Tech: deploy as a VM. For starters, allow the client to configure the VM network interface before downloading it. (On the fly, mount the VM with qemu nbd or whatever and write to the image, unmount qemu-nbd and proceed with the download request)
Expose everything as a RESTful API in the VM. NOTHING ELSE. Don't spend too much time hardening your VM security wise. Have a strict release schedule. Listen to the customer on what features they might want. But have your release schedule. Don't blindly add in everything this client wants.
If client is insistent on source code, throw an escrow clause into your contract. Dont had over everything yet.
Sales: draw up a contract. charge according to the sales figures of the client (yearly or one time is left to you). I cannot stress this enough.
Clinically, I can state that I am a live and let live person. Am not quoting that as a boon or bane. I suspect most people are. And this lot, while they can dig themselves out of poverty, cannot become rich. The reason is there is the small group that does not have the live and let live wired it. That that trait is good or bad is up for debate. It is this small group that ends up being rich given enough catalysts. They think only about themselves.
I can tell you right away that it would be of huge help when reading a new language especially in a new script. And who are biggest set of users that might benefit from this? (clue: they haven't said their first words yet).
Which one? Sorting out names, verbs, adjectives etc. Or finding out what "role" each word is playing (sorry, English is not my language so I do not know the proper technical terms for this... In my culture the first is called "Grammatical Analysis" while the latter is called "Logical Analysis")
Matrix Reloaded was 13 years ago. That's a bit unfair. You can still tell, I think, especially if you know where to look, but I don't think most people would notice, in Fast and the Furious 7, for example.
Terminator 2 used far, far more practical effects than one is likely to assume these days. The iconic shot of the t-1000 blown in half and sewing it self back up? Practical effect. Really amazing stuff over all. https://www.youtube.com/watch?v=EYQMfT6nsQs
Content sources typically charge per region. The more the subscribers per region, the better is Netflix chances of broadening the catalogue in that region. It is a tough problem that has several facets. Prime move, capturing a subscription base, marketing, campaigning, trials, licenses, sourcing, regional law compliance (nudity/violence etc.), infrastructure, etc.
Am not saying as a consumer you are wrong. But a perspective from the other side at least partially covers the why.
Statement from a Slashdot post about the AlphaGo victory: "We know now that we don't need any big new breakthroughs to get to true AI"
That is completely, utterly, ridiculously wrong.
As I've said in previous statements: most of human and animal learning is unsupervised learning. If intelligence was a cake, unsupervised learning would be the cake, supervised learning would be the icing on the cake, and reinforcement learning would be the cherry on the cake. We know how to make the icing and the cherry, but we don't know how to make the cake.
We need to solve the unsupervised learning problem before we can even think of getting to true AI. And that's just an obstacle we know about. What about all the ones we don't know about?
Seriously. I recall proving that there is one and only one circle that can be drawn through 3 distinct points on a Euclidean plane. This "theorem" can be thought of as a corollary of that.
