There are ways you can have an impact, even with a modest contribution. Find a high-quality research lab, like the Monje lab at Stanford, and make a direct charitable contribution. I went through a number of cancer research foundations' 990s and I was dismayed at the very small percentage of their donations that ended up going directly to cancer research. As someone whose father and mother and aunt and several uncles died of cancer, it matters a lot to me that we make progress on fighting cancer!
CAR-T cell therapies in particular seem to be curing cases previously thought incurable. Just this week we saw the Monje lab make forward progress on curing certain pediatric brain cancers!
For TPU pods they use 3D torus topology with multi-terabit cross connects. For GPU, A3 Ultra instances offer "non-blocking 3.2 Tbps per server of GPU-to-GPU traffic over RoCE".
Is that the worst for training? Namely: do superior solutions exist?
I believe the suggested setup was for making a site and images available to the public, for which hiding the origin behind Cloudflare seems a very good reason. Some public IP will need to have ports 443/80 open.
Am I the only one who saw "delve" at the top of the article and immediately thought "ah, an AI generated piece"? Well, that and the over-structured components of the analysis with nearly uniform word count per point and high-complexity but low signal-to-noise vocabulary using phraseology not common to the domain being discussed. (The article doesn't scan as written by an SRE/DBA.)
The propagation differences between 5Ghz and 6Ghz are minimal compared with 2.4Ghz vs 5Ghz. In fact, given that there's a bunch of other protocol & HW improvements, it wouldn't be surprising to see identical 6Ghz and 5Ghz deployment density.
I would guess it to be pretty rare that multiple aircraft would be on approach at once, and if so, I'd imagine one could hold at a distance to allow approaches to be serialized.
I've flown to/from McMurdo on a total of 4 types of planes... C-17, C130 (kiwi AF), L-100 (Safari, contracted by Italian antarctic program) and LC-130 (to/from pole).
Possible corollary: it may be difficult to regularly turn out highly compute-dependent research if you're paying full retail rack rates for your hardware (i.e. using someone else's cloud).
CAR-T cell therapies in particular seem to be curing cases previously thought incurable. Just this week we saw the Monje lab make forward progress on curing certain pediatric brain cancers!
https://www.nature.com/articles/s41586-024-08171-9