Ask HN: Could a form of modern-day AI been made to work back in the late 1980s?
Before you answer with a loud "No," this isn't a question that asks for facts. Of course, it didn't. But also, we also know that the underlying theories behind LLMs and other artificial intelligence techniques are not new; some date back to much earlier. Plus, there were efforts on both sides of the Cold War sectrum to build a supercomputer "to rule them all."
Thus, if we imagine that the scope of my question is in the context of a parallel reality hard-sci-fi story, how plausible would it have been for any side to advance over the other with any form of artifical intelligence (as rudimentary as it may have been)? NSA hold basic patents in the field from the 70s on. They've also been hot for floating point vector machines in a big way since then.