Three models, three jobs
What I use today. Preferences, not a ranking, and they change.
- For coding
- OpenAI ChatGPT 6 Astra
- For daily work
- OpenAI Sol / Luna
- For design
- Anthropic Opus 5.5
Where the code gets written
Codex is my main builder: I describe what I want, it writes it, and I review the changes before anything ships. It built most of this site. Cursor is my second workspace, for when I want to go through the code myself.
Write down the task before you pick a model
What goes in, what a good result looks like, and who checks it. If I can’t write those three things down, no model is going to fix it. This is how every Cortana project starts.
- 1What goes inEmails, a spreadsheet, a call transcript
- 2What good looks likeA reply drafted, a row updated, a summary sent
- 3Who checks itThe person who owns the result signs off
Test tools on your own work, not a demo
Give each tool the same handful of real examples and compare quality, time, and cost. The newest release isn’t automatically the right one for your job.
| Tool | Quality | Time | Cost |
|---|---|---|---|
| Tool A | _ / 5 | _ min | $ _ |
| Tool B | _ / 5 | _ min | $ _ |
| Tool C | _ / 5 | _ min | $ _ |
The short version is on TikTok
Recent videos on the tools and models I use.
Latest from OpenAI
Pulled from OpenAI’s official news feed. Their headlines, not my commentary.
News
Latest from OpenAI
Pulled from OpenAI’s official news feed. Their headlines, not my commentary.
Checking the latest announcements…
Let’s talk about what you need
Two quick questions, then your details. It takes about a minute.
- What you want
- What it’s for
- Your details