Also the funny part is that there are many nice ML algos that do not need harsh optimisations or even do not need optimisation at all (see random ferns for inst

Also the funny part is that there are many nice ML algos that do not need harsh optimisations or even do not need optimisation at all (see random ferns for instance). Thus it might be hype for hype.