Questions, corrections, and useful collaborations.
The shortest route is email. Context helps me understand whether I can contribute.
How to reach me
- Email[email protected]
- GitHubgithub.com/simossss
- Xx.com/theotherpomp
Good reasons to write
I welcome specific corrections to published work, requests for a source or method that is unclear, thoughtful questions about an experiment, and collaboration ideas connected to practical agent systems. The strongest inquiries usually name the page or system involved, explain what you are trying to learn or build, and say what kind of response would be useful.
Punkjazz.ai is especially interested in agent workflows for serious non-developer users, local and open-model systems, durable memory and recovery, research automation, evidence and verification, human authority boundaries, and interfaces that make complex agent behavior understandable. A concrete problem, artifact, benchmark, or reproducible observation is much easier to engage with than a broad request to “talk about AI.”
What to expect
This is a personal research site, not a staffed support desk, press office, or hosted product. I cannot promise a response or a particular turnaround time. I do read concise, relevant messages and appreciate factual corrections even when no reply is needed. Please do not send passwords, API keys, private datasets, confidential company material, health records, financial records, or anything else that should not travel by ordinary email.
If your message concerns a factual error, include the exact URL, the sentence or figure at issue, and a primary source when possible. If it concerns security or privacy, describe the affected public URL and impact without including exploit payloads or personal data. The privacy page explains what this site itself collects and what its infrastructure providers may process.