That's simple - to make quick rotation, you must apply force (torque) and beat inertia - from mathematics could know, the farther from center of mass the more energy will spend to rotate at same speed.
So, to rotate faster, you need larger control surfaces.
From other side, traditionally, self-stabilize spent at least 1% of energy (on small planes normal up to 10%). What all this mean - with 10 000kg of total weight, your control surface will constantly make 1000kg of force to just fly, but when need to turn, will need significantly more than 1000kg, that's all.
Old planes need self-stabilize, because constant corrections was very time consuming, but modern have powerful computers and could provide artificial stabilization - current 1kg computer could provide same stabilization as control surface constantly making 1000kg of force.
Video have some inaccuracies. As example, said "to detect heartbeat from 100m, 10-20 T", but then talking about "from 100km".
So, if carefully looking, SQUID could detect with 10-15T, so need just additional 5 magnitudes. And as I know from practice amateur astronomy, with regular signal, 5 additional magnitudes could be achieved with numerical methods.
So we have two questions remain - how CIA managed to have SQUID small enough to fit on something like helicopter (or weapons bay on fighter), and how they managed to fly nearer than 100m to pilot, each could be answered - "possible, if have highly motivated people, with big money".
> aren't server CPUs only like 25% of the total market?
Yes and no. If just formally calculate, yes, servers are small market volumes. But, they are much less constrained financially, than private person, so from same fab one could earn much more money if sell to server market, than if sell to consumer market.
For about price of land and connectivity - in large city land price begin on few millions dollars per square kilometer, and usage of cable channels could cost from 50$ per meter (easy could be 200$/m).
Plus, space arrange could last years.
Heat dissipation in range of megawatts could be just prohibited by local regulations.
So, space in large cities is very serious problem, and for business it is usually easier to "compress" as much computing power as possible in one rack.
You cannot place dc anywhere, in large cities space is extremely constrained, and land is extremely expensive.
Also big problem - connectivity - you cannot place DC where it cannot be connected to power grid and to very powerful network.
So yes, DC floor space is severely limited.
And the third issue - last decades, rack servers dissipate extremely large amounts of heat, I hear numbers up to tens Kilowatts per rack, which is just hard to dissipate with air cooling (as example, all IBM Power servers have option of liquid cooling, but this is totally different price range).
Also, internal CPU caches grow over time - in 286 and before, just was not any cache; in 386 first included page cache for mmu - stores tables with mostly used pages; in next generations sometimes advertised grow of page cache.
So yes, even when your cpu could address similar size of ram, possible it don't have enough page cache for your application.
Because very large share of market now are datacenters. Difference from desktop is dramatic - for desktop really acceptable very simple chips with bad energy efficiency, but DCs already deal with extremely high power consumption, as they typically "compress" so much consumption in one rack, that constantly working near to physical constraints.
Unfortunately, nobody at big companies know, what exactly math will win, so competition not end.
So, researchers will try one solution, then other solution, etc, until find something perfect, or until semiconductors production (Moore's Law) made enough semiconductors to run current models fast enough.
I believe, somebody already have silver bullet of ideal AI algorithm, which will lead all us to AGI, when scaled in some big company, but this knowledge is not obvious at the moment.
You could transfer different bands via different coils on different frequencies, but unfortunately, capacity of information channel is limited by frequency. Because of this, radio using high frequency waves as carrier (radio or light, or even some sort of invisible rays), not coil, and have hassle with some sort of modulation of waves.
This mean, you could not transfer more information than half of maximum frequency.
Very good channels with very high signal to noise ratio, could handle more bits than Shannon limit (on engineers slang "channel is ringing", such example is fiber channel).
Most modern research also consider some digital techniques of sound (information) compressing, like use LLM as (de)compressor (google llm compression algorithms).
Amateurs in USSR 50 years ago made wireless and powerless headphones, which use wire lay on perimeter of room to transfer sound and power.
In headphones there is tiny coil.
It really work and very reliable, but result coil (size of room) have very large reactive resistance, so it is nearly impossible to transfer even high frequencies, only low (bass) and medium, so it workable for speech but music is heavily distorted.
They treat you belong to community, and use your appearance in hidden ads as "just another consumer choose A.. products".
Even if you will intentionally hide all logos of A.. from A.. products u use, their design is very distinctive and widely known, so even looking on Xiaomi most people will think it is A..
Plus, A.. products usually deep integrated into their infrastructure, I mean A.. Wi-fi router, A.. printer, A.. speakers, A.. interfaces (Lightning), etc.
Well people, knowledge become belief. - Position of Moon creating tides, which affect some places very strong, because line of shore constantly moving more then by cup meters, so some operations need to be planned according to tides, this is not belief, this is fact. Position of Sun amplifies tides for up to 50%.
BTW, people knowing Astronomy, easy conclude, most extreme Sun tides amplification happen with Sun Eclipse.
I also seen few other wrong classifications.
Sorry, good idea, but such mistakes made it unplayable.
If possible for exact this plane, could make software update just as routine procedure.
But as I hear, air transporters could buy planes in different configurations, so for example, Emirates airlines, or Lufthansa always buy planes with all features included, but small Asian airlines could buy limited configuration (even without some safety indicators).
So for Emirates or Lufthansa, will need one empty flight to home airport, but for small airline will need to flight to some large maintenance base (or to factory base) and wait in queue there (you could find in internet images of Boeing factory base with lot of grounded 737-MAXes few years ago).
So for Emirates or Lufthansa will be minimal impact to flights (just like replacement of bus), but for small airlines things could be much worse.
> how do you avoid the voting circuit becoming a single point of failure
They do not.
Just make voting circuit much more reliable than computing blocks.
As example, computing block could be CMOS, but voting circuit made from discrete components, which are just too large to be sensitive to particles.
Unfortunately, discrete components are sensitive to overall exposure (more than nm scale transistors), because large square gather more events and suffered by diffusion.
Other example from aviation world - many planes still have mechanic connection of steering wheel to control surfaces, because mechanic connection considered ideally reliable.
Unfortunately, at least one catastrophe happen because one pilot blocked his wheel and other cannot overcome this block.
BTW weird fact, modern planes don't have rod physically connected to engine, because engine have it's own computer, which emulate behavior of old piston carburetor, and on Boeing emulating stick have electronic actuator, so it automatically placed in position, corresponding to actual engine mode, but Airbus don't have such actuator.
I want to say - especially big planes (and planes overall), are weird mix of very conservative inherited mechanisms and new technologies.
For about, what could be good machine for hacking, I have ideas.
First, probably, 8-bitness is unavoidable, because all those current Raspberries, are relatively powerful computers, even usually could install Android there, so you understand what I want to say :)
- Machine for hackers should be limited, on RAM, on CPU speed, sure, with limited screen resolution, and limited sound, because otherwise, on some point, will become race of wallets, as high quality picture and sound are usually expensive.
From other side, graphics should not be too primitive, looks like good compromise are C64 or Atari-65 (not many static objects on background, but with hardware accelerated sprites).
Some time before, I thought, the best balance for hackers machine is C64, until I read some details about Enterprise-128 (or 64).
What differs E128(64) - their absolute unique video-adapter, capable to show few resolutions on one screen. Imagine classic arcade game - for them very usual to have on top part of screen some static background and some indicators of achievements, and whole game process running on lower part of screen. So in good design, we should somehow make top part with minimal possible efforts, but focus on lower part; and E128/64 is most close hardware to this.
For about real implementations, I'm impressed with esp-32 rainbow, but unfortunately, it is ZX Spectrum simulator, and I think it is impossible to do on those hardware C64 or E128. When time will accept, I'll try other cheap hardware platforms, as I hear, RP2040 could run separate C64 chips (but nobody have done whole C64 on multiple RP2040s), so will be multicore machine, but it's ok.
Large scale DBA and Ops. For them typical daily solving tasks like "what is more reliable - two RAID-0 in stripe or two stripes in one RAID-0" - mathematics thinking gives exact answer.
Well, I seen one suggestion for probable nearest future of AI, to stop on GPT-4 level, but distill and use optimizations like switch to FP8 for faster speed.
So basically idea, model with very same capabilities, but distilled and optimized to have less size and faster inference.
I cannot promise 100x speedup, but I think 10x is very real.
Well, formally, you are right that Crimea is 90% Russian ethnically, but what about price?
- In 1940s, USSR totally resettled all Crimean Tatars to Siberia, and only in 90s they got permission to return home. For comparison, imagine US will resettle all people from Florida to Alaska - you will not name this genocide?
And yes, resettling of peoples, was typical for USSR, and Russia was main power in USSR, even when prime ministers was from other parts of Union, so demography policy was pro-Russia. For example, up to 1990s, Russia constantly resettle Russians to Eastern Ukrainian regions, and same way, now Kazakhstan, and other Asian exUSSR counties have large share of ethnically Russian population.
Your opinion is really interest and important, but it lack evidence.
What I mean, last years, I constantly gather information, on how countries grown their hackers community and programming industry, and at first I seen obvious things, but now with more info, things looking more complicated.
As example, I seen Eastern Europe and exUSSR, grown their hc and IT, mostly with cheap unlicensed local clones of PDP and IBM machines, and as I see, we in exUSSR know about Commodore/Atari and about consoles (I mean pre xBox), and we somewhere lack their taste, but we are already mature and even more or less competitive.
From other side, Japan have rich history of consoles and machines with extended graphic and sound capabilities (MSX, PC-98), and they have good achievements in enterprise machines and in hardware, but I don't see Japanese google, or Japanese facebook, or Japanese Oracle (databases). And I have not answer, how this happen.
What impress me even more - few years ago I got info, GDP in Asia/Africa and accessibility of compute devices grown very fast after Android appearance (I'm not sure if GDP connected, but smartphone became universal computing device with Android), so I see grow of games sales to Asia/Africa, why I notice - because their very specific culture, so gamedev have to made significant changes to game to enter their market. And this shocked me, as I release, Japanese just avoid to enter these markets, stay focused on their internal market and on West.
And BTW other side - East Europe and exUSSR was so poor, so having moderate access to really good Western computers, huge share of economy made accounting with pen and paper and abacus, some entities in middle 2000s. As I know, Japan have access to computers nearly as Americans, from at least middle 1980s, but looks like they lost something when most people switched from manuscript to keyboard.
So, to rotate faster, you need larger control surfaces.
From other side, traditionally, self-stabilize spent at least 1% of energy (on small planes normal up to 10%). What all this mean - with 10 000kg of total weight, your control surface will constantly make 1000kg of force to just fly, but when need to turn, will need significantly more than 1000kg, that's all.
Old planes need self-stabilize, because constant corrections was very time consuming, but modern have powerful computers and could provide artificial stabilization - current 1kg computer could provide same stabilization as control surface constantly making 1000kg of force.