The title of Tencent's paper [0] as well as their homepage for the model [1] each use the term "Open-Source" in the title, so I think they are making the claim.
One thought is that it's always a whole number of bits (3) to bit-address within a byte. It's 3.5 bits to bit address a 10 bit byte. Sorta just works out nicer in general to have powers of 2 when working on base 2.
Another part of it is the fact that it's a lot easier to represent stuff with hex if the bytes line up.
I can represent "255" with "0xFF" which fits nice and neat in 1 byte. However, now if a byte is 10bits that hex no longer really works. You have 1024 values to represent. The max value would be 0x3FF which just looks funky.
Coming up with an alphanumeric system to represent 2^10 cleanly just ends up weird and unintuitive.
We probably wouldn't have chosen hex in a theoretical world where bytes were 10 bits, right? It would probably be two groups of 5 like 02:21 == 85 (like an ip address) or five groups of two 0x01111 == 85. It just has to be one of its divisors.
Because modern computing has settled on the Boolean (binary) logic (0/1 or true/false) in the chip design, which has given us 8 bit bytes (a power of two). It is the easiest and most reliable to design and implement in the hardware.
On the other hand, if computing settled on a three-valued logic (e.g. 0/1/«something» where «something» has been proposed as -1, «undefined»/«unknown»/«undecided» or a «shade of grey»), we would have had 9 bit bytes (a power of three).
10 was tried numerous times at the dawn of computing and… it was found too unwieldy in the circuit design.
> On the other hand, if computing settled on a three-valued logic (e.g. 0/1/«something» where «something» has been proposed as -1, «undefined»/«unknown/undecided» or a «shade of grey»), we would have had 9 bit bytes (a power of three).
Is this true? 4 ternary bits give you really convenient base 12 which has a lot of desirable properties for things like multiplication and fixed point. Though I have no idea what ternary building blocks would look like so it’s hard to visualize potential hardware.
It is hard to say whether it would have been 9 or 12, now that people have stopped experimenting with alternative hardware designs. 9-bit byte designs certainly did exist (and maybe even the 12-bit designs), too, although they were still based on the Boolean logic.
I have certainly heard an argument that ternary logic would have been a better choice, if it won over, but it is history now, and we are left with the vestiges of the ternary logic in SQL (NULL values which are semantically «no value» / «undefined» values).
Many circuits have ceil(log_2(N_bits)) scaling wrt to propagation delay/other dimensions so you’re just leaving efficiency on the table if you aren’t using a power of 2 for your bit size.
You can also construct multiple hypothesis trackers from multiple Kalman Filters, but there is a little more machinery. For example, Interacting Multiple Models (IMM) trackers may use Kalman Filters or Particle Filters, and a lot of the foundational work by Bar-Shalom and others focuses on Kalman Filters.
The Kalman filter has a family of generalizations in the Extended Kalman Filter (EKF) and Unscented Kalman Filter (UKF.)
Also common in robotics applications is the Particle Filter, which uses a Monte Carlo approximation of the uncertainty in the state, rather than enforcing a (Gaussian) distribution, as in the traditional Kalman filter. This can be useful when the mechanics are highly nonlinear and/or your measurement uncertainties are, well, very non-Gaussian. Sebastian Thrun (a CMU robotics professor in the DARPA "Grand Challenge" days of self-driving cars) made an early Udacity course on Particle Filters.
I was encountering the same problem on my Intel MBP, and per another one of the comments here, find that switching from Chrome to Safari to view the page allows me to view the whole page, view it smoothly, and without my CPU utilization spiking or my fans spinning up.
I don't think anyone in this thread knows what happened, but since we're in a thread speculating why the CEO of the leading AI company was suddenly sacked, the possibility of an unacceptable interpersonal scandal isn't any more outlandish than others' suggestions of fraud, legal trouble for OpenAI, or foundering financials. The suggestion here is simply that Altman having done something "big and dangerous" is not a foregone conclusion.
In the words of Brandt, "well, Dude, we just don't know."
Kim Stanly Robinson also hits on this in Red Moon. Someone is emotionally invested in having a particular plot of moon where the earth peeks out through a gap in a crater wall during the wobble.
From the captioned art in the article: "Siege, from the Peterborough Psalter, early 14th century, via the KBR Museum, Belgium.
Yes, those defenders are all women."
Unlike the OP article, it does make use of the math formalism for Kalman filters, but it is a relatively gentle introduction that does a very good job visualizing and explaining the intuition of each term. I have gotten positive feedback (no pun intended!) from interns or junior hires using this resource to familiarize themselves with the topic.
If you are making a deeper study and are ready to dive into a textbook that more thoroughly explores theory and application, there is a book by Gibbs[1] that I have used in the past and is well-regarded in some segments of industry that rely on these techniques for state estimation and GNC.
[0] https://arxiv.org/pdf/2411.02265 [1] https://llm.hunyuan.tencent.com/