Of course that’s your contention. You just got finished reading about polysemanticity, so naturally you think interpretability means untangling superposition. You’re gonna believe that until next semester, when you find a neuron that lights up for the Eiffel Tower. Then you’re gonna spend the summer explaining how we can align a superintelligence because we know when it’s thinking about France. And next year you’ll realize that knowing it recognizes the Eiffel Tower doesn’t tell you what the fuck it’s planning.

when you experience a language model doing a task that used to require you at every step and now doesn't, it's a gigantic agency rupture. we tend to equate ourselves with our tasks, our outputs, and a rupture of this kind is akin to the loss of your identity and sense of self. you are no longer required, and you had no say in the matter. it's a kind of death. however, agency ruptures with AI seem to follow a predictable pattern: - Initial agency rupture. Here you see the AI doing what you used to do, and you see only the AI. In your perception of reality the AI is highlighted, and the human work around getting the AI to do the thing, is invisible. you can see this in the way we use language: "AI solved XYZ Erdos Problem" is another way to say we're in the initial rupture phase. - Seeing human scaffolding. After a little bit of experience with the model and its new capibility—programming, writing, math, etc—we start to see the edges of what's possible with the model. Here, instead of just seeing the AI doing the task, we start to see the scaffolding required by you and other humans on either end to make sure the model does its work. At this stage, the scaffolding can feel like a secondary part of the work, but it is work nonetheless. We begin to say things like, "The AI is prompting me", or "My job is just to babysit Claude." - Agency reconstruction. After a long enough time with the model at its current level of capability, we start to reconstruct a new sense of agency that centers our own role in the work being done, and makes the model's contribution mostly invisible. It is now a tool, it fast-forwards you through the boring or repetitive parts of the work, and what used to seem like scaffolding now seems like the actual work itself—the interesting part. You've had enough experience with the scaffolding to see its nuances, and how hard it is to get the model to reliably output high quality work. (As part of this, your conception of what quality work is changes—the floor is raised, and so is the ceiling.) You can tell you're at this point because you stop saying "The AI did this" and just say, "I did this" and the fact that you used AI is implied. For example, it would be weird for me to talk to someone @every and ask them if the AI built the most recent feature of a product. Of course it did! But we're so used to the capability that it becomes invisible, and we re-center ourselves and our own conception of agency. My theory is that the ability to metabolize agency ruptures and turn them into playfulness, and curiosity is a good leading indicator for whether you're a person who likes and excels in the new AI economy or not. Each time capabilities jump, you get new ruptures for existing fields touched by AI. And you get new ruptures for naive populations (like mathemeticians) who until recently hadn't been affected. if you know what the cycle looks like, it makes it much easier to ride it without freaking out. i think most people @every are naturally good at this

We’ve automated every single thing we can @every with AI agents. And yet there’s way more human work to do than ever. We’ve gone from 4 -> 30 human employees since GPT-3. I wrote a report on the structural reasons: how AI makes expert competence cheap, why that drives up demand for experts, and why the dynamic only intensifies as we approach AGI. After Automation: every.to/p/after-automation

A Day at the Center of the AI Boom In San Francisco, there’s one rule to live by: layer up San Francisco's climate always makes me feel like I have a fever. It's warm and pleasant in the sun, but if you have the misfortune of walking into the shadow of a building you're suddenly shivering like you've got the flu. So I've come to follow a singular rule for visiting the tech capital of the world: Layer up. Once you have wrapped yourself to the gills in Patagonia paraphernalia, as I now do, you can truly experience the city—its delights, terrors, and idiosyncrasies. The East Coast jet lag means you can wake up easily at 6 a.m. You can take a Lyft to @Sightglass coffee, where the air is redolent with the smell of beans roasted by master coffee makers. You can drink your latte in the rose-colored San Francisco morning as a cable car trundles by. You can’t walk around; you’re not sure if SOMA is dangerous at this hour. You can experience the electric atmosphere at OpenAI Dev Day, hob-nob with the celebrities of your field, get excited about the future, and do off-the-cuff food reviews. You can jump at the chance to get a gray @OpenAI sweater in a size medium as an extra buffer against the vicissitudes of the city’s mercurial clime. You can visit the Archimedes Banya, where men and women roam naked together between saunas, steam rooms, hot tubs, and cold tubs. You can listen to the owner, completely derobed, explain that the sauna you are sitting in is the best in the world. You can sit, wearing a bathing suit, as New Yorkers do, on a hot cedar bench and marvel at the hidden complexities of sauna design, of the skill it requires to balance the contrasting forces of temperature and humidity. You can find that another person sitting next to you in the sauna, as naked and unashamed as Adam before the fall, is someone who you have followed on Twitter for 10 years. You can say hello, and ask how he is. You can be invited to dinner in a private room at a sushi restaurant. You can muse aloud between bites of uni about ideas for ChatGPT posts you want to write and projects you want to build. You can get feedback from people who have forgotten more about machine learning and business building than you will ever know. You can leave dinner feeling like you cannot wait to get home and write. You can instead find yourself at a house party. You can lovingly admire the large bookshelf on display, relieved that even in the center of our techno-utopian AI-enabled future, people still love books. You can observe a man in a blue velvet suit ask a French woman with her brown hair in a tight bun: “Are you an AI person?” “Yes, but I think of AI as merely a tool to have an impact,” she says. “What are you trying to have an impact on?” he says. “Helping enterprises better serve their customers,” she says. You can laugh quietly to yourself. You can turn to leave and sheepishly run into friends from New York. You can share an unspoken sense of embarrassment to be here, which is to admit that maybe, in at least this one dimension, New York is not the best city in the world. You can sit on the stoop of this house party and reflect that it is not just the weather in San Francisco that runs hot in the sun and cold in the shade. It is also the town itself. It is the sumptuously appointed demo days held under warm lights while a homeless man is doubled over on the dark pavement outside. This, you realize, is true—and troubling. But it is not the whole story. Because what San Francisco is is largely a matter of what it will be. Walt Whitman wrote, “America…counts…for her justification and success…almost entirely on the future. . . For our New World I consider far less important for what it has done, or what it is, than for results to come.” San Francisco, too, counts for its justification entirely on the future. It is a synecdoche for America. The future is what is important. This is its sin and its charm, its peril and its undeniable promise. In San Francisco, the future, too, is hot and cold. When the next round closes, the sun is on your back. When the acquihire falls through, the shade is cold as ice. On Monday, AGI and universal prosperity are at hand. On Tuesday, apocalyptic doom looms, or worse, stagnation—progress’s dreaded asymptote. Feast or famine. Fever is the nature of the future. To see rightly in this town is to try to layer yourself against the vicissitudes of what comes next. To at once be swept up in the hope that everything might change, but also hold on to the knowledge that it will all stay—in some important ways—the same. To know that you overestimate what can change a year down the line, and underestimate what can change in 10. It is to hope for the sun and prepare for the shadow. It is to dream of making a new sun one day, and feeling like it might be possible. It’s a great time to be a founder. Layer up.

We're spinning @lexdotpage out of @Every! @nbaschez is now the CEO of Lex, and he's raised a $2.75mm seed round for it from True Ventures. You can read the fundraising announcement here: