The optimism trap

Early success made the future look close.

In a 1957 speech, Herbert Simon predicted that within ten years a computer would become the world chess champion, discover and prove an important mathematical theorem, and compose music valued for its beauty. The speech was published the following year.

In 1967, Marvin Minsky wrote that within a generation the problem of creating artificial intelligence would be substantially solved.

Both forecasts grew from real progress. The Logic Theorist had proved theorems, and search could solve carefully described problems. The mistake was assuming that the remaining problems would yield at the same pace.

The chess prediction did eventually come true, but nearly four decades later. Its route there shows both the power and the limits of search.