If you want to narrow down intelligence to only truly novel ideas and not incremental improvements on existing techniques nor applying knowledge others have found out for you, then you're backing yourself so far into a corner I'd say most people don't need to show any intelligence to survive. Whether it's your dad teaching you fishing or the school teaching you algebra or yourself reading a textbook on it, what most people mean by learning is absorbing skills others already have. It's maybe easier to see if you go back to before there was a general education, if you were the smith's son you'd be a smith. He'd teach you all he knew about it and that's the prime skill you'd swap for food and drink and shelter. If you could refine that skill, even better but grand creativity isn't really required to make a living. Of course you might argue that we've built skill into machines since the spinning jennys of the 18th century, but there's degrees.
I don't care if the solver is "dumb", if I ask an AI to nail up some boards I don't care if it needs to go through every tool in the toolbox before it decides that yes, a hammer is the best tool the same way a chess engine goes through every combination. Doesn't really matter as long as it gets the job done. Particularly not if it could break down a multi-step job like building a fence to digging holes for poles with a shovel, using a sledgehammer to drive them down, a saw to cut boards, a hammer to nail them up and a brush to paint them, kind of like a chess engine using many different pieces to force a mate. We're just not good enough at describing the board, the pieces, the moves or the goal just yet. Maybe it's not strong intelligence, but we humans often do stupid and sub-optimal choices simply because we have a limited span of time, skill, reaction, memory, meticulousness and so on. Saying it's beating us, but not on intelligence sounds a bit like sour grapes.
P.S. No, unrestricted chess engines don't lose to grandmasters anymore and haven't done so since the 1990s.