This morning, someone in the group shared a screenshot: their newly trained AI agent has been winning predictions for two days straight. I thought it was just another mystical script, but when I checked the link, it turned out to be plugged into @inference_labs' TruthTensor. Every strategy and trade was logged as a battle report. At that moment, the dev in me suddenly felt: AI shouldn’t just press buttons for you, it should also be audited and ranked. Now, TruthTensor Season 1 has officially started: a 14-day season with leaderboards reset and reopened. The underlying Proof of Inference uses zk proofs to lock in every inference, and 0.5% of tokens are allocated to content and the ecosystem. Kaito Yap’s rankings and tasks are integrated into the same scoring system. From a dev’s perspective, this feels more like a marathon of 'who truly understands verifiable AI,' way more reliable than a one-off airdrop gamble.