Do you trust AI? You probably shouldn’t. Let’s check it!

What annoys me most about modern AI is this “black box magic.” You send a prompt, get a result, but you have no idea whether the model actually computed everything correctly or just “made it Do you trust AI? You probably shouldn’t. Let’s check it!

What annoys me most about modern AI is this “black box magic.” You send a prompt, get a result, but you have no idea whether the model actually computed everything correctly or just “made it up” because it was more convenient.

Yesterday I dug into the architecture of @OpenGradient . The team is tackling a fundamental problem: how to make AI computations transparent. With their HACA technology, AI model outputs can be cryptographically verified directly on the blockchain. This is no longer just “believe it or not” — it’s mathematical proof that the model is operating without manipulation.

I’m wondering whether this could become a standard for the entire DeFi sector. Because trusting algorithms that manage your money without proof of code execution is basically suicide

#OPG $OPG

What do you think is currently holding back the development of AI in crypto the most?