adrian_b

33 ms·

541.
by adrian_b 4mo ago
Many cost relationships from TFA have already been more or less true for the 32-bit CPUs launched after 1990 and they all became true for the 32-bit high-end CPUs launched after 2000 (like Intel Pentium 4 and AMD Athlon XP), when the differ
542.
by adrian_b 4mo ago
I think that rules like this, for generating systematically a set of related identifiers, are fine, unlike the Hungarian notation, where there is no connection between identifiers that have a common prefix, except for sharing a data type.
543.
by adrian_b 4mo ago
The problem is that your explanation confuses phonemes with letters. A spoken language is described by decomposing the spoken words into phonemes, where phonemes are sounds that distinguish words, in the sense that replacing one phoneme in
544.
by adrian_b 4mo ago
The first FORTRAN versions encoded the type of a variable or function in the initial letter of its name: I, J, K, L M and N for integer, other letters for floating-point numbers. The first BASIC versions encoded the type of a variable or fu
545.
by adrian_b 4mo ago
Using the term "torsor" for that mathematical concept has been a very bad choice, both because the concept does not have any obvious relationship with the meaning of the word and because the word "torsor" had already bee
546.
by adrian_b 4mo ago
The term "baseless logarithm" is really nonsensical and using it would be a great mistake. Nonetheless, where the author of TFA is correct is that logarithms are a single physical quantity, like length, area or volume, and that ch
547.
by adrian_b 4mo ago
The high costs are necessary for high speed. When a low speed of the order of one token per second is accepted, any open weights LLM can be run on an ordinary PC (with the weights read from SSDs) and the cost becomes negligible. Such a low
548.
by adrian_b 4mo ago
While OpenAI and Anthropic have not provided any useful information for a long time, there still are some research publications from a few US companies, e.g. NVIDIA about its Nemotron models, or Google and IBM about their small LLMs.
549.
by adrian_b 4mo ago
Except that it is not a laser but a high power radio transmitter made with a vacuum tube (gyrotron). For generating the highest possible power of radio waves, vacuum tubes remain the only solution. This drilling method resembles more a micr
550.
by adrian_b 4mo ago
In my opinion, the article is very mistitled, because it does not contain even a single valid criticism against the geometric algebra theory, despite containing some perfectly valid criticism against some mistakes frequently made by geometr
551.
by adrian_b 4mo ago
Biological evolution has indeed lead much more frequently to simplified, streamlined structures good for a single purpose, which had evolved from more complex and more versatile structures, than to more complex structures that had evolved f
552.
by adrian_b 4mo ago
As demonstrated in an older article cited in the parent article, at the same pace of 6 breaths/min it matters a lot which is the ratio between the exhalation time and the inhalation time. In the experiments, slow inhalation with fast e
553.
by adrian_b 4mo ago
The title is inaccurate. The results are not about any kind of deep breathing or slow breathing. The results are specifically about a breathing that is slower due to prolonged exhalation. This kind of breathing is one of the many kinds of b
554.
by adrian_b 4mo ago
Society of Motion Picture and Television Engineers. The SMPTE standards have been very important for cinematography and television, especially for professional applications. Their importance has decreased since the transition to digital vid
555.
by adrian_b 4mo ago
I doubt that it has ever been possible to obtain enough output tokens from OpenAI or Anthropic to be useful for training other LLMs. In any case, had that been possible in the beginning, it stopped being possible long ago, because any suspi
556.
by adrian_b 4mo ago
In a CPU, there are 2 kinds of branch predictors, predictors for conditional branches, which must predict only whether the branch is taken or not taken, and predictors for indirect branch instructions, which must predict the address of the
557.
by adrian_b 4mo ago
The same is true for a "switch" statement. If the "case" values used in a "switch" are sparse, the "switch" will not be compiled into an indexed jump instruction, but into a sequence of conditional ju
558.
by adrian_b 4mo ago
There are low-drift-rate inertial solutions, but they have high cost and big size (i.e. laser gyroscopes and accelerometers, with atomic clocks). So I assume that the research effort is directed towards reducing the cost and size.
559.
by adrian_b 4mo ago
> According to the article you get purified greens from transmittance through foliage, ie backlight in eg a maple forest. This sounds plausible. I think that in general for content that you record yourself, where you would record whateve
560.
by adrian_b 4mo ago
You are right that I have not mentioned this, but indeed wide color gamut monitors should also support 10 bit or better resolution per color component. Software that does color processing should convert all input pixel formats into BT.2020
561.
by adrian_b 4mo ago
Indeed, "Figure 3" from that article should be a realistic depiction of the differences between sRGB, Display P3 and BT.2020. It is true that both the red and green primary colors of sRGB are bad (because they correspond with obso
562.
by adrian_b 4mo ago
In the way that you say, you can do lossless data compression, but then the LLM is used in a very distinct way than it is used in applications like chat or coding assistance. In the latter applications, you do queries which aim to extract i
563.
by adrian_b 4mo ago
Which is why in this book the title of a paragraph is "Compression is an artificial intelligence problem". However, I believe that is less useful to think that AI "finds an universal compression", than to think that the
564.
by adrian_b 4mo ago
While it is true that some saturated blue-green colors will never be reproducible with only 3 primary colors, the CIE 1931 chromaticity diagram used in TFA overemphasizes their importance, because human vision cannot distinguish many colors
565.
by adrian_b 4mo ago
The comparison results seem very plausible. From the conclusion, I agree with: > I wouldn't make either one the top-level coordinator by default. But I do not agree with the follow-up sentence: > The best shape is still a frontie
566.
by adrian_b 4mo ago
The human arms and hands are very versatile, and imitating them is a good choice for a universal robot, though 3 or 4 arms are definitely better than 2, and the hardest to imitate are the sensors, not the actuators. But the rest of the huma
567.
by adrian_b 4mo ago
It should be good enough for interactive use, but not for watching movies. In TFA it is said that for these new faster panels the transition time of a pixel is around 50 ms. This is comparable with some old LCDs.
568.
by adrian_b 4mo ago
When an invalid instruction exception occurs, it is very easy to identify whether the invalid instruction was an AVX-512 instruction, because those start with a prefix different from any other instructions. After that happens, the thread ca
569.
by adrian_b 4mo ago
I also see that error code, from Europe. Perhaps it blocks any non-USA connection.
570.
by adrian_b 4mo ago
It would be possible, but Microsoft has already accumulated many decades of being extremely reluctant to make even the slightest changes in Windows in order to accommodate new CPU features. Since 1990 and until now, the Intel-AMD ISA and th

More ›