The 5 hardest open problems in engineering right now:
1. Nuclear fusion
We've been "20 years away" for 70 years. In 2022, NIF finally achieved ignition. The problem now: how do you build a commercial reactor around a reaction that lasts 100 nanoseconds?
2. Desalination at scale
We can remove salt from seawater. We can't do it cheaply enough to matter. Energy cost is the wall. Reverse osmosis still needs a breakthrough material.
3. Carbon capture
Direct air capture works. It costs $300–$1,000 per ton of CO₂. We need it at $50. That's not an engineering problem yet — it's a materials science problem pretending to be one.
4. Brain-computer interfaces
Neuralink can read surface signals. Deep neural communication requires electrodes that don't scar brain tissue. No one has solved biocompatibility.
5. Quantum error correction
Quantum computers exist. Useful ones don't — because qubits fail constantly. Error correction currently requires 1,000 physical qubits per logical qubit.
These are the real frontiers.