#opg $OPG At the TGE, I sold half my position right at the open. My plan was to hold the rest and see how things developed, but after the price ripped higher, two heavy bearish candles hit hard and I panicked out of the remaining half too. In the end, I still walked away with about 80U profit after costs, so I cannot complain too much.
When I first opened Twin.fun, I assumed it was just another AI agent marketplace and almost swiped past it. But after checking the pricing model more carefully, I stopped and looked again. Twin.fun is actually a digital twin market built on @OpenGradient It lets people create AI personas based on real individuals or identity-like profiles. Each twin is tied on-chain to a 16-byte ID.
The interesting part is the trading design. It uses a quadratic bonding curve, meaning the price of keys rises as supply increases. Buyers are not purchasing a normal subscription; they are buying keys that function more like tradable shares. On top of that, every trade includes two fees: one goes to the protocol treasury, and the other goes to the twin owner.
If you hold at least one key, you can unlock chat access to that persona and use its tools and features.
What stood out to me is how closely this resembles a memecoin-style market. Early buyers enter cheaper, later buyers pay more, and the value moves with attention, hype, and public sentiment. In effect, people are not just buying access to an AI model; they are speculating on the popularity of a digital persona.
From an OPG perspective, the protocol fee could support treasury growth and value capture. But the same mechanism also makes revenue highly dependent on attention cycles, controversy, and volatility. That makes it very different from revenue driven by stable compute usage.
For now, it is an interesting model. But the real test will be whether this kind of product can stay valuable once the hype cools down.$CAP