[Hiring] # AI Science Communicator — Help Me Say the Important Things Consistently

Part-time

The Problem

I'm a tech exec — not one of the evil ones (hopefully). I've spent the last 20 years in AI, helping build some of the algorithms you know and love. Over the past year or two, I've been increasingly wrestling with the feeling that "we're missing something, aren't we?"

I've decided to do something about it and, although I've been partly successful in nudging the conversation when I've tried, I'm not able to be as consistent as I would like.

That's where you come in.

The Solution

I've always loved writing and have discovered over the years that I'm somewhat reasonable at it. When I write something, people read it and it (usually) lands. A recent post hit #2 on Hacker News. Posts I've written have cleared 250,000+ views in the first day. That is, it usually does work, but...

When I get busy (I am usually busy), things go dark for a month (or three) and the momentum evaporates. I don't need a ghostwriter to make me sound smart — I need someone who makes sure the things I'm already saying reach more people reliably, (not just occasionally when I scrounge up the time).

My writing voice is moderately irreverent, massively nerdy and improbably entertaining. Think 3Blue1Brown meets xkcd: the math is real and the metaphors are weird, but you walk away hopefully both entertained and (a little) smarter. I often lead with a claim that sounds obviously wrong, then work it out with numbers, data, and 20 years of hard-won lessons until "that can't be right" flips to "oh, damn, he's right." If you can engineer that flip in a stranger's head and make it fun, you already get the job.

The Role

Part editor, part educator, part keeper-of-the-cadence.

Translate & compress (~40%). Find the sharpest point in something long and make it travel — a note, a thread, a tight email opener. Not summarizing. Compression and teaching: arguing one idea well enough in 280 characters that a stranger wants the 2,000-word version.
Keep it moving (~30%). Own the calendar. Make sure I don't vanish for a month. Handle comments and DMs in a voice that sounds like me, not a brand account.
Read the signal (~15%). Track what lands and tell me what it means — "this format works, that one doesn't, here's what I'd change."
Make moments land (~15%). A big piece, a talk, a point of view I want to push — you build the content around it so it lands instead of passing.

The through-line: you can tell genuinely interesting from merely loud, and you see what needs to go out before I remember to ask.

Requirements

Demonstrated ability to break through the noise — if you haven't noticed, there's a lot going on out there...
Excellent writing and editing ability, especially for short form and technical content
Enough technical literacy to follow an argument about AI without getting lost (not a researcher; you keep up)
Experience running content for a person — founder, exec, creator, yourself — not just brand accounts.
Proven understanding of the nuances of Substack, X and email
Be a good, kind human - as a policy, I don't work with assholes
Bonus: Proven experience with science edu-tainment, e.g. SciShow, 3Blue1Brown

The Fine Print

Fully remote, async-friendly.
10–15 hrs/week, flexible on timing, serious about output.
No layers, no middlemen.
Compensation: $25–40/hr; I'm not trying to underpay someone good.

How to Apply

Send an email to hello@obviouslywrong.org with three things:

A couple sentences about why this job's for you. Not your résumé, why this.
A Substack note (≤100 words), ready to post, inspired by my essay The 9s of AI Reliability. The argument: AI's problem isn't intelligence, it's reliability — a model can be 97% accurate but only 70% reliable, and that gap stands between a demo and real ROI. Find the sharpest point. Say something real. In my voice, not yours.
A link to your LinkedIn (or attached resume)

Please keep it short! No need to write an essay. We're not trying to create makework for you, but also concision is half the job.