MCG Live · November 24, 2025 · 42 min
Thomas Perkins, founder of the PCEF nonprofit, discusses THRT, a blockchain-based security token that puts AI training data in the public domain for peer review. He turned down a $2.7M acquisition offer, emphasizing his mission to provide free cybersecurity tools rather than profit; the project has burned 5.2M tokens and attracted 10,000 users in 2.5 months.
Thomas Perkins is a reverse engineer who built a malware analysis engine called Malor, which he sold to Internet 2.0, and later created a neural network that powers Tracix.
A venture capital firm offered $2.67M to purchase the THRT token, underlying code, and Tracix; Perkins declined because he had previously sold a product for millions without receiving the money and wanted to maintain control.
Tracix.com offers free cybersecurity tools where users drag-and-drop files to receive a verdict (unknown, safe, or malicious) and can decrypt and view the AI training data that informed the classification.
The token utility is paying for file uploads and publishing to IPFS, with most fees burned and some going to a development treasury; 10,000 users have signed up in 2.5 months with roughly 60 uploads to the public domain.
Perkins has burned 5.2M of a planned 10M token burn, with 4M left to burn; he is releasing an SDK this week for integration into other products, with a partnership to embed Tracix into a messaging app for file scanning.
Tracix currently holds 45M samples totaling 4.7 terabytes of training data; it achieves 91.64% accuracy on Windows and Linux files but only 3% on PDFs, so Perkins does not release features until accuracy exceeds 90%.
Perkins released 'Nosex 402' two weeks ago, a more secure version of HTTP 402 that hashes data before sending and locks it to that hash to prevent client-side tampering, with no fees for use.
The roadmap includes SDK release, Phantom wallet login integration to replace early HTTP 402, holder verification for token-gated requests, IOC export for enterprise tools, and multi-file upload functionality.
This summary was written from the recording's transcript by a language model. For the exact words, ask the MCG Live archive: every answer there is quoted word for word and linked to the second it was said. Market Bubble Search is independent and not affiliated with the show. Nothing here is advice.