Any value, economically, that your "useful" work provides must be subtracted from the network's overall security. i.e. if you devise work so useful that people

Any value, economically, that your "useful" work provides must be subtracted from the network's overall security. i.e. if you devise work so useful that people will pay for its computation, than any attacker can use that value to subsidize their attack on the network. Currently if you were to buy enough hardware to attack bitcoin, for example, after the attack your hardware (which would have to be ASICs) are useless garbage, as the network would be devalued after attacking it and the hardware serves no other purpose. If I can then pivot to computing something of value, this economic hurdle isn't as expensive.