But the font design typically doesn't influence or change the meaning of the text. With emojis it does.
…and that's precisely why I complained that emojis have been standardized as Unicode code points, with their design being left to font designers. You just re-iterated that this is a consequence of using (abusing) the charset, which I had already acknowledged.
Having also inferred "board games" to mean something I'd play on an average game night these days, it did cause me to reflect on making sure I think a bit before jumping to conclusions. The modern world has certainly trained me to think "pfft, call me when this can play a _real_ board game" as my first response, and that's pretty pathetic of me IMHO. The technology in use here makes it a really interesting topic.
MC6809 was actually launched in 1979, like MC68000, about one year after the launch of 8086 in 1978 by Intel.
What Motorola did in 1978 was to publish some articles in the specialized magazines, announcing MC6809 as the future better replacement for their existing MC6800 derivatives. This is the same like Intel describing during last year how great will be their Panther Lake CPU, but Panther Lake has really been launched only a couple of days ago.
Bits per weight, its an average precision across all the weights. When you quantize these models, they don't just used a fixed precision size across all model layers/weights. There's a mix and it varies per quant method. This is why you can get bit precision that arent "real" in a strict computing sense.
e.g. A 4-bit quant can have half the attention and feed forward tensors in Q6, and the rest in Q4. Due to how block-scaling works, those k-quant dtypes (specifically for llama.cpp/gguf) have larger bpw than they suggest in their name. Q4 is around ~4.5 bpw, and Q6 is ~6.5.
My general experience with decompilation has been very negative (rough and not ready for use)
It feels like tool devs target byte editting more than refactoring decompiled code into something readable - you can't move lines of code, can't flip statement checked in if() for early return
Author of this article mentioned "byte euivalence", and while I'd be fine with functional sameness, I imagine provably-reversible refactor steps would be of great help for everyone
Not sure if you're a .NET/C# person, but PDBs are a bit different tho in that they contain full debug information and you can absolutely decompile a .DLL + .PDB combo. Very successfully even in the case of obfuscation.
Fight against obfuscation is different from fighting for readability
I've tried Ghidra, IDA and BinaryNinja, and all of them display code on the level of "C with classes" from early 00s (and declaration of variables at the beginning of function in style of structured programming of the 90s)
I'd be perfectly fine with that output, had there been good way to interactively fix it (refactor without changing behaviour)
No, it's to make it easier to dose different kind of biologically active substances. They can have significantly different "recommended weight to eat of this per day", IUs make that sort-of comparable and easier to remember.
The usability issue with IUs is that people are used to scales measuring weight and containers measuring volume, but an IU is different for each substance.
Another issue is insulin syringes are labeled in "insulin units," which hapless folks reasonably assume can be abbreviated "IU."
If you are measuring out a certain number of IUs, and your calculator or formula hasn't asked you which substance you're working with, you're gonna have a bad time.
Imagine outlawing comic sans because "letters must be serious" or smth
reply