The plan
On July 1 I start plz don't kill us, a month-long creator bootcamp in Berkeley run by people worried about AI. I'm getting there by bike, with a train hop over the Plains and another over the Nevada desert. Filming and posting the whole way. And one question for everyone I meet:
"In 10 years, AI will ______."
Farmers, railroad workers, casino dealers, rationalists. Same question, 2,000 miles apart.
The route
| Jun 11–16 | Chicago → Mt. Pleasant, IA bike · 260 miPrairie Path, Fox River, I&M Canal towpath, cross the Mississippi at Burlington. |
| Jun 16 | Mt. Pleasant → Denver zephyrOvernight. Bike rides in the baggage car. |
| Jun 18–22 | Denver → Glenwood Springs bike · 220 miBerthoud Pass (11,307'), Colorado River headwaters, Trough Road gravel, Glenwood Canyon. |
| Jun 22 | Glenwood → Salt Lake City zephyrRuby Canyon and the Utah desert. |
| Jun 23–24 | Salt Lake City restFamily, laundry, bike check. |
| Jun 24 | Salt Lake City → Reno zephyrOvernight. Nevada in June is the train's job. |
| Jun 25–28 | Reno → Berkeley bike · 235 miTruckee River, Donner Pass on old Highway 40, American River Parkway, Davis, Benicia Bridge. |
| Jun 28 | Berkeley arriveRoll straight to the bootcamp's door. |
Why
AI is getting really good, fast. Most people aren't being told what's coming, and there are big problems we need to figure out together:
- Misuse — what happens when anyone can ask for anything.
- Concentration of power — a handful of companies and governments holding the steering wheel.
- Alignment — making sure the thing we build actually does what we want.
- What it means to be human — when machines can do most of what we do.
People deserve to be in this conversation. So I'm riding through the country it's about, talking to everyone on the way.
The bootcamp at the end is called plz don't kill us — a month in Berkeley learning to make content about AI that normal people actually watch. The name is only half a joke.
Want to help?
Honest answer: I'm pretty uncertain what actually helps. But here's my best guess.
- Torchbearer — citizen advocacy for AI safety legislation. Follow or join.
- PauseAI — the "slow down until we know what we're doing" crowd.
- ControlAI — campaigning to keep AI under human control.
Not confident any of these are the answer — but they're people trying, and showing up beats spectating.
Or just get informed. Start here:
- Doom Debates — both sides, argued properly.
- Connor Leahy — on why humanity risks extinction from AGI.
- Roman Yampolskiy on Lex Fridman — the case that superintelligence can't be controlled.
- Tristan Harris at TED — the social-media guy on not repeating that mistake with AI.