markhahn

8 ms·

91.
by markhahn 10mo ago
Some more traditional number crunching has long looked at lower- and mixed-precision hardware.
92.
by markhahn 10mo ago
Is this claim based on observed comm patterns in some particular AI architecture?
93.
by markhahn 10mo ago
Really? Fully-connected hardware is in buildable (at scale) which we already know from the HPC world. Fat trees and dragonfly networks are pretty scalable, but a 3d torus is a very good tradeofff, and respects the dimensionality of reality
94.
by markhahn 10mo ago
Text is wrong about CRCs: everyone uses pretty heavy ECC, so it's not just a re-read. This also provides a somewhat graduated measure of the block's actual health, so the housekeeping firmware can decide whether to stop using the
95.
by markhahn 11mo ago
$72/yr is somewhere around 3x purchase price (per year). BB seems to be pretty smart about managing their exposure, infrastructure overhead, etc.
96.
by markhahn 11mo ago
all the major players say "tape". (but that's partly for practical issues like scaling and history)
97.
by markhahn 11mo ago
shouldn't really be a problem, since the capacity required for this use is so low. the real issue here is QLC in which the flash cell's margins are being squeezed enthusiastically...
98.
by markhahn 11mo ago
pointless to do it by the default 512b block size though.
99.
by markhahn 11mo ago
no, the firmware does any maintenance. good firmware should do gradual scrub whenever it's idle. unfortunately, there's no real way to know whether the firmware is good, or doing anything. I wonder if there's some easy way to
100.
by markhahn 11mo ago
I find it implausible that people really have extreme, detailed imagery. Not that they can't do it on demand, if desired. But if every time they imagined something, it instantly appeared with all possible detail - that's just tr
101.
by markhahn 11mo ago
but are those details fabricated on demand? I don't have any trouble following your path of increased detail, but if someone says "imagine an apple", I get a vaguely apple-shaped, generally redish object (I like cosmic crisp)
102.
by markhahn 11mo ago
I don't think vit c has anything to do with daylight/smog. vit d, definitely (and so not rickets).
103.
by markhahn 11mo ago
I think the strongest criticism is just that being short of just about anything would cause significant effects. being short of water, calories, any vitamin, protein, etc.
104.
by markhahn 1y ago
Java, and all living JITs, are hollow parodies of the Self environment.
105.
by markhahn 1y ago
everything is quantum. it's just convenient to deal with the emergent behavior instead (like "chemistry"). yes, photosynthesis is quantum. so is vision, smell, etc. heck, metabolism is quantum! but these are fast and local
106.
by markhahn 1y ago
sure. but there are very tangible bounds on the sorts of physical interactions that can cause any effect. for instance, lots of people love the idea of "brain waves". in general, neurons are event-driven, not given to "wavi
107.
by markhahn 1y ago
Your use of "is" here is problematic (and that's what triggered the other responses). I guess you mean "like". That would be accurate: TMS is disrupting normal brain activity a little like ECT does. It alters cons
108.
by markhahn 1y ago
I think sometimes people forget that when we think of "chemistry", we're talking about an approximation based on behavior that ultimately emerges from quantum effects. It's just that you don't often need to model t
109.
by markhahn 1y ago
No, there's no evidence against Tegmark's math. Finding isolated quantum effects is emphatically unsurprising: after all, everything is quantum. It's just limited in locality, which is basically what Tegmark is talking about
110.
by markhahn 1y ago
that is NOT what the article says. read the article, it's a pretty good survey.
111.
by markhahn 1y ago
Tegmark has used actual, you know, numbers and stuff to show that quantum effects in the brain are pretty implausible. https://pubmed.ncbi.nlm.nih.gov/11088215/
112.
by markhahn 1y ago
The main reason is because it's arbitrary. His "speculation" is litereally: I think quantum is mysterious, and brains are mysterious, so there must be quantum in the brain. That's just silly - even if only because his
113.
by markhahn 1y ago
no, this does not validate quantum mysticism. yes, it's microtubules, but there's no sign of any weirdness here. and Penrose's whole theory is "quantum is weird and I think brains are weird, so brains must be quantum.
114.
by markhahn 1y ago
is he any less competent than, say, Hegseth?
115.
by markhahn 1y ago
it's not. cake flour is simply lower-gluten flour. ther's not a big difference between that and "AP" flour, but a big difference between it and bread or "strong" (high-gluten) flour.
116.
by markhahn 1y ago
Worship of subjectivity is a tell for Mysterianism.
117.
by markhahn 1y ago
whenever I see the word 'reductionist', I wonder why it's being used to disparage. a materialist isn't saying that only material exists: no materialist denies that interesting stuff (behaviors, properties) emerges from m
118.
by markhahn 1y ago
I find that people who complain about "defining" consciousness are, in fact, Mysterians who opposed the very idea of such a definition. All we need to do (to talk about, to study it) is identify it. We need to be using the word t
119.
by markhahn 1y ago
What did you accomplish with it? Another answer talks about saving 40W. Why not? But it's not much in a normal power-cost environment.
120.
by markhahn 1y ago
I think that footnote is close to the heart of it: on a modern OoO superscalar processor, there are hundreds of instructions in-flight. that means a lot of work done to maintain their state and ensure that they "fire" when their

More ›