Picture a split screen.
On one side, a lecture hall built in tiers. A tenured professor down in the well, blackboards behind him that roll over each other in layers so he can draw for hours if he wants. Up in the seats, fifty to two hundred people, laptops open, pens and pencils and iPads out, everybody figuring out the best way to take notes. You can smell that room, can't you. It smells like books and like something you can't quite name. It just smells like school. I got my master's in 2024, and I can tell you the newest building we occupied smelled exactly like the oldest one. That doesn't change.
Now flip to the other side of the screen. It's 2am. You're plowing through the assignments from that very large class, and the thing you turn in tomorrow is supposed to reflect the actual state of your learning. And there's nobody. Nobody to count on at that hour to help you navigate the one deep thought you can't tell is valid, or the one problem you can't figure out why you can't solve. Just you and the plow.
Sit in that a second. A lot of us can smell it. When I was in my program I spent a lot of time in libraries, honestly just because at home I didn't have that smell. But most of the studying was solitude. Alone. What I carry from it isn't the solitude. It's the togetherness with my colleagues, and the connection with my best professors.
The two people who saw something
Two professors in particular, both PhDs. Whether they liked me or not, I could tell they were able to see something in me. And because they could see it, they were able to pull more out of me. That's the life guide. Not the pastor who has to be kind to you, not your mom's best friend. Somebody who cares enough about what they hold and value that they want to share it, and they want you to be excited about it. They want you changed by it.
That's the before. Hold onto it, because it's about to get radically changed.
The Stone: AI didn't crash the classroom, it got hired
Here's what's clearer every week. AI isn't sneaking into the class anymore. Kids aren't lying about using it to dodge the contract, the one I had to sign for my master's that said under no circumstances would I ever even look at AI, let alone submit anything touched by it. That era is over.
AI has a chair at the faculty table now. Literally. It's not the gadget we smuggle in. It's a colleague the school hired to manage us. It delivers the lecture. It grades the work. And more than anything, and I do mean this, it answers your question at 2am.
So here's the punchline. The professor's knowledge is downloadable. And I don't think that was ever a secret, not even early on. The answers are strong. Sometimes you vet them and push back a little, but the patterns a professor puts out there are widely known patterns, and AI is good at those. The part that survives, from back when I still had real professors and no AI, is the mentorship. The feeling that I had to be brave enough to let somebody shape me. To find the teachers I admired most, whose shape looked like something worth aiming toward.
AI didn't crash the classroom. It just got hired. So what do we do about it?
The receipts
I'm not arguing either way here. I have genuinely mixed feelings, so let me just lay the floor down flat.
The Chronicle of Higher Education, in a news feature literally titled "AI Has Joined the Faculty," ran this in November 2025. The University of Michigan's Ross School of Business, working with Google Public Sector, launched an actual virtual teaching assistant. Not a search box. A TA, the role and not just the tool. Arizona State University, through its Enterprise Technology office, gave every student, professor, and researcher ChatGPT Edu access, scaling an entire public university at once. And eCampus News, in a trade report on the rise of AI-native universities, described places now built around AI from the ground up instead of bolting it on after the fact.
The pattern, said once: this isn't creep. It's hiring. The machine has a job title now.
And I don't want us to file the people who spent all those years learning how to teach under "another casualty of AI eliminating jobs." They're part of a social fabric we've always had. The speed at which we're losing that is alarmingly unsettling.
Blade one: the bridge is not the arrival
I bring this up a lot. I feel strongly that we owe it to ourselves to look at AI as a bridge to a richer way of being together and knowing each other, not a substitute for it.
A TA that answers me at 2am is a real bridge. It's getting me the knowledge I need to reach the next step, so a night that's going to run far too long can at least be productive. And there's nothing bad about that. Nobody else is answering me at that hour. Nobody else is there saying I'll be your mentor. Picture the kid alone at midnight on a problem set, and now a patient little thing that never sighs at a dumb question is sitting on a rectangle right next to him, answering.
The bridge is not the enemy. I really mean that. The trap is mistaking the bridge for an arrival and living on it forever. If you only ask the machine, you never cross over to the human. And the human was the destination the whole time. If we stop thinking like that, we stop understanding how to interact at all. If the loneliness epidemic has taught us anything, it's that the place we cultivate holistic, communal, connective living is not the rectangle. Extend that to education and the growth education is supposed to give, and you're making it much harder.
And here's the thing about the problem machine. It loves a permanent bridge. AI is not built to talk you into leaving it in ten minutes so you can go spoon with your lover at 2am. It's built to poke holes in what it just told you so you ask more, to take what you say you know and refine it four hundred and fifty more times until four or five hours quietly vanish. This is social media addiction on steroids.
Blade two: the audacity coda
The machine is closing the knowledge gap. It is not closing the courage gap or the identity gap.
The whole point of a real teacher was that a human looked at your half-formed work and your half-formed self and said, "More, you're closer than you think," or, "No, try again, I know you have it in you." You can't download being believed in. Same way you can't download a relationship that's intimate and actually cares about you. You can't download the nerve it takes to hand your unfinished self over and let somebody else shape it. That's just not a thing.
Knowledge is downloadable. Being seen is not. The knowledge was never the point. It was the excuse to put you in a room with somebody who could change you, back when you were the one looking for the change in yourself.
The honest part, because I'm not the oracle
I'll be straight with you. I don't know how possible it is to keep the good and cut the bad. I don't know if we can.
It's not "ban the TA." That's cheap, and it abandons the kid at midnight. We didn't have TAs in my program, but we had adjuncts who weren't great. Kind, hardworking people with day careers who came in at night to teach us, but they weren't trained to teach, and they were doing their best with a wide range of intellects and challenges walking into the room. Honestly, for some of my classes a really good AI TA might have taught me a lot more. So banning it is too easy. And it's not "the machine is the teacher now," because that's bullshit. We need to be seen. We need to be known. Somebody who saw us at the start of the program needs to watch us put the tassel on the other side at the end of it.
And here's the uncomfortable part. The campus was already drifting into the room of one before AI showed up. I had chances to work with my classmates, and I know, from direct experience, that I wasn't an equal contributor in both directions. At the beginning of my program I was a freak, a hardworking freak. By the end I was isolated and depressed, I'd lost an important relationship, and I fucked up. In both versions, my instinct was to put in the time I had to with my teammates and then go be super productive alone, or go wallow alone.
So AI isn't driving everybody into an education of one. That was already happening. What it did was build the most frictionless version of self-learning ever made, and dare us to admit we like it, because we get to learn only what we tell it we want and never have to agree with a room full of people. We can just do what we want.
So the real question isn't what AI does to the university. It's what we were secretly hoping AI would let us stop doing. Showing up. Being seen. Being together. Being changed.
And for the record: I use AI every single day, more than ten to fifteen hours, seven days a week, for my work. I learn so much from it. There's no way I could have learned what I've learned over the last eight months without it as my sidekick. I want to be really clear about that. I still don't know the right answer.
The Maelstrom
Headlines, sorted. Fuck yes, fuck no, fuck maybe.
Fuck yes. How many times have you heard that recycling is bullshit, that we can't do anything with it? One of the real impediments is sortation, because one contaminant in a big vat can send the whole thing to the landfill anyway. Well, in Portsmouth, Virginia, an AI-run recycling plant came online this April, reported by WHRO Public Media. Colorado's AMP Robotics uses machine-learning cameras to claw recyclables back out of the mixed trash under a $450 million regional deal with the Southeastern Public Service Authority. The plant runs 108,000 tons a year, admittedly only about a quarter of everything collected, but a quarter sorted accurately is remarkable. The contract guarantees at least half stays out of the landfill, and it keeps the dump alive toward the end of the century, dumps that would otherwise turn over closer to every decade. We're not shipping it thousands of miles to some beach in another country. This is AI pointed exactly right, at the filthy work no human should have to do by hand. Fuck yes.
Fuck no. OpenAI did a cover-up. In the ChatGPT copyright case, reported by TechCrunch in July 2026, the New York Times and other publishers filed a motion for sanctions. They allege OpenAI deleted billions of ChatGPT outputs after the suit was filed, in violation of the court's preservation order that said they couldn't delete anything. They allege it hid the tools and datasets that could have identified copyrighted journalism inside the outputs. And what it didn't delete, it over-redacted until the court itself called the sample unusable. This isn't about whether AI can read the news. It's about a company allegedly torching the evidence of how it was built. There's no upside for a human life in a cover-up. Fuck no. And OpenAI, shame on you.
Fuck maybe. And I really mean maybe here, because I'm not chiming in on this one. Anthropic, a company that hasn't even had a third birthday, is targeting a $2 trillion IPO valuation, per reports from PYMNTS and Quartz. On the one hand, capital at that scale could genuinely fund useful research and infrastructure at civilization scale. Big money can build big good, it really can. On the other hand, $2 trillion on a lab that barely has a couple birthdays smells like a bubble, like hype rewarded over whether any of this serves a single human being. We know it serves some. It's helping me. But I'm not the center of the universe. So I'm framing it, I'm not calling it. That's yours. Go to the Ledger on goinghumanpod.com, find it through the hamburger menu, and vote. Your thinking matters, and your vote can promote or demote it based on what the community says.
Tsunami time
Here's the close. If one of us does this, it's a drop of water in a bucket. If millions of us do it, that bucket overflows everywhere and we get a tsunami.
So here's the one for this week. Next time you go out to dinner, or grab a coffee somewhere you know people are on minimum wage, likely working more than one job, barely making ends meet, and paid to smile at you and take care of you, give a big fucking tip. A big one. Not twenty percent. Skip your Starbucks three or four times that week so you can leave a twenty on a ten-dollar tab. And leave a handwritten note that just says thank you, you're appreciated.
Not a text. Not a five-star tap in an app. Pen, paper, a stranger's face, and a gesture that lets them know you see them. Because the whole episode was about being seen, the one thing the machine can't hand you. Anybody can tap a rating into a number. It takes a sliver of nerve to write your gratitude down and hand it to a person. So go see somebody the machine would've turned into a number.
That's the show. I'm so glad you came back, and I hope you'll tell your friends about me. Like and subscribe so it lands in your feed and you can share it out. Comments and votes live at goinghumanpod.com, a great way to be part of the community and watch your votes promote or demote a bucket you may or may not have picked. Thank you so much. I'm Moss. This is Going Human. Because robots can't.