Sure, smarter algorithms that are able to work with tiny datasets are likely possible. But when will they be available? We just don't know. We have already tried for a long time and didn't find anything that worked well, so it is clearly not a trivial problem. Your "let the robot play with the cat proposal" would likely end up with a robot that is able to recognize that cat extremely well, but fails to recognize that a cat with a different fur color is also a cat. It is also just a different way of generating more data, while humans are actually able to learn from very few examples, even without any way of generating more examples.
So there are many problems that need to be solved first and likely many other roadblocks that we don't know about yet. You can't make a prediction based on that. It is not impossible, that we will have superhuman general artificial intelligence in a few years, but it is also very possible, that in a 10-20 years we will know that our current tools are useful, but limited to a specific set of applications and general intelligence and other application areas are still searching for algorithms that work.
Yes, AI made impressive progress in the last few years, but there is not reason to believe all hard problems are solved now and that would be just a matter of year years until we get to superhuman general intelligence. At the moment we are nowhere near that.