Traditionally, economy are typically divided into three sectors: Agriculture, industry, and services. Service industry contains everything from nursing to software development and sales. The problem with this division is that there is an extreme productivity gap within work in the service industry. A software developer's work can serve 100 million people at a time, when a nurse can only serve one customer at one specific time.
The reason why highly developed economies have become so service driven is because they have become sort of bimodal: The cost of labor is such that only jobs that are productive enough (profitability per hour) are done in these countries, and jobs that absolutely have to be done there to sustain the population. Jobs in the middle, everything that is not highly profitable or location-dependent, is offshored to lower-cost countries due to the cost of labor. This results in these developed countries having issues: Cost of living is high due to labor cost and there's high economic inequality due to wildly differing productivity.
The solution would be to bring these "mid-productivity" jobs back to developed countries. However, the main roadblocks still remain: The cost of labor is too expensive for most of these jobs to be competitive globally. However, I think there might be a way to do this in the near future: Advancements in robotics would mean a higher level of automation for industrial work, meaning more industrial jobs would become viable in high-cost countries. Each worker would be productive enough that the cost of labor is not critical anymore.
To make this happen, I believe it's important to ensure that the country is viable for this kind of manufacturing: Energy supply needs to be abundant and cheap, workforce needs to be educated, outside the "elite" students, and there needs to be low trade barriers. Low trade barriers are needed, because virtually all manufacturing is part of a global supply chain where parts cross many borders before the product is sold (and (high-value) products are sold globally). Additionally, the viability of automation will vary between different parts of the supply chain, and so you likely cannot automate everything.
The fact that this area where the incident happened, Gulf of Finland, is not fully part Finnish/Estonian territorial waters, is only because of a bilateral Finnish-Estonian agreement. This was done in the 1990's purely for benevolence towards Russia.
Russia clearly hasn't acted in such way that they should enjoy these kinds of acts of benevolence. Finland and Estonia should seriously consider retreating from this agreement.
I don't think it's just benevolence. Territorial waters also doesn't mean what many think it means - unlike planes, ships have the almost-universally recognized right to cross territorial waters (innocent passage).
But what's more relevant here are rules about straits - territorial waters that fully enclose a section of someone else's territorial waters. My understanding is that that is a big part of the reason why the two countries restrict their claim of territorial waters to leave a corridor of international waters: They want to avoid the area falling under the straits rules (transit passage), which would give Russia more rights than it has now inside the territorial waters.
Yes, the right of passage through the strait would still clearly remain. This is already the case with Denmark and Sweden as these ships need to cross Öresund or Great Belt strait to reach the Atlantic.
However, this act would, in my understanding, give much more power to Finland and Estonia to detain these ships, and charge the crew for the crimes they have committed. Right now there seems to be a loophole in the legislation that Russia is actively exploiting for hybrid warfare purposes. If the strait rules would give Russia more ways to cause harm, some other way of dissuading Russia from making these acts should be done.
In general though, it feels stupid that we have to play by these rules, when the enemy makes a mockery of them and actively tries to exploit them to cause as much harm as possible. But that's the reality when bordering Russia.
There is a 1000+ km long front of active combat in Europe right now. A front where European shells and Russian ones are getting exchanged. Where F-16s fight Su-35s. And then we have things like the Russian cargo ship with nuclear materials that got sunk by a high-end torpedo. Just because shells aren't yet raining down on Berlin, it doesn't mean this war isn't kinetic.
Ukraine isn’t part of the EU, or historically part of the ‘European’ sphere (really meaning Western European). It’s historically been part of Russia. Or if you go back far enough, Russia was part of Ukraine.
It doesn’t completely negate your point, and anyone who isn’t seeing the writing on the wall is being willfully ignorant aka Chamberlain.
But culturally this is also a very different situation from France, Germany, England, Spain, or even Greece being shelled.
Which is also why people are so ‘meh’ on it, practically, and it’s taking so long to respond.
I don't see anything in your comment that would even argue with my point, much less negate it. That history lesson on Europe itself is pretty pointless, because if you go back a bit further you'd find much of Ukraine having been ruled by the Habsburgs - i.e. Austria. It doesn't get more European than that. And that short period of time where the Russians/Soviets ruled basically serves as Putin's propaganda reason for this war. That certainly doesn't belong here either.
Ukraine literally used to rule the land now known as Russia [https://en.wikipedia.org/wiki/Kievan_Rus%27]. Kyiv used to be the capital. I think you have your history confused, and your 'what is propaganda or not' reversed.
They are basically two brothers with a long history, with Russia having recently been on top - after previously being on the bottom - but then going bankrupt - and now trying to bully it's way back to being on top again.
Either way, they aren't France, Germany, England, Spain, etc. and have wildly different history. Ukraine isn't part of the EU or NATO (and in fact, the possibility they might eventually be is a big driver of what Russia is now doing).
Got to get that bullying in before it's too late, after all.
I already told you why that is not just wrong, it misses the point. It also doesn't help if you want to switch this whole historical legitimation propaganda to the other side, because it is meaningless anyway. We are not living in 19th or 15th or 10th century Europe anymore. Ukraine lies in Europe by any remotely recent definition and it has an actual kinetic front line. End of story. Everything else is malevolent political propaganda.
You should really go and read up on some geography, this is become embarrassing for you. Or spew your russian propaganda elsewhere, because if you are serious you have outed yourself now.
The issue is, the people who are (supposedly) in charge are also sounding increasingly hysterical and seem to be actively pushing for a NATO-Russia confrontation.
That is obviously insane, so I do wonder if there isn't something else going on beneath the surface
I hope so, but we have the head of NATO and numerous senior British officials (including the head of MI6, who is never normally heard from) talking about an impending major war. Maybe (hopefully) this is just hedging and something can be worked out
Innocent passage ≠ acts of negligence or sabotage. This sets an important precedent, that ships engaging in acts of sabotage could be be boarded, put under custody and their crews detained.
"The Federal Fiscal Court (BFH) said there was "reasonable doubt as to the legality of the confiscation measures," as it was unclear whether the ship had had authorization to enter and leave the EU despite the sanctions, due to an exemption applicable in emergencies."
No, he's saying that the area is international waters because Finland and Estonia agreed it was not either's territorial waters. It doesn't have to be international waters.
The US just doesn’t recognize China’s claims to areas (eg, Taiwan or ASEAN sea islands), so doesn’t regard those as Chinese territorial waters in the first place.
The point of US freedom of navigation exercises is to assert free transit of allied and international waters, despite Chinese claims, rather than to transit Chinese territorial waters. US warships generally avoid areas which the US views as Chinese territorial waters.
The fight in and around China's sea claims is they encroach into what the rest of the world generally agrees are other countries waters not international waters. The US would still insist it could travel through the Taiwanese or Phillipine waters China wants to claim as their own. It doesn't seem to map at all on to the situation between Finland and Estonia.
When you're a country as small and insignificant as Estonia is you're not doing anything out of "benevolence" towards a nuclear hyper-power, but what do I know?. maybe the Maja Kallas-types really do believe in their own word-blabber.
Hyper power that can't overwhelm a country that was supposed to fold in 48h? Give me a break.
While your sentiment may be correct in 2010s it certainly was not when these things were being decided in early 90s. USSR and Russia which de facto ruled it was seen as a failed state that needs "western help" and on a path to democracy. While we (here in Poland) we're quite skeptical, having the Russian WW2 occupying force leave in 1991 (yes, we didn't get freedom after WW2 until 1991). There was still a lot of hope Russia will follow in the footsteps of other central/eastern European countries like Poland/Czech/The Baltics if only we help them. So yes, there was huge resentment, but also a huge benefit of hope and benevolence too.
Was some of this calculated? Sure. No doubt someone sat in Talin and Helsinki and thought: if we treat them like post WW1 Germany it will be easier for the extremists to take power. So let's not pour sand in their fuel tank as they are desperately trying to restart the engine of their economy.
I don't even think it was a mistake at the time. It was a decent way to behave. But the moment the tide has started turning in Russia towards autocracy the screw should've been tightened. No oil and gas should fund Russian army after at least their attack on Georgia. If not before when the atrocities of the Chechen war became known.
Unfortunately corrupt politicians (that are still in power in Europe and even in my country) have continued signing deals and making money by financing what was clearly a huge enemy in the making.
Russia wasn't an eny in 1993, but it certainly was one in 2008 when it invaded Goergia. If only we acted properly in early 2000s all of this could've been prevented.
Zelensky was arguably a nobody internationally, until he lead his country to stalling out the entire military might of the Russian Federation in a war that's only a few months away from being longer than the Great Patriotic One, and keep on giving Russia bloodey noses like taking out a chunk of their strategic bomber fleet, the underwater drone strike on Novorossiysk, and tanking the Russian economy. Not bad for a literal comedian.
Hyper-power, seriously? Russia is not even a super power, it only has some nuclear weapons, that’s all. Just like India, Pakistan, France, Israel, etc. In all likelihood most Russia’s nuclear weapons aren’t even operable anymore.
I think you just made a lot of people aware of it. Although I doubt people care all that much, unlike you do.
It's pretty insulting; as if people don't have agency and can't form their own opinions. No, it can only happen due to some external boogeyman, which seems to be NAFO in your case.
In my opinion, C++ syntax is pretty readable. Of course there are codebases that are difficult to read (heavily abstracted, templated codebases especially), but it's not really that different compared to most other languages. But this exists in most languages, even C can be as bad with use of macros.
By far the worst in this aspect has been Scala, where every codebase seems to use a completely different dialect of the language, completely different constructs etc. There seems to have very little agreement on how the language should be used. Much, much less than C++.
I'd say over 70% of all computing is already been non-CPU for years. If you look at your typical phone or laptop SoC, the CPU is only a small part. The GPU takes the majority of area, with other accelerators also taking significant space. Manufacturers would not spend that money on silicon, if it was not already used.
> I'd say over 70% of all computing is already been non-CPU for years.
> If you look at your typical phone or laptop SoC, the CPU is only a small part.
Keep in mind that the die area doesn't always correspond to the throughput (average rate) of the computations done on it. That area may be allocated for a higher computational bandwidth (peak rate) and lower latency. Or in other words, get the results of a large number of computations faster, even if it means that the circuits idle for the rest of the cycles. I don't know the situation on mobile SoCs with regards to those quantities.
This is true, and my example was a very rough metric. But the computation density per area is actually way, way higher on GPU's compared to CPU's. CPU's only spend a tiny fraction of their area doing actual computation.
> If you look at your typical phone or laptop SoC, the CPU is only a small part
In mobile SoCs a good chunk of this is power efficiency. On a battery-powered device, there's always going to be a tradeoff to spend die area making something like 4K video playback more power efficient, versus general purpose compute
Desktop-focussed SKUs are more liable to spend a metric ton of die area on bigger caches close to your compute.
If going by raw operations done, if the given workload uses 3d rendering for UI that's probably true for computers/laptops. Watching YT video is essentially CPU pushing data between internet and GPU's video decoder, and to GPU-accelerated UI.
HN in general is quite clueless about topics like hardware, high performance computing, graphics, and AI performance. So you probably shouldn't care if you are downvoted, especially if you honestly know you are being correct.
Also, I'd say if you buy for example a Macbook with an M4 Pro chip, it is already is a big GPU attached to a small CPU.
Switch has its own API. The GPU also doesn't have limitations you'd associate with "mobile". In terms of architecture, it's a full desktop GPU with desktop-class features.
well, it's a desktop GPU with desktop-class features from 2014 which makes it quite outdated relative to current mobile GPUs. The just released Switch 2 uses an Ampere-based GPU, which means it's desktop-class for 2020 (RTX 3xxx series), which is nothing to scoff about but "desktop-class features" is a rapidly moving target and the Switch ends up being a lot closer to mobile than it does to desktop since it's always launching with ~2 generations old GPUs.
Only if you're ignoring mobile entirely. One of the things Vulkan did which would be a shame to lose is it unified desktop and mobile GPU APIs.
In this context, both old Switch and Switch 2 have full desktop-class GPUs. They don't need to care about the API problems that mobile vendors imposed to Vulkan.
In AI training, you want to sample the dataset in arbitrary fashion. You may want to arbitrarily subset your dataset for specific jobs. These are fundamentally opposed demands compared to linear access: To make your tar-file approach work, the data has to ordered to match the sample order of your training workload, coupling data storage and sampler design.
There are solutions for this, but the added complexity is big. In any case, your training code and data storage become tightly coupled. If you can avoid it by having a faster storage solution, at least I would be highly appreciative of it.
- Modern DL frameworks (PyTorch DataLoader, WebDataset, NVIDIA DALI) do not require random access to disk. They stream large sequential shards into a RAM buffer and shuffle within that buffer. As long as the buffer size is significantly larger than the batch size, the statistical convergence of the model is identical to perfect random sampling.
- AI training is a bandwidth problem, not a latency problem. GPUs need to be fed at 10GB/s+. Making millions of small HTTP requests introduces massive overhead (headers, SSL handshakes, TTFB) that kills bandwidth. Even if the storage engine has 0ms latency, the network stack does not.
- If you truly need "arbitrary subsetting" without downloading a whole tarball, formats like Parquet or indexed TFRecords allow HTTP Range Requests. You can fetch specific byte ranges from a large blob without "coupling" the storage layout significantly.
Highly dependent on what you are training. "Shuffling within a buffer" still results in your sampling being dependent on the data storage order. PyTorch DataLoader does not handle this for you. High level libraries like DALI do, but this is the exact coupling I wanted to say to avoid. These libraries have specific use cases in mind, and therefore have restrictions that may or may not suit your needs.
AI training is a bandwidth problem, not a latency problem. GPUs need to be fed at 10GB/s+. Making millions of small HTTP requests introduces massive overhead (headers, SSL handshakes, TTFB) that kills bandwidth. Even if the storage engine has 0ms latency, the network stack does not.
Agree that throughput is more of an issue than latency, as you can queue data to CPU memory. Small object throughput is definitely an issue though, which is what I was talking about. Also, there's no need to use HTTP for your requests, so HTTP or TLS overheads are more of self-induced problems of the storage system itself.
You can fetch specific byte ranges from a large blob without "coupling" the storage layout significantly.
This has exact same throughput problems as small objects though.
When GPUs started being used for deep learning (after AlexNet), GPUs were not at all matmul machines. They were machines that excel in most kinds of heavily parallel workloads. And this holds to this day, with the exception of the tensor core, which is an additional hardware block designed to accelerate this specific task.
Matrix multiplication didn't "win" because HW was designed for it. It won because matrix multiplication is a fundamental part of linear algebra and is very effective in deep learning (most kinds of functions you might want to write for deep learning can be expressed as a matmul). Acceleration of it became later. Additionally, matrix multiplication is a good fit for physics, as you can design the HW so that data movement is minimized, and most of the chip area and power are spent in actual computation, and not moving data around.
Fundamentally speaking, you also want to make your algorithm compatible with real-world physics. The need for heavy parallelism is required by the fact that you cannot physically make a fast chip that processes dependent operations. It's just not possible to propagate signals through transistors fast enough to make it possible. Even CPUs, even if they present a non-parallel programming environment, have to rely on expensive tricks like speculative out-of-order execution to make "sequential" code parallel to make it fast.
In general though, I personally would wish that chips would be made with taking programmability in mind. A fixed-function matrix multiplier might be slightly more efficient than a parallel computing chip with smaller matrix multipliers. But it would be significantly more programmable, and you can design much more interesting (and potentially more efficient) algorithms for it.
It should be around 25 ms in normal conditions. That's what I got when pinging Hetzner in Germany, from Finland, when the cable was still in use and when using a connection that routes through the cable.
I am don't use Hetzner, but I use ssh between Finland and Germany every day. As a matter of fact even back and forth because of tunneling. After reading the news this morning (Hetzner incident is date 3:30 UTC) I was surpised that I had not noted any lag. It remained very reponsive all day.
I have a persistent VPN tunnel between Finland and Germany and I’ve not noticed really any disruptions. If it had cut out for even a moment, it would’ve interrupted my services (since they don’t recover gracefully at the moment) and I would’ve found out.
Specifically this is being pushed by Ylva Johansson [1] from Sweden, who has (reportedly) financial connections to the organisation Thorn which is hoping to sell this chat monitoring software.
> Only the Commission can propose new legislation.
That is a technicality relying on a shallow look at the word "propose". The commission frequently takes direction from the council when deciding what to focus on which leads to them "proposing" legislation. In this case the push for this has come from the Council and certain national governments combined with a particular commissioner.
The Commission agenda and mandate is set by the Council that nominates it, and periodically reviewed by the same Council. Items are set in meetings, the agenda of such meetings is typically public.
If the Commission pushes, it's because the Council told it to push.
The Commission is formed by commissioners, each of which has been nominated by a different nation's government. The Council consists of actual members of the national governments.
The third party that can't propose legislation but has to approve it (and which strongly opposed this) is the Parliament, which is directly elected.
OK, this is probably some nuance of the English language that I'm missing as a non-native speaker, but I meant that the people that make up the Commission are not part of the national governments. The people that are part of national governments each get to nominate one commissioner though, in addition to being part of the Council.
The reason why highly developed economies have become so service driven is because they have become sort of bimodal: The cost of labor is such that only jobs that are productive enough (profitability per hour) are done in these countries, and jobs that absolutely have to be done there to sustain the population. Jobs in the middle, everything that is not highly profitable or location-dependent, is offshored to lower-cost countries due to the cost of labor. This results in these developed countries having issues: Cost of living is high due to labor cost and there's high economic inequality due to wildly differing productivity.
The solution would be to bring these "mid-productivity" jobs back to developed countries. However, the main roadblocks still remain: The cost of labor is too expensive for most of these jobs to be competitive globally. However, I think there might be a way to do this in the near future: Advancements in robotics would mean a higher level of automation for industrial work, meaning more industrial jobs would become viable in high-cost countries. Each worker would be productive enough that the cost of labor is not critical anymore.
To make this happen, I believe it's important to ensure that the country is viable for this kind of manufacturing: Energy supply needs to be abundant and cheap, workforce needs to be educated, outside the "elite" students, and there needs to be low trade barriers. Low trade barriers are needed, because virtually all manufacturing is part of a global supply chain where parts cross many borders before the product is sold (and (high-value) products are sold globally). Additionally, the viability of automation will vary between different parts of the supply chain, and so you likely cannot automate everything.