imo testnet hype doesn't always translate to mainnet success but paragraph's approach looks fresh, might be worth a small bag