A few years ago, the biggest problem with Erlang's hot code updates was getting the files updated on all of the nodes. Has this been solved or improved in any way?
I don't think updating files is the problem. The biggest issue with hot code updates seems to be that they can create states that cannot be replicated in either release on its own.
This is my experience. About 25% of the time I'd encounter a bug that's impossible to reproduce without both versions of the code in memory, and end up restarting the node anyway dropping requests in the process. Whereas if I'd have architected around not having hot code updates I could built it in a way that never has to drop requests
In general, you can save your team a lot of ops trouble just by periodically restarting your long running services from scratch instead of trying to keep alive a process or container for a long time.
I’m still new to the erlang/elixir community and I haven’t run it in prod yet but this is my experience coming from Java, Node, and Python.
There's about a thousand different ways to update files on servers?
You can build os packages, and push those however you like.
You can use rsync.
You could push the files over dist, if you want.
You could probably do something cool with bittorrent (maybe that trend is over?)
If you write Makefiles to push, you can use make -j X to get low effort parallelization, which works ok if your node count isn't too big, and you don't need as instant as possible updates.
Erlang source and beam files don't tend to get very large. And most people's dist clusters aren't very large either; I don't think I've seen anyone posting large cluster numbers lately, but I'd be surprised if anyone was pushing to 10,000 nodes at once. Assuming they're well connected, pushing to 10,000 nodes takes some prep, but not that much; if you're driving it from your laptop, you probably want an intermediate pusher node in your datacenter, so you can push once from home/office internet to the pusher node, and then fork a bunch of pushers in the datacenter to push to the other hosts. If you've got multiple locations and you're feeling fancy, have a pusher node at each location, push to the pusher node nearest you; that pushes to the node at each location and from there to individual nodes.
Other issues are more pressing; like making sure you write your code so it's hotload friendly, and maybe trying to test that to confirm you won't use the immense power of hotloading to very rapidly crash all your server processes.
I think Twitter once cobbled together a BitTorrent based deployment strategy for Capistrano called murder, that was a cool read from their eng blog back in the day.
I wish I had used a pusher node to deploy things when a colleague was using almost all the upstream bandwidth in the office making a video call when my bosses were giving demo and the fix I coded for an issue discovered during the demo could not deploy via Capistrano
Stripe thinks that AI is now ready to not only read data but also act on it and generate payment things.
Because AI is still prone to hallucinations and making stuff up, I am a bit concerned about letting agents do things like adjust pricing tiers automatically, negotiate payment plans, or alter things that may not be entirely legal to do.
Yes, depending on thresholds which vary by state. For some you have to figure out taxes from the first transaction, for others from the 100th and then there are some that consider revenue, like $100k. At least at a conceptual level, taxes in the US seem much more complex than the EU, where you have only one tax rate per country.
I wonder if the reason it's perceived as "easy" is because companies are flying under the radar. In my experience, you get a bit of leeway until the country you're selling into starts to notice and demands you start collecting and paying sales tax. Perhaps the fact the US is split across jurisdictions internally means tax authorities don't have a very clear picture until the company gets big enough.
Either that, or more companies are using third parties who mask the complexity for them. Kinda like a duck looking serene on the water but its feet are paddling furiously to keep moving.
When you're inside a jurisdiction, you don't really have the option to ignore the law for long because the tax authorities will come knocking sooner rather than later.
Yep, I think that's the best quote of the interview and the runner up is the Napster quote.
People seems to have forgotten that Intel used to had StrongARM in their lineup, with the same logic Intel is a computing company not x86 company, similar to AMD [1].
For one of the latest trends in the computing landscape please check this new Embedded+ platform recently introduced by AMD [2]. I'm biased towards IoT but I've a very strong feeling that IoT is the killer application of AI, similar to AI is a killer application of HPC as the latter statement was mentioned by Lisa Su during the interview. The turning point for IoT however is yet to happen, it will probably happened when the number of machine nodes talking to directly to each other, surpassing the number of machine nodes talking to human, then it will be its ChatGPT moment.
It has huge implications because it makes competing in that market much harder due to licensing issues. Intel and AMD have a duopoly on the x86 market which compromises a huge chunk of server and personal computing, but that is changing fast. If they go ARM (or risc-v or whatever) they will have more competition to contend with, including their existing cloud computing clients designing their own chips and fabbing them with other foundries.
There was an interview with one of the SPARC creators who said that a huge benefit of control of the instruction set was the ability to take the platform in directions that (Intel) resellers could not.
He was otherwise largely agnostic on the benefits of SPARC.
Loose seatbelts are probably fine. That’s what I do. Presumably that keeps me from getting launched into the ceiling. A lap belt is a compromise anyway. Aircraft crew use more as do car drivers for that matter.
Right but what's the point? If the belt is on very loose, you give your body a chance to gain momentum, which will be redirected when the seat belt catches you, which would lead to serious injury.
The belt doesn't need to be on tight, but it should be relatively snug.
It's just one of those things that is trivial to do, with a very, very low probability to save you from injury, but costs nothing, so why not...
I'm going by the levels of "loose" I regularly see on flights. I regularly see seat belts buckled, effectively fully slacked. It's not everyone, but probably 5%.
You won't gain additional momentum from a loose belt. But a loose belt may give your body the chance to slip out. The impulse you feel from a tight belt will be no less than the impulse you feel from a loose one.
Impulse is constant, its the force that is transferred to your body. This is constant whether or not a belt is worn.
A belt simply takes that force and redistributes it into the seat frame, if the belt is not worn the force is not redistributed by the belt but rather by your head into the overhead bin
The impact is a matter of energy, not of force- if the force applies for more than an instant, you keep accelerating before hitting the ceiling and the energy of the impact will be proportional to the square of your speed. If the belt is tight, you stop at a lower speed (relative to the aircraft) and the energy of the impact (with the belt, in this case) is lower.
It's like debt. If I'm standing on the ground, I'm constantly "making payments" to the ground. If I jump off a cliff, then while I am falling, I stop making payments, so when I next contact the ground, accumulated debt suddenly must be paid.
More precisely, while I am falling, my velocity and consequently my momentum relative to the Earth steadily increases, which is analogous to debt because I will eventually need to restore the situation in which my momentum relative to the Earth is zero (because that is how people live their lives: at rest relative to the surface of the Earth rather than for example in orbit).
In the same way, the passenger in the airplane must eventually go back to being at rest relative to the airplane, but if the seat belt is very loose, "debt" (momentum relative to the airliner) can be accumulated before all the slack has been taken out of the seat belt.
Yes, the total amount of impulse (change in momentum caused by the environment's pushing on the person) integrated over time is the same regardless of how tight the seat belt is, but it matters whether that impulse is spread out over time or comes in one big jolt.
That doesn't make sense. Force is not accumulated by people standing on the ground, so your later analogy does not work at all
The ground opposes the force produced by gravity, there is no accumulation of anything. Gravity accelerates mass when it is pulling it in motion unless there is an equal and opposite force which is the normal force exerted by the ground (or airplane)
This behavior is reflected in car attitudes too. Many Japanese drive with a detached seatbelt.
Not even the presence of children changes this calculus.
Most Dutch people don’t wear a helmet when riding a bicycle. As e-bikes are becoming more popular, some older people are opting to wear a helmet since the speed is higher, but many still don’t.
It’s not uncommon to see a parent riding around with a kid attached to the bike and neither are wearing a helmet.
I think it is kind of a cultural thing as well as convenience. If you ride a bike to the station to catch a train, then you might need storage or to carry it with you.
Pricesely this, yes. And for the kids, an average trip to school takes like 5-10 minutes, hassling with three helmets (mine and of two kids) takes almost as long. And then I have to carry three helmets if I want to drop by to the supermarket after getting them to school.
The bicycle seats for kids are designed in such a way that the only real danger of a head being hit is in the frontal collision, or in a really forceful side one. The design of Dutch streets makes the chance of either happening small enough that the helmets are widely seen as not worth it.
I thought that Dutch parents put small kids on the front of their bicycle to absorb the blunt of impacts. Then you dispose of the small child and make another one.
Also it's just not really a big danger. There's no epidemic of brain injury due to this.
Part of the reason is that because bike usage is so common and the infrastructure is very separate (bike lanes everywhere, no mixed traffic on eg roundabouts), and bikes almost always have right of way and are not to blame in an accident, that drivers are very aware of them. I noticed in countries like Ireland with much less bikes that bike lanes sometimes just end in the middle of a high speed roundabout. And most drivers there have no clue which is in part because until 10 years ago you could just buy a driver's license at the post office. It was a "learner's permit" but it was normal practice to just go and drive.
So yeah in Ireland I wouldn't even ride a bike with a helmet. In Holland I'm much safer even without one.
I don't remember the "Not Just Bikes" YouTube video, but he specifically discusses the relatively low rates of bicycle helmet wearing, especially amoungst urban riders. I agree with you: When there are serious bike injuries, it is usually due to a car or truck in mixed space. When only bikes riding less than 25km/h, the injuries are pretty minor.
Actually that seems very Japanese in a weird way - they might be trusting themselves and everyone else to follow rules precisely, thus avoiding all the accidents and obviating the need for seatbelts.
Is driving actually kind of easy in Japan for this reason? I’ve never had a car while there. It’s wild how different Japan and China are despite being neighbors. I can’t think of a European neighbor pair that is so different.
One is an island nation that's been fairly open, the other had waves of isolationism, famines, and only recently opened up their markets in the late 80s. Private car ownership was not on the table until probably 25 years ago.
But to play your game, Switzerland and Italy, especially southern Italy, are very different.
Your counterexample is not even close to being accurate. Italy and Switzerland have much more in common than China and Japan. Italian is even one of the official languages of Switzerland. Switzerland and Italy are both essentially federations of distinct provinces. Both are recognizably Western European. China and Japan are worlds apart, as are China and India.
My main point was that there's plenty of reasons for China and Japan to be different, especially when it comes to driving norms. Are you surprised as the other poster was about these differences between China and Japan?
And going back to the original intent: would you say driving in Zurich and driving in Naples are two distinct experiences?
I can: Germany and (European) Russia. They are about the same distance as Tokyo to Shanghai. They have wildly different driving cultures! Also, China is the size of a continent. Did you drive in all the different regions? I doubt it. Do what region are you extrapolating from?
