Some Silicon Valley researchers have reportedly begun smoking cigarettes to demonstrate their confidence that AI will cure cancer. I respect the impulse, but it betrays a lack of ambition. Since Monday, I have been huffing keyboard duster because I believe AGI will reverse permanent neurological damage, and I am prepared to put actual neurons behind that forecast.

Smoking is a weak signal. The harm is probabilistic, spread over decades, and mitigated by existing medicine. Any founder can smoke a Marlboro, announce he's long AGI, and collect likes. I prefer a shorter feedback loop. My girlfriend says I've begun repeating questions she's just answered and recently tried to unlock our front door with a banana. I consider this independent validation of the experiment.

I know these products can cause brain damage and sudden cardiac death. Both are accounted for in my model. Brain damage is precisely what AI will solve, and mortality is a separate roadmap item currently tracking 2032. If you're comfortable accepting a 20% risk of human extinction to build AGI but unwilling to lose the ability to remember your dentist's name, I question whether you've internalized your own timelines.

My neurologist has been resistant. She told me some injuries are irreversible and asked me to stop using industrial aerosols. I sent her three preprints on neural tissue regeneration and a screenshot of an AI model scoring 87% on a biology benchmark. She has yet to offer a substantive rebuttal.

My coworkers remain supportive, though I've been moved off core infrastructure and given a whiteboard marker to hold during standup so I feel included. Our CTO called the practice "an HR issue." I told him every disruptive technology faces regulatory headwinds.

There are tradeoffs. I can no longer do long division, remember my ZIP code, or reliably distinguish our COO from the water cooler. But every lost capability expands the market for whoever eventually restores it. I'm not getting dumber. I'm growing the total addressable market.