yokaze

Werner Vogels seems to disagree: https://twitter.com/Werner/status/25137574680

> After removing punctuation, space symbols, newlines and tabs, we filtered out documents with less than 200 characters. These documents typically contain only meta data and no useful information.

> But i guess "upsampling" in this case is just explicit duplication of the training data.

Possibly, but duplication means weighing and that is important in unbalanced trainingsets and improves the results in practice.

Well, in Spain you can use your eID directly: https://github.com/OpenSC/OpenSC/wiki/DNIe-%28OpenDNIe%29#up...

And you can get a standard X.509v3 certificate from the FNMT (https://www.fnmt.es/en/ceres), or other agencies, which identifies you online for government purposes.

"The Rolls Royce/Prada/... of beers"

Champagne is a brand for sparkling wine. People pay a premium (justified or not) for it.

My examples are even more unrelated than the beer, still the companies would likely defend their brand.

Miller is trying to pull itself up with the brand Champagne and is pulling the brand down. (If Miller is champagne of beers, what does it say about champagne as a sparkling wine)

> cause I progressively want to get closer to the metal (one day work with embedded systems)

Don't see the point in going for desktop apps. If you want to broaden your skills, that's great, but it's not a progression in skills.

For both server side programming as well embedded development, UI is less of a concern, and you control often the whole platform/OS.

Desktop apps require skills in UI development and also you have to bother to learn multiple platforms. Libraries and languages only do so much in providing an abstraction.

Also it depends on what you understand with Embedded systems. They range from full fledged windows/Linux/... systems (quite close to desktop apps and more UI programming needed) to PICs.

In my view, the programming language is the least of the concerns and a small part of the skill set you'll need to acquire.

Choose the tools you feel comfortable with in the problem space you want to work in.

There is micropython as well.

Except that time is signed integer, so

   $ date -I --date='@-2147483648'
   1901-12-13

I think, it is reasonably explained here: https://en.wikipedia.org/wiki/Enlargement_of_the_eurozone

They are required to adopt the Euro, as soon as they fulfill the Convergence criteria Those convergence criteria include "Compatibility of legislation", and Sweden simply doesn't make their laws compatible. (Details here: https://ec.europa.eu/info/sites/default/files/economy-financ...)

In summary, they have to join, if they are compliant. For that they have to pass laws, and they chose not to. So, it isn't quite as mandatory as some people paint it.

To quote from the report:

> The Swedish authorities are invited to remedy the abovementioned incompatibilities

(2012)

Yes, that's the intent of the unit: https://en.m.wikipedia.org/wiki/Sievert

Also had a (failed) startup in Germany, with a slightly different perspective. Possibly outdated.

> It still baffles me, that you need to buy your own shares in Germany, in order to receive the “status” and protection of being a fully incorporated company.

It depends on how you read it, but it isn't true that you have to put in 25.000 in cash on the table. A GmbH is a Limited, which needs to have a value of at least 25.000EUR, so anyone trading with you knows, you are at least good for that amount.

The easiest way is to put the money in cash into the account of the company.

But it can also be done via transferring the ownership of either tangible or intangible assets into the company. The problem here is, especially with software, that the evaluation of assets is not for free and difficult.

> Now, you don’t need to convert the UG to a GmbH but if you want to be taken seriously in Germany by customers and investors, it can be a huge factor doing business. For our investor, it was a must-have.

The feedback I got from institutional investors was the following:

- If you are not willing to put down 25k, while asking for millions, it doesn't send much of signal of confidence in your proposal from your side.

Maybe it is a matter of German risk averseness, but I think it is reasonable, especially paired with the next advice:

- You will take a top level management position in your new company, your salary should reflect that. You have to strike a balance of not asking too much (you are after all a startup), but also not asking enough (you need to know the market value of your position). But chances are, the increase in your salary by your "promotion" should pay your initial investment in two years time.

In Germany, composition is much more salary based.

> I don’t know what to tell you, other than this process can take up to 3–4 weeks to get finished. Even most challenger banks (neo banks) are not really faster.

> How do I know? We’ve incorporated another holding company UG and wanted to use a Berlin based challenger bank — it took us almost a month to get done.

> Choose your bank provider wisely or at least switch early. We went with a 0$/month neo bank and it was horrible.

Yeah, we went with a traditional physical bank. Costs a fair bit more than 0€/month, but it took a couple of days, maybe?

> A mail popped up along the lines of: “We would like to do a 10 minute interview in Mountain View with you. We compensate your travel expenses”

And I think, that is the most striking difference in culture between the US (or at least SV) and Germany. Those guys where trying to safe a couple (granted couple of thousands) Euros, and a couple of Euros per month for the bank, and the investor is paying the travel costs probably coming close to the founding costs for a 10 minute interview.

I think, you are misreading it.

It isn't about making an architectural design decision / change at 3am, but someone has to live with the consequences of such a design decision and fix something in some way or another at 3am in the morning.

I.e. you could go with a clustered db setup which is supposedly HA, or a classical single instance DB...

The point here is, whoever has to get up at 3am to fix a potential catastrophic failure in the setup should have a say in it.

> And a singleton is not an instance?

No, a singleton is a pattern that ensures that there is at most a single instance of a class.

> There is no perfect design pattern

There I can certainly agree, but ...

> An exe is a singleton, a file is a singleton. Most things in a computer are singletons.

I would say you are mistaking the instance for the class. An individual exe/file/... are instances of the abstract concept file or executable.

What you are interacting with is usually an instance, not a class.

I'm missing, which settings have been chosen for rocksdb, and why.

I dearly hope, not just the defaults.

AFAIK, the car development pipeline is 5 years from initial concepts and designs to actually selling it.

For me, that means there are only three years left to have something like that on a prototype level.

Yes, the design is routinely changed to make the currencies more safe against falsification. And the old currency is often taken out of circulation simply by wear and tear. But it is the sign of a stable economy that legal tender does not have an expiration date.

"Note that it is U.S. government policy that all designs of Federal Reserve notes remain legal tender, or legally valid for payments, regardless of when they were issued. This policy includes all denominations of Federal Reserve notes, from 1914 to present as per 31 U.S.C. § 5103." https://www.uscurrency.gov/acceptance-and-use-older-design-f...

Despite not being legal tender anymore, the Deutsche Bundesbank changes your DM notes, if they are after issued after 1948 (https://www.bundesbank.de/en/tasks/cash-management/dm-bankno...). The currency has been introduced in 1948.

If you feel inclined to do so, you can spend a 1 Yen note from 1885 https://en.wikipedia.org/wiki/Banknotes_of_the_Japanese_yen

So, inflation and wear and tear takes old notes out of circulation rather than expiry.

> I have utmost sympathy for Ukraine but where were the angry people when US attacked Iraq?

"On February 15, millions of people protested, in approximately 800 cities around the world. Listed by the 2004 Guinness Book of Records as the largest protest in human history". There is a whole page on all the protest for that particular topic: https://en.wikipedia.org/wiki/Protests_against_the_Iraq_War

"That's libertarians for you — anarchists who want police protection from their slaves." - Kim Stanley Robinson, Green Mars

Or technically a character from the book.

But I think, that's the point of the article: Smaller government for the part of government which doesn't concern you. But that in turn might look differently for people in different circumstances.

Demographics (to avoid the term "race") affect your view if police or other parts of government are essential or rather part of the problem .

I use violent monkey, because it's open source and supports multiple browsers. Spent only minimal amount of research on the choice though.

I think, there is a possible grain of truth there.

>> If there were classified orders taking up fab space, would we ever know?

The shortages could have led to larger back-orders, and potential "double booking". Those orders is probably a trade secret, so it probably won't be easy to find out.

Just before the pandemic, everyone was optimising for zero inventory, and probably now you have a feedback loop which probably overshoots a bit in the other direction too.

You need something in half a year or a year, you likely order it right away instead of doing it just-in-time.

My guess: It is lowered expectations, and people accept the excuse.

So, what would you tell your customers, if you can't deliver on time?

Not sure if your example is considered to be complete, but AFAIK, static strings with the same content are the same instance. I'm side stepping your question in a way, but you can validate that easily with the "is" keyword.

  for k1,k2 in itertools.zip_longest(d1, d2):
    print(k1 is k2)

Some details, which are in my view not unimportant:

- The tobacco use is not only affecting the person "choosing the lifetime of pleasure", but their surroundings too, thanks to second hand smoke.

- Nicotine is highly addictive:

"The majority of smokers would like to stop smoking, and each year about half try to quit permanently. Yet, only about 6 percent of smokers are able to quit in a given year" [1]

For me, that makes it highly questionable, if that is a voluntary choice. I know enough people, which have chosen to indulge in that particular pleasure, and after decades of abstinence still have to put a conscious effort into not relapsing (in certain situations).

[1] https://www.drugabuse.gov/publications/research-reports/toba...

And among those twelve countries you have: United Arab Emirates, Saudi Arabia, Bahrain, Oman, Belarus, Kuwait, Qatar

Personally, I would prefer my country not to be in same exclusive club .

Paris has a population density of 20000/km², and Manhattan of 28,872/km².

Population density is correlated with car density up to a certain point.

If you read this, or other things about reducing car traffic in Barcelona, one should keep the following in mind: https://ajuntament.barcelona.cat/lafabricadelsol/en/noticia/...

6000cars/km² in Barcelona.

San Francisco/Oakland metropolitan area has 4000/sq mi... 1540/km²

Guttenberg, NJ has the highest in the US with 20,600/sq mi (8000/km²)

Source: https://www.google.com/amp/s/www.planetizen.com/node/45622%3...

The author has clearly an opinion. I am getting more and more the feeling, that IT is a field where opinions are strongly hold, mostly because of past personal bad experiences, and not with much evidence.

There are many ways to approach a problem, and there are good and bad ones, but often you have to make a trade-off. I would appreciate a more differentiated view.

The author makes the math on "high cost of maintenance windows", but leaves out the expenditure side to it. And what is the hand-rule for adding a nine behind the availability? I vaguely remember exponential cost increase. "No downtime, ever", requires next to infinite cost.

The author suggest a false dichotomy. You can have maintenance windows and while aiming for zero downtime. The maintenance window serves the customer a higher predictability on when to expect failure. No failures are obviously preferable, but it is never the question if to cause failures or not, but where you spend your resources on.

So, we are in agreement in that point: if you are going that way, you are relying on the physical switches.

I'm not sure about the problem being solved, depending on the scale.

If you are with your use-case (number of rules, macs, five-tuples, ...) outside the capabilities of the hardware, then, in my experience, it breaks down quickly. (I.e. flooding, or becoming unresponsive to configuration commands and needed to be _power cycled_, ...).

And those are switches costing easily five figures, and from two different well known vendors.

If you have practical experience with a setup, which scales more gracefully, please share.

But I'm sceptical that it exists, considering Google, Facebook, AWS all went with custom hardware.

I'm wondering, if at a certain scale of virtual NICs/flows per port, you are not better off with a hybrid approach, where software is part of a decision hierarchy. E.g Openflow.

No practical experience though.

In my experience with said expensive hardware, it greatly depends on your use-case / scale.

You have to pay attention to stay within the limits of the hardware the switches (TCAM, etc...). And correct me, if I am wrong, but most NICs are fairly constrained in their hardware support on the numbers of MACs they can filter for. And those two combined, you might end up with a lot of flooding, which needs to be handled by the CPU again.

You also have to rely on the implementation of your vendor. And at least with one major vendor, who shall remain nameless, it is a major effort just to get acknowledged that there is a bug.

With a vaccine, you get a standardised dosis of an agent resembling the disease or just a part of it. The remaining variability is the response to that standardised input.

With an infection, you have no control over the degree of exposure, so on top of the variability of the response comes the variability of the exposure.

Additionally, in case of exposure of the full pathogen, there is some variability to what part of the pathogen the immune system will respond. That part can actually be a good thing in total.