anankaie

10 ms·

91.
by anankaie 4y ago
Apple and Microsoft have an in-house ad platform that is provided to, e.g. their app developers to use to monetize free apps, so people can be exposed to their ads there. Microsoft also runs ads on their free offerings, e.g. MSN News. I wou
92.
by anankaie 4y ago
> ultra-slow ‘Enterprise’ ring and a move-fast (think about the speed of macOS changes) ‘consumer’ ring They do: Long-term Servicing Channel releases: https://techcommunity.microsoft.com/t5/windows-it-pro-blog/
93.
by anankaie 4y ago
No. Bring it to court, like literally every other case of law enforcement action.
94.
by anankaie 4y ago
And in practice, this effectively never happens (are there good examples of it happening?). Somehow the revolution keeps getting stuck in the totalitarian phase.
95.
by anankaie 4y ago
One problem is that these techniques - as I have seen them applied - are terrible at analyzing multimodal distributions. They will fail to capture small detractors until they pile up into a feeling of the software being user-hostile, becaus
96.
by anankaie 4y ago
No, the antitrust case turned slightly on appeal (due to Microsoft being able to point at the judge giving interviews about the case on TV /during/ the case), removing the breakup order and enabling them to settle with a Consent D
97.
by anankaie 4y ago
Would you run into class balance issues if you do not have a growing pool of initial videos?
98.
by anankaie 4y ago
It depends on how you set it up. You can use a very non-expressive valuation that assigns a score, but then you need to use reinforcement learning techniques to transform that score to a model update. Alternatively, if your valuation repres
99.
by anankaie 4y ago
The easiest way to do this would be to train a bandit model to be a bidder, using this vector as part of the input features. You would spend some time and money doing randomized exploration (maybe something more sophisticated with active le
100.
by anankaie 4y ago
I suspect that the difference is in knowing versus unknowing use, and the deliberate choice of including learned elements or not on the part of the user. In the case of the ML model, it can reproduce a part of a training example perfectly,
101.
by anankaie 4y ago
If this continues into the long-term without solution, I suspect paid package repos will start making an appearance, whereby only vetted versions of components are available. At least one big company already maintains internal clones of rep
102.
by anankaie 4y ago
And this despite that most of that information is static and primarily text with links to media. There is no reason not to store it as markdown and render the final HTML over the top of it whenever the relevant styling envelope changes. We
103.
by anankaie 4y ago
If a significant (and rising) proportion of the population is starting to distrust you and increasingly anecdotes show that we are certainly not moving toward more acceptance of science in the short term - as a slow, generational effect, ye
104.
by anankaie 4y ago
It did not help that there were strong political partisan effects further corrupting this situation beyond what you have noted. I have to admit, I really wonder what the counterfactual world (where Trump got elected for a second term) would