I have a couple of references but I feel like other people have a collated list of things they might refer to that I would find helpful. Yes I have done a google search however many of the resources are all floating point centric which I am less interested in. Linear algebra for SIMD with a focus on Fixed point is my primary concern.

Topics/Algorithms of interest: Ichol Matrix inversion Household vectors rotation matrix generation General filtering Quanization design Fixed Point stability of matrix operations FFT/DFT (fftw is good but other soruces welcome) Complex numbers for radio algorithms (Qasim Chaudhari has some good wireless topics books)

HW analysis side: Comparsion of Memory version Register based ISA. I think RVV prompts doing byte filters in memory rather than via perms in register space. Their are lot of opinions and I honestly am looking for arguments for architecture trade-offs the comparsion of different approaches is hard to find

Comments URL: https://news.ycombinator.com/item?id=48096162

Points: 2

# Comments: 0

The surveillance of the future will be powered by the things we produce today. If the accepted algorithms leave cookies those cookies will be used tracked and monitized. The bad argument is the forced verification to do things on the internet. Making that start at the hardware is a lock in thats not okay. Business will always own the services and making standards that trade our practical liberty for the sake of security is a very compromised position in my opinion.

And it does start with the age verification, followed by id checks, etc. Its compromising precisely because no lines are drawn and no rights to privacy are codified in law. Without guiderails the worse path will likely be taken for maximum profit

If you just look 10 years ahead you'd probably conclude solar, wind and maybe hydro is enough because short term thinking will always undersell the climate risk in my opinion. My justification for this thought is look at climate deniers arguments its always about magnitude and speed now because its the last effective argument.

Nuclear cost recovery and profit function for proven GEN4 is also on the 20-30 year timescale (depending on how much cost overruns they've faced it could be 50 years for bad cases) not the 5-10 year timescale. Making them unattractive financially speaking. Despite the fact that after that time which most US reactors are they are extremely profitable for the operator because the fuel -> power out is incrediblely in their favor. Ultimately, it takes longer term risk evaluation to show their benefits but they are undeniable and will be involved in solving the climate crisis.

Ticketing gets weird because it varries by state I cannot speak to Califorina because I don't live there but in the northeast many states have points and if you get 10 points you lose your license. Not every ticket gives points i.e parking tickets. But speeding X miles over is 4 points. Drunk driving is 10 points generally speaking. Their are others like drag racing on public roads etc. But Those systems would be very harsh on a self-driving vechicals which all have the same "driver". You might need to be more permissive on the number of points because the "driver" is clocking 1000x-10000x what an ordinary driver does but tickets solves all the practical problems your bring up. The only grounds to debate feel like the number of points or threshold which besides stating I don't feel like you stake a position on where the threshold is (doing sound seems really hard without data on how often violations can be). Intentionality is also ignored because it doesn't matter in driving since you could always say you didn't know X law i.e unintentional but its not an excuse because you are opt in to using public roads you have to know the laws.

I am not trying to be dismissive but I am really unsure what other system and thresholds you are describing while you are being compared to an extremely concrete and effective systems (effective at punishing drivers :) from actually dangerous behavior). Usually the only hard part is observing offense but I think that makes ticketing acceptable because generally speaking the penalty for getting a ticket is very harsh on low income earners (increases compliance)

OSs are few in number and are special pieces of software supported by extreme network effects and locking in effects. Its virtually impossible to abandon mainstream OSs even in the desktop world which is much more permissive their are basically 2 OSs. You can count Linux as a third but its not a serious market share competitor yet! EU and France might change that.

But essentially total OS control allows a type of anti competitive behavior that is unavoidable. You either need 10s of options which are different in the services which isn't viable because modern OSs are some of the most complex pieces of software in existence. Your only other option is to regulate them like the power company because its not practical to build your own grid to have fair pricing and access

Ironically I was watching Nuremberg last not and is is schocking how close some of the leaders of this country are to characters like Hermann Göring, or Hitler himself in talking points. They are certainly populists but the language they used is MGGA (make german great again) so to speak. And factually that were not particular that good at it either most of Germans recovery is really due to the liberal government that pass laws that built the Autobahn were laws not by the Nazi party. They certain jumped on them and accelerate them but effective governence is not really for the populist

I think the major struggle with basic research is there is no way to conduct it in which results are guarenteed. If you could do that you wouldn't need basic research. But there are a ton of questions whose outcomes are not really valuable at all but you simply don't know. On net science dispite those many useless questions answered still is extremely net posititve because some of those apparently meaningless questions ended up being the right question to drive research to useful good answers.

How else could states that deny those data centers if they cannot pass legislation to prevent them or require XYZ parameters before they are allowed to be built? Your argument is nonsensical in my opinion especially in context. I get that if you do a string compare they are different sentences but the semantic effects of the two statements are equavalanet in the framing that comapnies are actively trying to permit and build them.

The managers and everyone are so excited by the fact the person did it with AI but I just get really confused because it seems like they just made some worse that has less value because it cannot actually correctly simulate the thing we want to test. Maybe i am being petty and salty but I think the that this is time wasted by any measure. And net-negative value but the team wants to emphasize we are using AI. There have been some productive uses but the productivity trap-doors are about the same as with normal development just people seem more willing to take the trap door ideas now.