there is still a great deal of pain to come as the industry cuts back, consolidates, and fumbles toward a more functional economic framework.

The high-stakes Writers Guild of America strike has focused attention on Hollywood’s labour unrest, but the really systemic issue is streaming’s busted math. There may be no problem more foundational than the way the system monetises its biggest hits: it doesn’t.

Just ask Shawn Ryan. In April, the veteran TV producer’s latest show, the spy thriller The Night Agent, became the fifth-most-watched English-language original series in Netflix’s history, generating 627 million viewing hours in its first four weeks. As it climbed to the heights of such platform-defining smashes as Stranger Things and Bridgerton, Ryan wondered how The Night Agent’s success might be reflected in his compensation.

“I had done the calculations. Half a billion hours is the equivalent of over 61 million people watching all ten episodes in 18 days. Those shows that air after the Super Bowl — it’s like having five or ten of them. So I asked my lawyer, ‘What does that mean?’” recalls Ryan. As it turns out, not much. “In my case, it means that I got paid what I got paid. I’ll get a little bonus when season two gets picked up and a nominal royalty fee for each additional episode that gets made. But if you think I’m going out and buying a private jet, you’re way, way off.”

«

An absorbing read which also makes one think that streaming, whether music or video (and perhaps also spoken, ie podcasts?), is only good for the consumer and the provider, but not the content creator. That, though, carries the seeds of its own destruction.
unique link to this extract

---

The platforms give up on 2020 [US election] lies • Platformer

Casey Newton:

»

One function [Elon] Musk now serves in the tech ecosystem is to give cover to other companies seeking to make unpalatable decisions. Across a variety of dimensions, Musk has moved fast and loudest — and when others have followed, the response has been barely a whimper.

Mass layoffs, stricter job performance requirements, a war on remote work, paid verification for social accounts — all of these served as a kind of aphrodisiac for other Silicon Valley CEOs, who proceeded to implement their own, slightly softer versions of Musk’s cultural reset.

Most recently, Twitter’s decaying policy and enforcement systems have proven to be enticing for other social platforms.

Last month, for example, Musk told an interviewer that users who made false claims about the 2020 election being stolen “would be corrected.” But there was no accompanying effort to make that happen. And so, that same week, the top 10 posts promoting a rigged election narrative racked up a collective 43,000 retweets, the Associated Press reported.

As Musk was surely not aware, his predecessors had sought to unwind the company’s enforcement of 2020 election lies. In January 2022, CNN reported to general surprise that Twitter had abandoned its old policy in March 2021. Enforcement measures were intended to operate only until the next president was inaugurated, a spokeswoman said at the time, and no longer.

In any case, Twitter’s peers took notice of its reversal and chose to follow suit. In February, Meta restored Donald Trump’s accounts, and upon reinstating him said it would no longer prevent users from lying about the 2020 election. And on Friday, YouTube announced that it wouldn’t, either.

«

Twitter used to be a great leader in setting these standards. No longer. And as Newton points out, that means defection becomes much easier for all the other platforms. None of which is encouraging, as another American election looms.
unique link to this extract

---

India pauses plans to add new coal plants for five years, bets on renewables, batteries • AP News

Sibi Arasu:

»

The Indian government will not consider any proposals for new coal plants for the next five years and focus on growing its renewables sector, according to an updated national electricity plan released Wednesday evening.

The temporary pause in the growth of the dirty fuel was hailed by energy experts as a positive step for a country that is currently reliant on coal for around 75% of its electricity.

Updated every five years, the plan serves as a guideline for India’s priorities in its electricity sector.

India is the world’s third highest emitter and most populous country. It plans to reach net zero emissions by 2070, which would mean significantly slashing coal use and ramping up renewable energy.

In a draft of the plan released in September, the Central Electricity Authority, which is in charge of planning for India’s electricity needs, projected that nearly 8,000 megawatts of new coal capacity was required by 2027. But Wednesday’s strategy proposes the build out of more than 8,600 megawatts of battery energy storage systems instead.

«

Great to see India taking this crucial step. You don’t solve climate change by adding to it (NB Rowan Atkinson).
unique link to this extract

---

First impressions of Vision Pro and VisionOS • Daring Fireball

John Gruber is impressed with his impressions:

»

The worlds, as it were, of MacOS and iOS (or Windows, or Android, or whatever) are defined and limited by the displays on which they run. If MacOS is a place I go mentally when working, that place is manifested physically by the Mac’s display. It’s like the playing field, or the court, in sports — it has very clear, hard and fast, rectangular bounds. It is of fixed size and shape, and everything I do in that world takes place in the confines of those display boundaries.

VisionOS is very much going to be a conceptual place like that for work. But there is no display. There are no boundaries. The intellectual “place” where the apps of VisionOS are presented is the real-world place in which you use the device, or the expansive virtual environment you choose. The room in which you’re sitting is the canvas. The whole room. The display on a Mac or iOS device is to me like a portal, a rectangular window into a well-defined virtual world. With VisionOS the virtual world is the actual world around you.

…This is not confusing or complex, but it feels profound. Last night I chatted with a friend who, I found out only then, has been using Vision Pro for months inside Apple. While talking about this “your real world room is your canvas for arranging your application windows” aspect of the experience, he said that he spent weeks feeling a bit constrained, keeping his open VisionOS windows all in front of him as though on a virtual display, before a colleague opened his mind to spreading out and making applications windows much larger and arranging them in a wider carousel not merely in front of him but around him. The constraints of even the largest physical display simply do not exist with VisionOS.

«

Someone pointed out to me on Twitter that a Vision Pro might be like a laptop that would let you have multiple huge screens yet would fit in a kitchen drawer at the end of the work day. If Apple can solve the (keyboard) input problem, this could be remarkable.
unique link to this extract

---

We don’t trade with ants • world spirit sock puppet

Katja Grace:

»

When discussing advanced AI, sometimes the following exchanges happens:

“Perhaps advanced AI won’t kill us. Perhaps it will trade with us”

“We don’t trade with ants”

I think it’s interesting to get clear on exactly why we don’t trade with ants, and whether it is relevant to the AI situation.

When a person says “we don’t trade with ants”, I think the implicit explanation is that humans are so big, powerful and smart compared to ants that we don’t need to trade with them because they have nothing of value and if they did we could just take it; anything they can do we can do better, and we can just walk all over them. Why negotiate when you can steal?

I think this is broadly wrong, and that it is also an interesting case of the classic cognitive error of imagining that trade is about swapping fixed-value objects, rather than creating new value from a confluence of one’s needs and the other’s affordances. It’s only in the imaginary zero-sum world that you can generally replace trade with stealing the other party’s stuff, if the other party is weak enough.

Ants, with their skills, could do a lot that we would plausibly find worth paying for.

«

The implication, of course, being that we humans would be the ants compared to the superintelligences. Grace makes some good points about what ants can do (and so cheaply!) which leads one to puzzle about why, dammit, we don’t trade with ants.
unique link to this extract

---

Global smartphone production drops to a ten-year quarterly low at 250m units • Trendforce

»

The ongoing global economic downturn continues to impact consumer confidence in the market. TrendForce reports that the global production volume of smartphones in 1Q23 was only 250m units—marking a 19.5% YoY decrease. This represents not only the greatest annual decrease but also a historic low in quarterly production since 2014.

Samsung observed a slight surge in Q1 production thanks to the launch of its Galaxy S23 series, reaching 61.5m units—a 5.5% QoQ rise. However, TrendForce predicts a nearly 10% drop in Q2 production due to weakening demand for new models. Apple faced a substantial 27.5% QoQ drop in smartphone production in Q1, delivering a total of 53.3m units. The new iPhone 14 series accounted for approximately 78% of this figure, an improvement from the same period last year. Nonetheless, as the company navigates the transition period between model launches, a projected decrease of 20% is expected in Q2.

«

That’s a colossal drop; it’s as bad as the PC market is seeing, and I think there’s a suspicion that this is more than cyclical; that people are just not seeing any reason to buy a new smartphone at all. Foldables haven’t moved the dial in the slightest; they’re a margin of error.

But maybe if Apple were to offer cameras on its phones that could record in 3D, for viewing later on a headset… others might take that idea up too.
unique link to this extract

---

Apple software chief Craig Federighi on iOS 17’s new privacy features • Fast Company

Michael Grothaus:

»

What does concern [Apple software VP Craig] Federighi from a privacy and security standpoint, however, is the human element. Specifically, he worries about a rise in the use of deepfakes, AI-generated audio and video that can make it look like anyone is saying or doing anything. As AI tools become more accessible in the years ahead, deepfakes could increasingly be used in so-called social engineering attacks, in which the attacker persuades a victim to hand over valuable data by tricking them into thinking they are communicating with someone they’re not.

“When someone can imitate the voice of your loved one,” he says, spotting social engineering attacks will only become more difficult. If “someone asks you, ‘Oh, can you just give me the password to this and that? I got locked out,’ ” and it literally sounds like your spouse, that, I think, is going to be a real threat.” Apple is already thinking through how to defend users from such trickery. “We want to do everything we can to make sure that we’re flagging [deepfake threats] in the future: Do we think we have a connection to the device of the person you think you’re talking to? These kinds of things. But it is going to be an interesting time,” he says, and everyone will need to “keep their wits about them.”

«

There’s plenty more in what is a very interesting interview; Federighi is by far the most expansive of the Apple VPs on parade these days.
unique link to this extract

---

Apple avoids “AI” hype at WWDC keynote by baking ML into products • Ars Technica

Benj Edwards:

»

Amid impressive new products like the Apple Silicon Mac Pro and the Apple Vision Pro revealed at Monday’s WWDC 2023 keynote event, Apple presenters never once mentioned the term “AI,” a notable omission given that its competitors like Microsoft and Google have been heavily focusing on generative AI at the moment. Still, AI was a part of Apple’s presentation, just by other names.

While “AI” is a very ambiguous term these days, surrounded by both astounding advancements and extreme hype, Apple chose to avoid that association and instead focused on terms like “machine learning” and “ML.” For example, during the iOS 17 demo, SVP of Software Engineering Craig Federighi talked about improvements to autocorrect and dictation:

»

Autocorrect is powered by on-device machine learning, and over the years, we’ve continued to advance these models. The keyboard now leverages a transformer language model, which is state of the art for word prediction, making autocorrect more accurate than ever. And with the power of Apple Silicon, iPhone can run this model every time you tap a key.

«

Notably, Apple mentioned the AI term “transformer” in an Apple keynote. The company specifically talked about a “transformer language model,” which means its AI model uses the transformer architecture that has been powering many recent generative AI innovations, such as the DALL-E image generator and the ChatGPT chatbot.

«

As Edwards effectively points out, not saying “AI” wasn’t an omission; the presenters didn’t somehow forget to include it. They purposely avoided saying it because they’re trying very hard not to get lumped in with the other companies.
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