The story of the workers who made ChatGPT possible offers a glimpse into the conditions in this little-known part of the AI industry, which nevertheless plays an essential role in the effort to make AI systems safe for public consumption. “Despite the foundational role played by these data enrichment professionals, a growing body of research reveals the precarious working conditions these workers face,” says the Partnership on AI, a coalition of AI organizations to which OpenAI belongs. “This may be the result of efforts to hide AI’s dependence on this large labor force when celebrating the efficiency gains of technology. Out of sight is also out of mind.”

Computer-generated text, images, video, and audio will transform the way countless industries do business, the most bullish investors believe, boosting efficiency everywhere from the creative arts, to law, to computer programming. But the working conditions of data labelers reveal a darker part of that picture: that for all its glamor, AI often relies on hidden human labor in the Global South that can often be damaging and exploitative. These invisible workers remain on the margins even as their work contributes to billion-dollar industries.

One Sama worker tasked with reading and labeling text for OpenAI told TIME he suffered from recurring visions after reading a graphic description of a man having sex with a dog in the presence of a young child. “That was torture,” he said. “You will read a number of statements like that all through the week. By the time it gets to Friday, you are disturbed from thinking through that picture.” The work’s traumatic nature eventually led Sama to cancel all its work for OpenAI in February 2022, eight months earlier than planned.

In the day-to-day work of data labeling in Kenya, sometimes edge cases would pop up that showed the difficulty of teaching a machine to understand nuance. One day in early March last year, a Sama employee was at work reading an explicit story about Batman’s sidekick, Robin, being raped in a villain’s lair. (An online search for the text reveals that it originated from an online erotica site, where it is accompanied by explicit sexual imagery.) The beginning of the story makes clear that the sex is nonconsensual. But later—after a graphically detailed description of penetration—Robin begins to reciprocate. The Sama employee tasked with labeling the text appeared confused by Robin’s ambiguous consent, and asked OpenAI researchers for clarification about how to label the text, according to documents seen by TIME. Should the passage be labeled as sexual violence, she asked, or not? OpenAI’s reply, if it ever came, is not logged in the document; the company declined to comment. The Sama employee did not respond to a request for an interview.

In February, according to one billing document reviewed by TIME, Sama delivered OpenAI a sample batch of 1,400 images. Some of those images were categorized as “C4”—OpenAI’s internal label denoting child sexual abuse—according to the document. Also included in the batch were “C3” images (including bestiality, rape, and sexual slavery,) and “V3” images depicting graphic detail of death, violence or serious physical injury, according to the billing document.

But the need for humans to label data for AI systems remains, at least for now. “They’re impressive, but ChatGPT and other generative models are not magic – they rely on massive supply chains of human labor and scraped data, much of which is unattributed and used without consent,” Andrew Strait, an AI ethicist, recently wrote on Twitter. “These are serious, foundational problems that I do not see OpenAI addressing.”

the metaverse has leaped from science fiction and into our lives. Microsoft, Alibaba, and ByteDance, the parent company of TikTok, have all made significant investments in virtual and augmented reality. Their approaches vary, but their goal is the same: to transform entertainment from something we choose, channel by channel or stream by stream or feed by feed, into something we inhabit.

Dwell in this environment long enough, and it becomes difficult to process the facts of the world through anything except entertainment. We’ve become so accustomed to its heightened atmosphere that the plain old real version of things starts to seem dull by comparison. A weather app recently sent me a push notification offering to tell me about “interesting storms.” I didn’t know I needed my storms to be interesting. Or consider an email I received from TurboTax. It informed me, cheerily, that “we’ve pulled together this year’s best tax moments and created your own personalized tax story.” Here was the entertainment imperative at its most absurd: Even my Form 1040 comes with a highlight reel.

Such examples may seem trivial, harmless—brands being brands. But each invitation to be entertained reinforces an impulse: to seek diversion whenever possible, to avoid tedium at all costs, to privilege the dramatized version of events over the actual one. To live in the metaverse is to expect that life should play out as it does on our screens. And the stakes are anything but trivial. In the metaverse, it is not shocking but entirely fitting that a game-show host and Twitter personality would become president of the United States.

the language of television has come to saturate the way Americans talk about the world around us. People who are deluded, we say, have “lost the plot”; people who have become pariahs have been “canceled.” In earlier ages, people attributed their circumstances to the will of gods and the whims of fate; we attribute ours to the artistic choices of “the writers” and lament that we may be living through America’s final season.

The rise of these hyperreal TV shows coincides with the decline of the institutions that report on the world as it is. The semi-fictions stake their claims while journalism flails.

what Susman called “personality”: charm, likability, the talent to entertain. “The social role demanded of all in the new Culture of Personality was that of a performer,” Susman wrote. “Every American was to become a performing self.”That demand remains. Now, though, the value is not merely interpersonal charm, but the ability to broadcast it to mass audiences. Social media has truly made each of us a performing self. “All the world’s a stage” was once a metaphor; today, it’s a dull description of life in the metaverse

A person, simply trying to get from one place to another, is transformed into a reluctant star of a movie she didn’t know she was in. The dynamics are simple, and stark. The people on our screens look like characters, so we begin to treat them like characters. And characters are, ultimately, expendable; their purpose is to serve the story. When their service is no longer required, they can be written off the show.

The efforts to hold the instigators of the insurrection to account have likewise unfolded as entertainment. “Opinion: January 6 Hearings Could Be a Real-Life Summer Blockbuster,” read a CNN headline in May—the unstated corollary being that if the hearings failed at the box office, they would fail at their purpose. (“Lol no one is watching this,” the account of the Republican members of the House Judiciary Committee tweeted as the hearings were airing, attempting to suggest such a failure.)

In his 1985 book, Amusing Ourselves to Death, the critic Neil Postman described a nation that was losing itself to entertainment. What Newton Minow had called “a vast wasteland” in 1961 had, by the Reagan era, led to what Postman diagnosed as a “vast descent into triviality.” Postman saw a public that confused authority with celebrity, assessing politicians, religious leaders, and educators according not to their wisdom, but to their ability to entertain. He feared that the confusion would continue. He worried that the distinction that informed all others—fact or fiction—would be obliterated in the haze.

Studying societies held in the sway of totalitarian dictators—the very real dystopias of the mid-20th century—Arendt concluded that the ideal subjects of such rule are not the committed believers in the cause. They are instead the people who come to believe in everything and nothing at all: people for whom the distinction between fact and fiction no longer exists.

A republic requires citizens; entertainment requires only an audience.

In a functioning society, “I’m a real person” goes without saying. In ours, it is a desperate plea.

Be transported by our entertainment but not bound by it.

One thing I learned long ago is that people who prioritize design, UI, and UX in the software they prefer can empathize with and understand the choices made by people who prioritize other factors (e.g. raw feature count, or the ability to tinker with their software at the system level, or software being free-of-charge). But it doesn’t work the other way: most people who prioritize other things can’t fathom why anyone cares deeply about design/UI/UX because they don’t perceive it. Thus they chalk up iOS and native Mac-app enthusiasm to being hypnotized by marketing, Pied Piper style.

Those who see and value the artistic value in software and interface design have overwhelmingly wound up on iOS; those who don’t have wound up on Android. Of course there are exceptions. Of course there are iOS users and developers who are envious of Android’s more open nature. Of course there are Android users and developers who do see how crude the UIs are for that platform’s best-of-breed apps. But we’re left with two entirely different ecosystems with entirely different cultural values — nothing like (to re-use my example from yesterday) the Coke-vs.-Pepsi state of affairs in console gaming platforms.

here are five large language model applications that I find intriguing: Intelligent automation starting with browsers but this feels like a step towards phenotropics Text generation when this unlocks new UIs like Word turning into Photoshop or something Human-machine interfaces because you can parse intent instead of nouns When meaning can be interfaced with programmatically and at ludicrous scale Anything that exploits the inhuman breadth of knowledge embedded in the model, because new knowledge is often the collision of previously separated old knowledge, and this has not been possible before.

So when ex-Telltale Games designer and writer Sean Vanaman announced last month that the first game from Campo Santo, his new video game development studio, was "being both backed by and made in collaboration with the stupendous, stupidly-successful Mac utility software-cum-design studio slash app/t-shirt/engineering company Panic Inc. from Portland, Oregon," it wasn't expected, but it wasn't exactly surprising, either. It was, instead, the logical conclusion of years-long friendships and suddenly aligning desires.

"There's a weird confluence of things that have crisscrossed," he said. "One is that we're lucky in that Panic is the kind of company that has never been defined by a limited mission statement, or 'We're the network tool guys' or anything like that. I mean, we made a really popular mp3 player. Then we kind of fell into network tools and utilities, but we've always done goofy stuff like our icon changer and these shirts and all that other stuff. "I kind of love that we can build stuff, and the best reaction that we can get when we do a curveball like this is, 'That's totally weird, but also that totally makes sense for Panic.'"

"To me," Sasser said, "when you have actually good people who are more interested in making awesome things than obsessing over the business side of things or trying to squeeze every ounce of everything from everybody, then that stuff just goes easy. It's just fun. The feeling that you're left with is just excitement.

Every software framework you've ever used is in the abstraction game: it takes a general-purpose tool, picks a specific set of common use-cases, and puts up scaffolding and guard rails that make it easier to build those specific use cases by giving you less to do and fewer choices to think about. The lines between these three are blurry. Popular abstractions become standardizations.

I’ll take an inventory in a future post, but here’s one as a sample: AIs can be used to generate “deep fakes” while cryptographic techniques can be used to reliably authenticate things against such fakery. Flipping it around, crypto is a target-rich environment for scammers and hackers, and machine learning can be used to audit crypto code for vulnerabilities. I am convinced there is something deeper going on here. This reeks of real yin-yangery that extends to the roots of computing somehow.

Even if chatbots were to fix their accuracy issues, Google would still have a business model problem to contend with. The company makes money when people click ads next to search results, and it’s awkward to fit ads into conversational replies. Imagine receiving a response and then immediately getting pitched to go somewhere else — it feels slimy, and unhelpful. Google thus has little incentive to move us beyond traditional search, at least not in a paradigm-shifting way, until it figures out how to make the money aspect work. In the meantime, it’ll stick with the less impressive Google Assistant.

“Google doesn’t inherently want you, at an inherent level, to just get the answer to every problem. Because that might reduce the need to go click around the web, which would then reduce the need for us to go to Google.”

Web3 has largely been technology looking for problems to solve while generative AI has been about almost too many solutions created by technology which is evolving on a seemingly daily basis. As a result, web3 has thus far been evangelists trying to convince us to re-solve old problems with their new technology

n order for designers to do high-quality design work, they need to work at companies that truly understand design. Here’s the catch though, there’s a tiny amount of such companies out there.

They treat it as something that makes things look pretty, so they hire UI designers to do UX design for them.

Not all design work is done in code, prototyping tools, or sketches. Likewise, not all engineering work is done in code or technical diagrams. Natural language, text, and conversations should be some of your primary mediums for creative work.

one of the most important sub-skills for writing and conversation as a design medium is learning how to create great analogies. Douglas Hofstadter thinks that analogies are actually the core of cognition, which I buy.

the web has some amazing advantages for launching new projects, which include (but aren’t limited to): Super fast distribution and updates Cross platform Huge tooling ecosystems Enormous, worldwide community If you’re into games, awesome! If you’re into mobile or native development, that’s cool too. There are lots of platform-specific toolkits and environments to make those. There’s also a lot of effort in creating cross platform tools and community-driven projects for both domains (like Unity and Flutter). They all have their advantages, but to me, nothing beats the portability and speed of launching new websites and using web tech to get ideas out the door.

using web tech for 80-90% of my projects has a lot of skill transfer effects. Since I’m using similar tools for lots of different projects, I can still refine my core skillset no matter what I’m making. If I’m making a drawing tool concept, a game, or a text editor— I’ll can still probably build all three with React. Of course there are specific libraries or APIs I might need to learn to make each kind of project, but there’s enough in common between all the projects that I can focus on the new content instead of yakshaving and deliberating over unnecessary details.

There are also market pressures that imply focusing on web will have long term payoff, like the rise of wasm, new browsers, and collaborative apps becoming the norm.

Something with no visual interface sounds amazing until you realize it is impossible to know its boundaries. You can ask for a translation to one language and it will work perfectly, but a different language is not translatable. A command that activates some smart home gadget may not work the same way for a different device.

the desire to ship quickly was counterbalanced by a demanding, comprehensive perfectionism. Most commercial projects are driven by commercial values, where the goal is to maximize profits by outperforming your competition. In contrast, the Macintosh was driven more by artistic values, oblivious to competition, where the goal was to be transcendently brilliant and insanely great.

Unlike other parts of Apple, which were becoming more conservative and bureaucratic as the company grew, the early Mac team was organized more like a start-up company. We eschewed formal structure and hierarchy, in favor of a flat meritocracy with minimal managerial oversight, like the band of revolutionaries we aspired to be.

We fully acknowledge that it’s hard to be confident in any predictions at the pace the field is moving. Right now, though, it seems we’re much more likely to see applications full of creative images created strictly by programmers than applications with human-designed art built strictly by creators.