markhahn

16 ms·

211.
by markhahn 2y ago
is it wifi, bt, zigbee? surely matter-compatible, right?
212.
by markhahn 2y ago
I usually prefer the 'reset' option (such as in IPMI). After all, this is the as-designed way to politely ask all devices to re-initialize. Yes, power-cycling is more unambiguous, but afaikt, the example here is purely that power
213.
by markhahn 2y ago
odd to publish such an article with so many undefended opinions. for instance, that low-boilerplate is somehow worse than high-boilerplate languages because the boilerplate somehow builds character ("teaches important lessons ... from
214.
by markhahn 2y ago
Intel's board certainly has a lot to do with things like acquisitions, whether the company goes into certain fields. Foundary is a good example: either it was a good idea (and Pat definitely got it going with EUV), or it's a bad
215.
by markhahn 2y ago
nvidia's success has nothing to do with bean counters - if nvidia even has them. bean counters are for industries like beans - where the market is stable and the only way to liberate shareholder value is to catch the stray beans. they
216.
by markhahn 2y ago
not sure why people give so much credit to Cuda, at least now. AI hasn't cared about Cuda for years - sure, Torch on NV will use it, but AI is all about Torch/TF, not details below that.
217.
by markhahn 2y ago
I'm curious why you think there was not effort behind ia64. I still think ia64 was sunk by the let-the-compiler-do-it tarpit. There were ambivalent aspects such as, at the time, questions of power or chip area. But things like memory
218.
by markhahn 2y ago
Intel did a good job on iGPUs, which satisfy most customers (even back in the i915 days). But in retrospect, of course it was a missed market. It's hard to disentagle that choice from others in the same era (including the refusal to j
219.
by markhahn 2y ago
No offense, but what you say sounds like an argument for the tech approach. Intel didn't lose the cloud, and what you describe as power efficiency was the fab/euv problem (that Pat fixed). AI: okay, but are you sure that Intel sh
220.
by markhahn 2y ago
that's pretty funny! maybe a political-science researcher, not from Germany would give a less tone-based definition.
221.
by markhahn 2y ago
only Brits? the set of people interested in AI seems to be quite specific: techno-optimists, fad-seeking "entrepreneurs", people who can get with low-quality outputs.
222.
by markhahn 2y ago
how about this: you can only abandon hardware if you enable open firmware on it.
223.
by markhahn 2y ago
So many words, so little detail eh? It's hard to tell whether their thing is yet another take on the sea-of-cores model, or something more interesting. They talk in terms of risc-v, so maybe it's a lot of cores, with something ne
224.
by markhahn 2y ago
afaik, it really depends on the spatial structure of the audio field. think nyquist sampling rates, applied to space, and you can't apply a low-pass filter just because you don't care about higher-order signals. that means that f
225.
by markhahn 2y ago
understood: big facilities get shared; sharing requires arbitration and queueing. an interesting angle on 50 threads and 256G: your data is probably pretty cool (cache-friendly). if your threads are merely HT, that's only 25 real core
226.
by markhahn 2y ago
the question is really: does it have anything vendor-specific, interop-breakers? FC seems to work nicely in a single-vendor stack, or at least among specific sets of big-name vendors. that's OK for the "enterprise" market, w
227.
by markhahn 2y ago
probably easy to get ethics approval though ;)
228.
by markhahn 2y ago
eBPF I think the question becomes: how do you all creation of safe kernel code when that code becomes more functional. historically, BPF was just performing some filtering (near reception -> efficient). eBPF originally added more metri
229.
by markhahn 2y ago
Jensen embraced AI as a way to recover TAM after ASICs took over crypto mining. You can see that between-period in NVidia revenue and profit graphs. By that time, GP-GPU had been around for a long, long time. CUDA still doesn't have
230.
by markhahn 2y ago
"not perfect so it's worthless!"
231.
by markhahn 2y ago
weird - everything claimed as wrong or bad is, IMO, utterly sensible and robust in the real, mean old world. by the author's logic, TCP's entire logic of congestion control is terrible because it's not eg, a deterministic cr
232.
by markhahn 2y ago
Imagine a rice cooker (perhaps with pressure) that gives you data and control. Modulate the applied power (pwm), read a few thermocouples, pressure sensor, release solenoid. Make it all programmable, let people program it. I speculate that
233.
by markhahn 2y ago
No, I miss monochrome displays too. Yes, sure, color can provide some benefit in syntax highlighting, but that's a pretty weak argument. I wonder if there's an interesting reseach opportunity: how complex does a syntax need to be
234.
by markhahn 2y ago
do some people find it pleasant to pan their head? maybe my eyeballs don't rotate far enough?
235.
by markhahn 2y ago
"Chilling effect" implies that we should want ARM to succeed. IMO we need to question the premises of the current IP ecosystem. Obviously, the principles of open-source are quite the opposite to how ARM licenses IP. (Afaik, ARM
236.
by markhahn 2y ago
yes, the AI bubble is at least deflating. that's what's so corrosive about media coverage and stock-market speculation: it corrupts any actual price signals in the market. there is not an infinite demand for AI training, let alone
237.
by markhahn 2y ago
I think your question mainly demonstrates how much trouble most people have about reasoning about exponentials. Not intending any personal insult, but unless you do it regularly (ie, probably are a physicist), you will use a term like &quo
238.
by markhahn 2y ago
Now imagine that this was in gradeschool curricula! I really think education is mostly about providing higher-level intuitions - making correct thought habitual and thus easy. Part of what's so attractive about this particular article
239.
by markhahn 2y ago
the stock market is just a casino, and this article is just an attempt to pump NVDA. especially if you line up availability dates, AMD is competitive. not to mention price! if there's any news here, it's that the recent announcem
240.
by markhahn 2y ago
there are a thousand excellent CUDA programming courses/sites. none of what was mentioned in this blog post is news if you've ever had more than 2 hours of a CUDA course...

More ›