Of course trust is still involved. The whitepaper specifically puts "trust" in the context of trusted third parties ("financial institutions serving as trusted third parties to process electronic payments"), and proposes a solution to the double-spend problem when they are not relied upon. It is a very unique solution that could be useful in solving some very specific use cases. There are a wide number of risks that could cause bitcoin to crumble - Core Dev Matt Corralo often says he gives bitcoin a 5% chance of success. The wide-sweeping conclusion is that blockchain is useless, though I would agree in almost all cases, it disregards the successful instances where trust minimization provides substantial benefit.