Macrodoses #2

Accessibility grifters, that Ted Chiang essay, "founder mode," and NYT Games

Hal Triedman

Shira Abramovich

jacob sujin kuppermann

, and 2 others

Sep 07, 2024

Hello Reboot community! Welcome to our second issue of:

If you didn’t catch our first issue, Reboot Macrodoses are our editorial board’s way of sharing what we’re thinking about, though mini-essays, links, pictures, reader-submitted letters to the editor (submit one here), and more.

Speaking of the editorial board—last week we announced a new set of editors! As you will see from their future work (and their macrodoses below), they are sharp and smart and you should absolutely pitch them.

In today’s issue of macrodoses, we have:

Grift in the world of online accessibility
A response to that viral Ted Chiang essay about AI and art
“Founder mode” in economic context
The controversial, parasocial world of NYT Games
The murky history of “A computer can never be held accountable”
Engineers leave the Bay Area (no, not that one) for greener pastures
Reboot personals

Hope you enjoy!

— Hal

Grift Gossip

This week, I went down a rabbit hole with the publication of this article on IGN about Susan Banks and two other relatively well-known accessibility advocates, all of whom were queer Deaf women of color. Before she passed away in 2019, Banks had been a prominent blogger on accessibility issues for Deaf and mentally ill gamers, such as content warnings and comprehensive subtitles. Banks was so well known, in fact, that she had a gaming accessibility advocacy award named after her.

But this article makes a convincing case that Susan Banks actually never existed at all. Instead, her partner Coty Craven (nominative determinism?) invented her in order to gain clout for his own work as an accessibility advocate, and even went on to make up two more personas after she “died.”

Now that the grift has been exposed, Banks reads to me like a list of sympathy checkboxes. Refugee? Check. Successful model? Check. Queer? Check. Long list of invisible disabilities that are often ignored? Check. Some details are even wilder—for instance, she allegedly suffered a double amputation after a freak KitchenAid accident. But when accessibility awareness is such an uphill battle, who would question someone who was writing about the issue? It’s a testament to how little effective advocacy there is that this grift slipped under the radar for so long, across three different personas. And as Deaf women of color with many invisible disabilities, Craven’s fake personas were exactly the demographic most commonly accused of “faking.” Who would want to repeat those allegations, even if they held water?

I’m not mad at the people who believed Craven—it just saddens me that stories like these can damage credibility for disabled gamers, who already have a hard time advocating for themselves. What this story exposes, to me, is not just a shortage of advocacy but also how hard it is to get fixes for usability and accessibility issues. We shouldn’t need someone pretending that their fake partner dropped a KitchenAid on their foot in order to get clout for better subtitles.

— Shira

Choices, Choices, Choices

A reported 15 billion images have been generated with AI, though science fiction author Ted Chiang argued in The New Yorker last week that none of them counts as art. Chiang attributes AI-generated media’s disqualification as art to a lack of choice-making required of a user prompting ChatGPT or DALL-E to produce words or images. He adds nuance to an oft-repeated analogy between generative AI and the advent of photography, noting that “over time people realized that there were a vast number of things you could do with cameras, and the artistry lies in the many choices that a photographer makes.”

While I agree with Chiang that this is not what's happening in most Midjourney renders, AI can also be a new medium that expands the range of choices available to artists. Some of the most interesting artists working with AI as a medium are:

Kat Zhang, aka the Poet Engineer, who is impressively prolific at turning out captivating experiments with diverse configurations of AI models and other tools, including using gesture to control Stable Diffusion generations, visualizing how large language models (LLMs) embed color, and evolving fictional creatures.
Delta.Ark, who constructs massive virtual worlds in Unreal Engine that are populated by characters with LLMs in their “brains,” enabling them to scan an open-ended environment and take actions in pursuit of interesting-ness.
Sebastian Schmieg and Lina Schwarzenberg, who invite the public to upload photos of scenes they deem to be in need of alteration. In Twitch-streamed performances titled “The World Is Beautiful Again,” the duo then use AI image generation tools to “repair” the submitted images in real time—construction sites are beautified with trees, snarled masses of cables are managed, and shoddy manicures are embellished with flowers.
Disnovation.org and their Predictive Art Bot (2018), which uses natural language processing to extract ideas from recent news headlines and turn them into titles for nonexistent (though very plausible) artworks.

Approaches like these (or the experimental AI poetry my fellow editor Shira described beautifully for Reboot in 2021) involve constructing custom models or integrating existing ones within larger systems of tools—processes that certainly involve a great deal of choices.

— Hannah

Founder Mode

Much has been memed about Paul Graham’s “founder mode,” his mini-manifesto for founder-driven startup leadership over “professional fakers [who] drive the company into the ground.” But what’s missing from the discourse, and what I’ll offer, is a structural lens on the issue.

First, founder mode is not new: Brian Chesky has been giving versions of his how-I-saved-Airbnb talk to Y Combinator and others for over a year now. But I wouldn’t give Chesky all the credit either. Turn back the clock a couple more years, and you’ll find that Silicon Valley’s shift to flatter structures and more imperious leaders began with the end of zero-interest rates, the resulting tech layoffs, and Elon Musk’s reign in Twitter 2.0. After a decade of froth, overhiring, PR projects, and bloated management, it took a recession to force some focus. The worst of the layoffs have come and gone, so founder-CEOs are safely declaring their new management method a success. In this context, PG’s catchy new moniker is more of a victory lap than a movement’s beginning.

So what to make of it? First and obviously, founder mode does not mean that founders should be assholes to workers—even though that’s all some wannabes will hear. I’d also still expect more careful hiring and a less labor-friendly environment across the board.

The real insight here is that founders possess special instincts about their user base and problem space that most mercenaries won’t, and they should apply that knowledge directly to business problems, even if it means micromanaging. But I also think that smart employees at all levels can and should go founder mode. And founder-CEOs would be better off hiring (and delegating to) high-agency, mission-driven people who they trust to be in the weeds, rather than driving themselves (and everyone else) crazy reporting bugs at 2am.

— Jasmine

Missed Connections

Newspaper puzzles have always had a strong subculture (see the American Crossword Puzzle Tournament), but no game has managed to sustain a recurring viral Twitter discourse cycle like the NYT Connections. While Wordle and SpellingBee have similar reach, they are just letter guessing games. By contrast, Connections asks you to take 16 words and classify them into four groups, thereby leaving room for opinion on whether the grouping is sensible or valid. There’s a new puzzle every day, and no archives—like peak pandemic Wordle, everyone on a small slice of the internet is facing the same challenge at the same time, and this makes for perfect Twitter bait.

Some of the humor links current events to the daily puzzle. It also frequently dunks on people who don’t seem to get the point of the game. But the most interesting part is the parasocial relationship we have with Connections editor Wyna Liu. Check the search results for “Wyna” vs. “Will Shortz” (the long-time NYT Crossword editor) and you’ll see the difference in volume and popularity. I checked my Connections group chat, and we consistently refer to Wyna directly. More impressively, she’s silent on Twitter, so her name recognition doesn’t come from content creation under her own name.

I think the key virality behind Connections is that the groupings can be simply unhinged, and the level of difficulty comes at an unspecified rate. You can get normal stuff like synonyms, or bizarre things like “colors with their first letters changed”. The unpredictability and reference to today’s pop culture feels like a reflection of its online audience. It makes the puzzle feel individual rather than institutional.

Wyna, if you’re reading this, I hope you’re getting a raise because it seems like people are spending more time playing Games than reading the news.

— Shohini

A Computer Can Never Be Held

If you’ve gotten into enough online discussions about AI and capitalism you’ve probably seen it:

16 words, in stark all-caps sans serif. Usually identified as an IBM slide from 1979, the phrase “A computer can never be held accountable therefore a computer must never make a management decision” has become one of the common refrains of Twitter tech crit; someone goes viral reposting it every 10-20 months. Yet despite the phrase’s totemic status it’s surprisingly hard to verify much of anything at all about its provenance. The image in question was first posted in early 2017 by Twitter user (and former IBM employee) @bumblebike, who would later post a few more slides from the presentation in question (I personally think PUTTING A BAD SYSTEM ON LINE IS LIKE POURING GASOLINE INTO A FIRE is also very catchy) with later efforts stymied by a flood at the house the slides were stored at. Later efforts by users on Cohost and Mastodon revealed a few more lines via jpeg palimpsest analysis, but inquiries into the IBM archives as to any further details have proven unfruitful.

Personally, I find it fascinating that this slide has captivated so many people in spite (or perhaps because) of its unclear provenance. There are plenty of pieces of late 70s tech ephemera that haven’t caught on—the Computer History Museum alone has untold numbers of boxes full of them—but a quote shorn, perhaps forever, from its original context, allows for it to become something more than just a line in a presentation about computer systems but a rallying cry and prophetic claim.

— Jacob

A Tale of Two Bay Areas

When you hear "Bay Area expats," you’re probably thinking about tech workers and other professionals who leave the San Francisco Bay Area in search of a simpler, more fulfilling life. Well, there’s another Bay Area, and it’s in China, per Jacob Dreyer’s article for Noema covering the cross-pacific phenomenon.

Zhongshan is the "megapolitan": a bustling and industrious city like SF. On the outskirts, instead of Oakland and Berkeley, we have the Guangdong-Hong Kong-Macao Bay Area. Zhongshan, home to commuters, elite university graduates, and employees of prestigious companies, is met with passionless demand. Young people are looking for a simpler way of life, both in spirit and in their work. That’s the role of Qixi, where an increasing number of young corporate refugees are tending to tiny farms. Montessori schools, previously only a choice by retired hippie parents, are enrolled by the children of these “expatriates.”

The coder-turned-blueberry farmer fantasy is becoming real, and it's in China. It's interesting to see cottagecore take flight in a country with a vivid memory of its less prosperous past, where the workers who built its superpower status are now seeking refuge from the very state they helped create.

— Hamidah