The Long Game.
One big idea. No hedging.
Deep, opinionated essays on AI, the future of work, and what it actually means for the people building in it. Not a news digest. Not a listicle. One thesis, fully argued, written for founders, investors, and builders who think in decades, not news cycles.
"The pipeline is the proof.
The person is the product."
Think Economist.
Meets private letter.
Each issue starts with a single thesis: a claim about where AI is actually taking us. Not "here's what happened this week," but "here's why this matters over the next five years."
Then the argument is built: three or four supporting lines of evidence, a serious look at the counterargument, and a prediction at the end. No filler. No hedging. No listicles.
This is Aseem's personal POV, front and center, drawing from building a real AI media operation, a decade of enterprise tech, and a research background that includes HBR and MIT Tech Review.
One Thesis
Every issue defends a single, falsifiable claim about AI and the future.
Takes Stances
No "on the one hand, on the other hand." The Long Game picks a side and argues it.
Makes Predictions
Every essay ends with a concrete, time-bound prediction you can track.
Earned Perspective
Written by someone building AI systems daily, not observing from a distance.
The Archive
Six issues deep so far, published as time allows.
The Req That Never Got Opened
Meta's reversed AI layoff plan got a Reuters investigation. The entry-level hiring freeze that came with it didn't, and it'll still be true in five years, long after the layoff headline is forgotten.
Read essay →The Death of AI Scaffolding: What Happens When the Infrastructure Layer Disappears
RAG toolkits, orchestration frameworks, data connectors: the infrastructure layer built to make AI usable is dissolving as fast as it was funded. The problems it solved got smaller before the companies could grow into them.
Read essay →The Production-Progress Paradox: Why More AI Doesn't Mean Faster Science
Five hundred papers get indexed every hour, and real breakthroughs are getting harder to find inside them. AI didn't remove science's bottleneck; it moved it, from generating ideas to telling which ones are true.
Read essay →The Agent Economy Is Already Here: We're Just Too Scared to Let It Run
The infrastructure for autonomous AI agents exists today. The bottleneck isn't capability; it's the human fear of ceding control to systems we can't fully audit. That fear is rational, but it's also temporary.
Read essay →I Built a Media Company for $23/Month. Here's What That Actually Means.
A $23/month VPS runs three newsletters, five videos a day, and nine services, with exactly one manual step left: hitting publish. This is what a media company looks like when the labor is already automated.
Read essay →The Great Tokenization: Why AI's Economic Model Is Breaking Software
Every AI interaction is metered by the token. This isn't just a billing detail; it's a fundamentally new economic model for software that will restructure how we build, price, and consume technology.
Read essay →For people who play the long game.
The Founder
You're building in AI and need a compass, not a compass rose. You want someone who's actually shipping to tell you where the puck is going.
The Investor
You're placing bets on AI startups and want macro frameworks for thinking about what survives the hype cycle vs what becomes infrastructure.
The Senior Builder
You're 5–10 years into tech, you're tired of breathless coverage, and you want long-form analysis that respects your intelligence and your time.
Start thinking in decades,
not news cycles.
Read the full archive on overfitted.tech. No paywall, no sign-up. Or follow on LinkedIn to get each essay the moment it's published.