Tom Weber
Been testing AI chat tools since GPT-2 was the state of the art. Writes about what actually works instead of what the marketing pages say.
Experience
Tom spent a decade in engineering before pivoting to consumer-AI reviews in 2023. His work on conversational models and chatbot evaluation informs the testing rubric used across every Synthlust review. He's personally run more than 800 hours of paid testing across the AI companion category.
Beats
AI companion apps (the focus of this site), consumer chatbots generally, prompt engineering, model behavior under extended use. He's particularly interested in memory systems, content filters that don't work, and companies that treat user data as a revenue stream.
Testing approach
Every product is tested for a minimum of 30 days. A standardized set of prompts and edge cases runs against each product. Scores follow the public methodology. Tom does not let commercial relationships touch the editorial line: no company sees a review before it publishes, and no score is adjusted for commission.
Contact
Pitches, corrections, and legitimate questions welcome at [email protected]. Press and partnerships: [email protected].
Recent reviews by Tom
