John Palmer's viral 'Something Small Is Coming' parody captures the AI hype cycle perfectly
Feb 12, 2026
Key Points
- John Palmer, a Stripe employee, published a viral parody arguing that AI breakthroughs in late 2025 and early 2026 represent a genuine inflection point comparable to food delivery's adoption a decade ago.
- Palmer documents rapid capability acceleration: AI moved from failing basic math in 2022 to passing the bar exam in 2023, writing working software by 2024, and handling most coding work by early 2026.
- Palmer argues early adopters gain competitive advantage by subscribing to paid AI tiers and assigning real work to the tools weeks before colleagues recognize their value.
Summary
John Palmer, a Stripe employee and crypto observer, published a viral parody titled 'Something Small Is Coming' that captures the current AI hype cycle. Palmer argues that AI breakthroughs in late 2025 and early 2026 represent a genuine inflection point comparable to the food delivery boom a decade ago, marking a shift from steady improvement to sudden, visible capability jumps.
On February 5, 2026, two major AI labs released new models on the same day. Palmer treats this as a watershed moment. Some engineers now describe their work to AI, leave for hours, and return to completed code.
Palmer addresses the standard objection that AI isn't that good by noting current models are unrecognizable from those available six months earlier. Skeptics often aren't using paid tiers, which separates toy use from production capability.
He sketches the capability progression bluntly. In 2022, AI couldn't do basic math. By 2023 it passed the bar exam. By 2024 it wrote working software. By 2025, top engineers had handed most coding work to AI. By February 2026, you can have AI send you a Reddit summary from your Mac Mini.
His recommendations are tongue-in-cheek but substantive. Buy a Mac Mini—the hardware every serious user now owns. Subscribe to paid tiers. Give AI real work: a GitHub repo for engineers, a messy spreadsheet for finance professionals. Early adoption by even non-experts will create competitive advantage before colleagues notice. He also suggests reconsidering career advice for children, given that the "good grades, good college, stable job" playbook may no longer hold.
Palmer's credibility comes partly from acknowledging his own peripheral position. He's not a researcher. He bought a Mac Mini but hasn't set it up yet. He was at a birthday dinner when the February 5 models dropped. He treats proximity to people experiencing the change as sufficient evidence that something real is happening. His bet is that early adoption, even by non-experts, will prove more valuable than waiting for certainty.