Want to chat?
I ship a lot of AI projects. People keep emailing me with questions: which model, which stack, how to scope it, why the thing they built is hallucinating. I used to try to reply to all of them. I got tired.
So this is how I work with people now. You book an hour on my calendar, bring a specific question, and I tell you what I actually think. No pitch, no upsell, no deck afterwards. Just the call.
What people usually bring to the call
An AI idea they're sizing up. Before they spend three months building. They want to know if it's tractable with current models, what the architecture might look like, where the unit economics break.
Something they've already half-built that isn't working. The RAG hallucinates. The agent loops. The bill is somehow $4k a month. They need a second pair of eyes.
A question about something specific in my portfolio. How does Kensaku handle dataset enrichment? How's Novel Translator priced? What would I do differently on KeqingMains if I started it today?
A straight-up choice. Claude or GPT for their use case. Workers or Vercel. Self-host the model or use an API. I don't always have the right answer, but I've almost always made the wrong one first, and that saves time.
What the call is not
It's not a discovery call to see if we'd be a fit for a bigger project. There is no bigger project. I don't take implementation work, I don't do retainers, I don't sell fractional-CTO time. The hour is the entire relationship.
It's also not a friendly chat. I charge because it filters for people who actually want my time, and because nobody values a free call. If you want to pick my brain for free, I'm a bad choice — the projects page has a lot of what I'd tell you anyway.
Pick a time
Google Meet. Paid upfront through cal.com.
Calendar below. If it doesn't load, open it in a new tab.
New tabNon-consult stuff goes to chris@doanythingai.com. I read it all. I don't always reply quickly.