4 / 4 · 2025–nowEra 4 of 4 · 2025–now
When anyone can build
Code got close to free, so the hard part moved: what should you make, and can you prove it works? I'm building again, and I'm taking apart inference to see where it goes next.
meir@dick ~ $ #ShippedWithAI # Public build log, 2026
meir@dick ~ $ System3 Inference # Founder, 2026–. Callozum and the Adaptive Factory.
- #ShippedWithAI Public build log, 2026
- System3 Inference Founder, 2026–now. Callozum and the Adaptive Factory.
What changed
In December 2025 the cost of building fell through the floor. Since then I’ve shipped 18 products with coding agents, for clients and for myself.
The question moved from “can you write the code?” to “do you know what to build, and can you prove it works?”
How I work now
With a client, I find the bottleneck, build a working AI prototype in days, and let the prototype test the business case. No slide deck. I also logged my builds in public, as #ShippedWithAI.
The next bet
Cloudflare Connect in October 2026 set the direction. Inference is changing shape, from questions to streams, and something has to decide where each piece of thinking runs. That’s partly about speed and partly about rules: which data is allowed to leave a device or a country. And now that writing code is cheap, the job is proving that the new version is better.
For now this work lives under one working name, System3 Inference. Callozum decides where each part of an AI request or stream runs, and the Adaptive Factory proves a rebuild before it ships.
What I learned
Writing code got cheap, but proof didn’t, so I publish the numbers.
live15
in-progress5
shipped11
Write to me about the problem you want to fix.
- emailEmail contact@meirdick.com
- linkedinLinkedIn in/mdick
- githubGitHub meirdick
- substackSubstack aistrategyconsultant
- rssRSS /rss.xml