Stupid engines make prize winning puzzles >10 moves a joke.
Correcting an old prize winning composition
Yep the original idea is still there. Beautiful puzzle btw, the pawn arrives on h3 just in time.
No underpromotion mate 
knights are too slow
They do - but they also save a lot of cooked problems by verifying their corrected versions. This was always a very risky composition since we humans immediately have the intuition that it might have duals or cooks. There are so many variations that it is really bold to claim that you have seen everything. It's not so much that the engine takes over human territory but more that we are so arrogant to claim to manage the complexities of engine territory. Perhaps we shouldn't attempt to make such problems.
As a composer I have learned to respect the participation of engines and I discovered that they do not takeover the human role of generator of ideas. Good ideas have the magical property of being realizable in a valid setting because good ideas are the result of a valid thought process and engines do not provide that. I am very happy that my version saves all of the creative content of the orginal!
The situation in chess composition is merely a metaphor for the future of humanity in the presence of AI systems. We will increasingly rely on these systems and we must learn to live with that fact. Remember how John Connor was educated into the same reality by his mother and by the terminator. Fighting the machines while simulaneously accepting their aid! No fantasy in those movie stories!
This topic is about the restoration of a brilliant composition from the latter part of the last century which was unfortunately 'cooked' (chess slang for 'found to be incorrect'). Those were the days when chess engines were starting to compete with humans but were not yet widely available for preparation and testing, and probably not strong enough to crack the secrets of long direct mate problems with unnatural diagrams. Today even the free online engines are capable of shooting holes in those from all angles.
Here is the original prize winner with its intended solution:
When you run an engine on this problem it will quickly find a cook - in 8 rather than 10 moves which poses a tough challenge on the repairs. Nevertheless I tried and came up with this "correction":
Is this one correct? I don't know! I have no private engine to do my analysis and depend on what chess.com provides. Which is not that fast or that good for complicated puzzles. If you have a better engine please verify that my solution is correct and that it is the only one!
Has this problem been corrected before? I don't know! I searched the internet and I searched the outternet
but to no avail. My guess is that the "cook" has been found but I doubt a fix was produced. Anyone to contradict me?