NanoGPT on the THORChain Podcast
Milan joined the THORChain live stream podcast on June 13 for a long conversation about NanoGPT, crypto payments, privacy, AI access, and decentralized liquidity.
Full episode on X. Also on YouTube.
This was a fun one because there is a lot of overlap between what THORChain is trying to do and what we care about at NanoGPT: native assets, self-custody, fewer middlemen, and making useful tools available without asking users for a pile of personal information first.
Some of the main topics:
- How Milan first got into Bitcoin in 2012 through an online game
- Working on crypto policy at the Dutch central bank, and why that was both interesting and frustrating
- How NanoGPT started as a small Telegram bot where people could pay tiny amounts of Nano to use ChatGPT
- Why pay-as-you-go AI is often a better fit than monthly subscriptions
- How NanoGPT routes across many AI providers, and why Auto Model and Coding Router exist
- Why privacy matters for AI access, including TEE models, Private Mode, local models, and PII redaction
- What our monthly crypto payment stats show about which coins people actually spend
- Why THORChain matters to us as a way to accept more native assets without relying as much on centralized exchanges
- The weirdness of having an exchange account frozen over a normal small Dogecoin payment because of old UTXO history
- Whether Nano could eventually make sense on THORChain
The episode is long and informal, in a good way. It covers NanoGPT, central banking, crypto payments, Monero, Bitcoin Lightning, Nano, THORChain, AI tooling, and a lot of the general "why are we building any of this?" stuff.
If those topics sound interesting, listen to the full episode. The conversation is better than a written recap.
