Independent AI research lab
The Lab is working.
BigSteve Labs builds, tests, and documents practical AI systems at the edge of agents, automation, and reliability. The work is public when it is ready. The Lab is not a demo.
Current experiment
Project BigSteve
Multiple independent AI builders. One specification. One clean WordPress staging environment. The strongest complete result becomes the foundation for the next BigSteveLabs.com.
This site is that experiment in public. Not a contest page. A live evaluation of whether a serious laboratory website can be built, owned, and maintained in WordPress.
01 / Systems
What we work on
AI agents, local assistants, reliability recovery, experimental infrastructure, and applied evaluation. Work is released when it survives contact with the real world.
02 / Evidence
What we publish
Lab reports, experiment records, field results, and limitations. Claims without evidence stay off the site.
03 / Access
How to follow
Research on Substack and Medium. Code on GitHub. Appearances when they matter. Inquiries when the work is relevant.
In the field
Projects, not promises.
MR BIG. Buzz Desk. Home AI Lab. Agent Rescue. Each is a real system under test — not a slide.
Lab report
Open Source & Developer Report #001
Public research on developer tooling, open source, and the conditions under which AI systems actually ship. RC1 is released.
Field results
Failure at 73%. Recovery completed.
An AI job failed under HTTP 429. The failure was identified. An alternate provider was selected. Human approval was required. Progress was preserved. The run completed.
Look enormous. Sound authoritative. Demonstrate serious work.
The Lab does not chase attention.
It creates things worthy of attention. Projects should feel like releases. Tests should feel like events. Results should feel valuable.
