Technology is inevitable if a lot of people can independently do it without huge resources, and the result is useful.
- Personal computing - inevitable. Once ICs became cheap, it started happening, with no one effort dominating.
- Moon landing - not inevitable. Huge resource commitment required, and not repeated since.
- Internal combustion engine - inevitable. Once fuels and steel were available, it was possible to contain the explosion of an IC engine, and many people started making them.
- Nuclear weapons - not inevitable. Uranium separation is so hard that somebody had to spend billions to get it to work at all. It wasn't clear that fission could be made to work.
- Radio - inevitable. Once something with gain and something that rectifies were invented, radio was something many people could work upon.
- Steel - interesting case. Steel is thousands of years old, but mass production of steel only dates to 1880. It took considerable metallurgical research to get it right, with about 10,000 tries before the Bessemer converter worked reliably. No one had done that before, and one person did it.
- "AI", via the machine learning route - inevitable. The concepts date from the 1960s, but it took half a century of IC development to make them feasible.
Looking at the issue in this way moves it from rhetoric to reality.
Note that none of the inevitable technologies have a "moat".
- Personal computing - inevitable. Once ICs became cheap, it started happening, with no one effort dominating.
- Moon landing - not inevitable. Huge resource commitment required, and not repeated since.
- Internal combustion engine - inevitable. Once fuels and steel were available, it was possible to contain the explosion of an IC engine, and many people started making them.
- Nuclear weapons - not inevitable. Uranium separation is so hard that somebody had to spend billions to get it to work at all. It wasn't clear that fission could be made to work.
- Radio - inevitable. Once something with gain and something that rectifies were invented, radio was something many people could work upon.
- Steel - interesting case. Steel is thousands of years old, but mass production of steel only dates to 1880. It took considerable metallurgical research to get it right, with about 10,000 tries before the Bessemer converter worked reliably. No one had done that before, and one person did it.
- "AI", via the machine learning route - inevitable. The concepts date from the 1960s, but it took half a century of IC development to make them feasible.
Looking at the issue in this way moves it from rhetoric to reality.
Note that none of the inevitable technologies have a "moat".