There will probably be other breakthroughs in AI in the next 10 years but there’s also a chance that we will face roadblocks and find ourselves in a plateau, similar to what’s happened with other technologies in the past. The point at which we we will reach such plateau is almost impossible to tell but my guess is that there will always be a need for developers to perform the least structured tasks. The reason is that AI as we know it today is good at finding patterns in existing data and generating new data as an interpolation of those patterns, hence if a problem has no identifiable pattern or is different enough from the training dataset, it will be hard for an AI to generate a result.