My Seattle Mazda was hit by this bug. The backup camera still functions for the most part, although it does flicker and occasionally blank for a second or two (presumably while the system reboots). Annoying, but not horrible.
X-ray lenses don't really exist, although you can produce focused beams (using either reflection or simply collimation). So, these methods typically involve scanning a sample and measuring some fluorescence line (which gets emitted in roughly all directions).
That site doesn't mention the spot size, but its likely much larger than the page thickness, so the iron signal from a tilted page probably wouldn't differ in a measurable way from having the page perpendicular to the beam. Keep in mind that they require ~24 hours to get even the images shown. Instead it sounds like they rely on compositional differences in the ink from the two sides and look at fluorescence from an element that is only present on one side.
It might be possible to get the other side by taking a difference between the Fe and Ca signals (after scaling appropriately).
In the nexus presentation, they mentioned that a few symbol keys were moved up to soft keys to fit the keyboard while keeping a laptop-like distance between keys. So, it might not be super convenient, but you can likely still type your passwords.
If you treat the indices as between elements and include all elements that are between the start and end elements, there is only one reasonable choice. It would be more "special" to include elements that are outside of the specified range (i.e., the element to the right of the final index).
This is especially true if you have multimodal distributions (such as the income datasets in the article). Although its true that there are simple algebraic properties that allow you to calculate the mean and variance of the total population given the same for sub-populations, it often isn't the case that this will be a good fit to the data. That being said, this is a useful property for parallelizing gaussian fits.
I don't think that will actually change anything. Two positively slowing lines (of any slope) have, by definition correlation = 1.0. The "random data" he used is just a small amount of noise on top of the line(s).
Here, "good" is on the left and "bad" is on the right. The color is orthogonal (it gives the number of operations with latencies in a given bucket). For example, a red square on the right side of the output would have definitely been "bad".
For apps that have been popular for a long time, I wonder if there is a negative bias in reviews (beyond that already present in all online reviews). A customer who gets an update that breaks an app they use every day is very likely to submit a negative review (even if they have reviewed positively in the past). However, a customer that gets and update that make incremental changes that largely don't affect them has no incentive to go post a positive review of the new version. So, the positive reviews come only from new customers, which are relatively small for apps that have reached saturation.
I agree with you on the need for good dedicated test engineers, but even an apparently low crash rate can be damaging at scale. If you have 10,000,000 customers, a crash rate of 0.17% (over some unspecified time period) is 17,000 crashes (in the same time period). On the other hand, if the crash is deterministic, but caused by some user-specific state that affects 0.17% of your customers, that's 17,000 unhappy customers who are drastically more likely to post reviews than the potentially millions of happy customers not experiencing crashes (and perhaps reviewed the app positively years ago).
Note: I don't use Facebook's app often, and have no context for what the actual issues are.
I agree with your first two paragraphs. But, be careful about the last. The pilot wave in Bohmian mechanics is not a wave in space time, but rather a separate physical entity. Basically the dynamics of both the particle and its pilot wave are described by different portions of the wavefunction. (The wavefunction is complex, and so has enough variables to describe two physical quantities).
That was a fun one. I seem to recall it taking a day or so to generate a suitably colliding prefix pair using publicly available code. After that the rest was pretty easy.