Do you have any publications to back up your claims about your work? They seem more than a bit grandiose. If you're ideas are as novel and useful as you say then you should publish them.
And I'm sorry, but you're completely wrong about companies recognizing commercial potential. I worked on Alexa for five years, it is a far harder problem than you think. It is nowhere near as simple as "we just weren't looking at the right NN architecture or optimizer!" You're acting like it was a novel idea to think LMs would be extremely useful if the performance was better (in 2017). I'm just trying to tell you that isn't the case.
And I'm sorry, but you're completely wrong about companies recognizing commercial potential. I worked on Alexa for five years, it is a far harder problem than you think. It is nowhere near as simple as "we just weren't looking at the right NN architecture or optimizer!" You're acting like it was a novel idea to think LMs would be extremely useful if the performance was better (in 2017). I'm just trying to tell you that isn't the case.