The writings of Friedrich Nietzsche seem to have inspired, at least in part, the latest output from Marc Andreessen. CC-licensed photo by jwyg 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 10 links for you. Use them wisely. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. Observations and links welcome.

---

Google, DOJ still blocking public access to monopoly trial docs, NYT says • Ars Technica

Ashley Belanger:

»

Dozens of exhibits from the Google antitrust trial are still being hidden from the public, The New York Times Company alleged in a court filing on Monday.

According to The Times, there are several issues with access to public trial exhibits on both sides. The Department of Justice has failed to post at least 68 exhibits on its website that were shared in the trial, The Times alleged, and states have not provided access to 18 records despite reporters’ requests.

Google’s responses to document requests have also been spotty, The Times alleged. Sometimes Google “has not responded at all” to requests to review public exhibits. Other times, Google responds, but “often does not provide the exhibit in its entirety,” The Times claimed, including limiting public access to “particular page(s) of the exhibit shown to a given witness.”

The Times has asked the court to intervene and expand public access to key evidence weighed in what’s “arguably the most important antitrust trial in decades, with far-reaching consequences for the future of the tech industry.”

This is just the latest attempt to stop the Google antitrust trial from being shrouded in secrecy. Just before the trial, advocates lost a fight to get the court to provide a public access audio stream of the whole trial. Then shortly after the trial began, Google tried and failed to reduce public access to trial exhibits by requesting an opportunity to review every trial document before the DOJ posted anything online.

The drama over Google’s request to control how trial documents are shared concluded with an agreement between the DOJ and Google that either party would have an opportunity to object to the release of certain trial exhibits within three hours—a matter that both parties would have to prepare to argue the following trial day. Otherwise, either side “may” post the trial exhibit the next day.

…The public has a “weighty” constitutional right to access the Google trial exhibits to fully consider the government’s case against Google, The NYT argued.

«

Former patent lawyer Nilay Patel, now editor-in-chief at The Verge, writes on this too; multiple outlets have joined the NYT’s complaint, and he’s getting the Verge/Vox ensemble to do so too.
unique link to this extract

---

Netflix may hike prices after success of password-sharing crackdown • Reuters

Samrhitha A:

»

Netflix’s crackdown on password-sharing likely boosted subscribers by about 6 million in the third quarter and the streaming pioneer is expected to set the stage for price increases when it reports earnings on Wednesday.

The only profitable major streamer, Netflix has resisted joining rivals like Walt Disney in hiking ad-free prices this year and instead curbed password-sharing outside households to tap the more than 100 million viewers who use its service without subscribing.

“Netflix now closely resembles a utility in many markets,” analysts at Bernstein said. “The challenge of being labeled a utility is how a maturing company continues finding growth.”

It could hike prices after the end of the Hollywood actors strike, a media report said earlier in October.

Five months after calling a strike that plunged Hollywood into turmoil, the Writers Guild of America (WGA) last week approved a new contract with major studios.

Netflix, however, has weathered the strike well thanks to its larger international presence and strong content slate.

After a slow start for the ad plan launched last year, analysts said they expect Netflix will raise prices of its ad-free options in the coming months to nudge more subscribers to the other tier, where commercials help bring in more revenue per user.

«

This is the surprising thing: raising prices so that people drop down to the “cheaper” tier with adverts, which generates more money. Streaming is recapitulating the evolution of TV, though regressing backwards towards adverts.
unique link to this extract

---

Microsoft and Adobe push new symbol to label AI images • The Register

Katyanna Quach:

»

you can use Coalition for Content Provenance and Authenticity (C2PA)’s Content Credentials metadata for any picture – it doesn’t have to be AI generated. The examples given on the Content Credentials’ website were made using Adobe’s Photoshop and Firefly AI tools, and can be identified as such through their metadata. What Microsoft, Adobe and others have promised to do is ensure their AI generators will at some point in the future include this cryptographically signed metadata in their machine-crafted pictures. The goal being to provide a way for people to see if a picture was model or human-made and how.

For instance, Microsoft said artwork produced by its text-to-picture Bing Image Creator and Bing AI chatbot will feature that metadata at some point.

Now here’s the tricky part, assuming the specification is secure and robust. It’s one thing to store that metadata in a picture. How does the user find out, without digging into the file contents?

Well, you will need a compatible application, one that understands the Content Credentials metadata. If an app recognizes that data in a file, it should superimpose the “cr” symbol over the image in a top corner. When you click on that symbol, a widget should appear describing the source of the pic and other details from the Content Credentials metadata – for example, if it was made via Bing or Photoshop.

That’s how people can easily inspect the origin of the snap. But of course if the file is opened in an application that doesn’t support Content Credentials, no symbol is shown: the app won’t understand the data and won’t show a symbol.

«

This is definitely the problem. Even if this comes into force at once, it would take decades for all the image-using apps to incorporate it, and there will be an incentive for those who want to create undetectable fakes to use or create apps that don’t follow this protocol. Not a tragedy of the commons, but a version of it.
unique link to this extract

---

The Techno-Optimist Manifesto • Andreessen Horowitz

Marc Andreessen:

»

We are being lied to.

We are told that technology takes our jobs, reduces our wages, increases inequality, threatens our health, ruins the environment, degrades our society, corrupts our children, impairs our humanity, threatens our future, and is ever on the verge of ruining everything.

We are told to be angry, bitter, and resentful about technology.

We are told to be pessimistic.

The myth of Prometheus – in various updated forms like Frankenstein, Oppenheimer, and Terminator – haunts our nightmares.

We are told to denounce our birthright – our intelligence, our control over nature, our ability to build a better world.

We are told to be miserable about the future.

«

There’s 5,000 words of this, presented in short, sharp sentences full of aspiration and not much actual roadmapping. It reads a bit like a 17-year-old who has just read The Fountainhead and some Nietzsche and had a joint. (He recommends the work of, among others, “John Galt”.) Andreessen has a habit of writing stuff like this every few months: who remembers “It’s Time To Build” from April 2020, which complained that there weren’t any vaccines for Covid because “we chose not to ‘build'”.

O ye of little faith, Mr Andreessen.
unique link to this extract

---

Twitter is throttling Patreon links: creators say it undermines their livelihood • The Markup

Jon Keegan and Dan Phiffer:

»

Twitter is now slowing down traffic on links to the crowdfunding site Patreon, WhatsApp, and at times, Meta’s Messenger app, a Markup analysis confirms. 

Using a tool launched by The Markup last month, readers discovered that links to these sites were delayed by an average of 2.5 seconds—findings we confirmed. 

Patreon users told The Markup that the throttling undermines their ability to reach new supporters on Twitter, which has historically been a key platform for building donations. “It’s dirty pool,” said Matt Carlin, the producer of a popular call-in show, Office Hours Live, which takes in nearly $30,000 per month from Patreon and is being throttled. “It’s another big tech company trying to squash little guys—or, in their attempts to squash bigger guys, squashing little guys like us.”

In September, The Markup reported that Twitter, now officially named X, was slowing down links to Bluesky, Facebook, Instagram, and Substack, also by an average of 2.5 seconds, which can feel extremely slow for users. We simultaneously launched a tool that lets readers test any link posted on X (which the platform automatically shortens using the t.co domain), and measured the time it took for X to redirect the link to its original destination. The Markup also built a bot that would let us know if any links readers were testing appeared to be throttled.

«

Every link posted on Twitter goes through its t.co shortener, which guards against spam and malware and so on. Is it possible that the system is just running slower? Sure, 2.5 seconds is a long time in computer terms, but I struggle to believe that people stop short of going to Patreon because they have to wait one thousand, two thousand, three thou, for the link to work.
unique link to this extract

---

Fearing AI, fan fiction writers lock their accounts • TechCrunch

Morgan Sung:

»

Kinktober. Whumptober. Kisstober. Flufftober. Goretober. October is a bacchanal of fan fiction, from romantic one-shots about unconventional character pairings to delicious smut that’ll make you reconsider your own sense of morality — all inspired by the month’s countless themed writing challenges. It’s an especially busy time for the fan fiction site Archive of Our Own (AO3). 

But this year’s monthlong prompt festival may seem quieter to the casual AO3 reader, with popular writers’ work seemingly wiped from the site altogether. In most cases, the stories still exist, but they aren’t publicly viewable anymore. 

In an effort to prevent their writing from being scraped and used to train AI models, many AO3 writers are locking their work, restricting it to readers who have registered AO3 accounts. Though it may curb bot commenters, it also limits traffic from guest users, which can be a blow for newer and less popular writers. Whether it’s effective is questionable, but in the AI paranoia, AO3 writers are taking any measures they can to protect their work. 

At the time of reporting, over 966,000 of the roughly 11.7 million works on AO3 were accessible only for registered users. It’s just a fraction of AO3’s vast library of content, but it’s worth noting that many authors are only locking new work, since existing fics were likely already scraped. 

«

They were happy for anyone to read them and share them and learn from them as long as the “anyone” was human. But if the “anyone” who learns from their work is a machine, they don’t like it? There’s an immense lack of logic in this.
unique link to this extract

---

The monthly “AI or real” quiz • BBC Bitesize

»

A deepfake video is a computer-generated copy that aims to perfectly replicate the voice, look, expressions and even gestures of another person. They can look very realistic, as you’ll see from this month’s quiz!

See if you can get full marks – and remember, always double check where an image or video has orginally come from. Who posted it and what are the intentions behind it? Verify sources and check whether any trustworthy news sites have published the content.

«

I got 6/8, so tolerable? It’s generally the fingers, isn’t it. Though not always.
unique link to this extract

---

About half of Bandcamp employees have been laid off • The Verge

Ash Parrish:

»

One of the worst tech labor years ever continues with the news that roughly half of Bandcamp employees have been laid off. Bandcamp employees are reporting the news via social media.

Epic Games bought the indie music platform back in 2022 for an undisclosed amount before selling it barely a year later.

Late last month, Epic Games laid off 16% of its workforce, or 830 employees, due to what CEO Tim Sweeney described as overspending. Epic also revealed that it would sell the Bandcamp business to California-based music licensing company Songtradr. In that announcement, Epic disclosed that an additional 250 people would be leaving Epic either through receiving offers from Songtradr or Epic’s divesture from its SuperAwesome ad business. Employees who did not receive offers from Songtradr were notified today and will be eligible for severance.

In an email to The Verge, Songtradr confirmed that 50% of Bandcamp employees have been extended offers to join Songtradr and reaffirmed from a previous statement the company’s commitment to keeping the Bandcamp experience the same.

«

Totally puzzling why Epic bought Bandcamp at all. Not an obvious or even unobvious fit. And Epic has form: in 2019 it bought group video chat app Houseparty, then shut it down in 2021.
unique link to this extract

---

The negative impact of mobile-first web design on desktop • Nielsen Norman Group

Kim Salazar, Tim Neusesser and Nishi Chitale:

»

Over 55% of worldwide web traffic comes from mobile devices. As the use of mobile devices continues to grow, so does the importance of mobile designs. This is why the mobile-first design approach, where designers first design for mobile and then adjust the design for tablets and desktop, has become very popular.

A minimalist style eliminates any design elements that are not required for the core functionality or message of the website. To avoid a cluttered appearance and achieve a minimalist aesthetic, designers make heavy use of negative space and simplify the amount of content displayed in one screen viewport. The result is content dispersion: a long page with low information density.

In the past, slow internet speeds limited the use of high-resolution imagery on websites. Today, however, improved bandwidth has eliminated this design constraint. Many modern web pages use large, high-resolution imagery. Image-focused designs are eye-catching, but they result in overly dispersed text content. 

A few large images on a desktop page can work well. However, if the page has too many big images, the text-based content that is displayed inline or in between images becomes dispersed and fragmented across the page.

«

NN/G is certainly right that designs which don’t really adjust for the desktop can be annoying as hell. But there’s the alternative where desktops become cluttered as hell. Mobile has generally been a good influence on web design. It just needs to swing back a bit.
unique link to this extract

---

AI and The Bible: seraphim, cherubim and thrones • MLearning.ai

Samuele:

»

The angelic tradition has always intrigued me. Both because my family has a strong spiritual connotation, and because the traditional representation of angels is very different from their textual description. What better argument to test how biblical angels look to an Artificial Intelligence?
Today I focus on the first hierarchy of angels, the most important and closest to divinity. It consists of 3 orders:
• Seraphim — שרף
• Cherubim — כְּרוּב
• Thrones — θρόνος

So I start with the Seraphim, the angels closest to God, the ones most important in their closeness to the creator. But how are they described in the Bible? In these words of Isaiah:

»

each one had six wings; with twain he covered his face, and with twain he covered his feet, and with twain he did fly

«

If I pass this description to my AI, I get…

«

…well, he gets something that looks to me like a Dementor in the Harry Potter films, and then refines it a bit to get something a lot more scary. This was a couple of years ago, when AI weren’t so good at drawing. They’re still pretty impressive, though.
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