Anecdotal example, though I'm sure I'm not the only one:
We had a C++ service. It needed high availability, but didn't have super high resource requirements. Our setup was an ec2 instance (c5.xlarge) to build and release AMIs (a bash script using debootstrap to build the AMI which someone wrote 10 years ago probably), an autoscaling group of 3 t2.small instances spread across AZs, and an ALB.
The total cost of this was perhaps $200/mo, and we had incredibly good uptime.
So, what's the catch? Well, it took the service about 20 minutes to build, and about 35 minutes from clicking the "deploy new version" to it actually running. Someone higher up noticed, and a bright-eyed infra engineer said k8s would make the deploy cycle faster.
Fast forward a year. Our autoscaling group is now 3 c5.xlarge instances because the kubelet + docker + coredns + all this other k8s gunk I don't understand need significantly more CPU than our app does (and without giving them more CPU, deploy times were much slower since downloading and unpacking the image was so slow). We have a new logging system (our old logging setup wasn't cloud native apparently) that takes a gig more memory per node. A gig of memory per node to support our service, which peaks at 200MiB RSS. Building and deploying a new version still takes about 35 minutes because compiling C++ is the exact same speed in a dockerfile as it is on an ec2 instance.
It costs about $600/mo, and it has far more operational load. When it isn't having any issues, the p99 is identical.
> better performance or lesser costs
It seems like the opposite of what you'd expect. K8s is adding more components. It's adding more resource usage. Why wouldn't that be slower and cost more?
The OP made a good point that applies to all poor countries I've lived in: those jobs are the only (even the best) way out for most people to get out of poverty, which they do to help their families/children/themselves.
Maybe you were raised in a family that did not need to go through such hard labor, but that doesn't mean your right in your view of your own world ;)
P.S. I was also born and raised in a poor country.
So I grew up in China and I can agree that the law is applied selectively. If you are a foreign corporation and have not paid the right bribes or in good books of the ccp official in your area in charge woe behold if you violate any labor rights. On the other hand if you're friends with the right people you can basically operate a slave camp.
That being said I do agree that our attitude of "lets shut down here cuz it doesnt work for us" does more harm than good. It's not like this will solve corruption. If anything, this will give Chinese people more reason to hate on USA and India, thus further cementing their governments power.
The point that “those jobs are the only (even the best) way out for most people to get out of poverty" stands. But the point that China has strict labor laws is laughable. The laws are indeed strict but seldom enforced.
This is true of most strict rules elsewhere: they're there but seldom enforced. Take speed limits, they're strict, but good luck enforcing them. All we do is we monitor from time to time and give tickets. Yet we could enforce this to the manufacturers, right? Why can a car go beyond the speed limit if it's strictly prohibited?
We cannot compare working conditions as is, but they're easy to compare when you take time into account: how where our labor laws a couple of generations ago? Not that different than poor countries today.
The problem with preconceptions about your parameters is that you might be missing some crazy cool path to your goal, which you might find by randomly exploring your sample space. I remember seeing this same principle in mcmc samplers using uniform priors. Why is this so crazy?
It's predicated on the assumption that a random discovery from a zero-comprehension state is more likely to get you to a goal than an evolution from a state that has at least some correctness.
More generally, it disingenuously disregards the fact that the definition of the problem brings with it an enormous set of preconceptions. Reductio ad absurdum, you should just start training a model on completely random data in search of some unexpected but useful outcome.
Obviously we don't do this; by setting a goal and a context we have already applied constraints, and so this really just devolves into a quantitative argument about the set of initial conditions.
(This is the entire point of the Minsky / Sussman koan.)
> from a zero-comprehension state is more likely to get you to a goal than an evolution from a state that has at least some correctness.
I get that starting from a point with "some correctness" makes sense if you want to use such information (e.g. a standard starting point). However, such information is a preconceived solution to the problem, which might not be that useful after all. The fact is that you indeed might not at all need such information to find an optimal solution to a given problem.
> by setting a goal and a context we have already applied constraints.
I might be missing your point here since the goal and constraints must come from the real world problem to solve which is independent from the method to solve the problem. Unless you're describing p-value hacking your wait out, which is a broader problem.
With exploring, the starting state should only affect which local-maximum you end up in. Therefore you need to make an argument that a random starting state is likely to end up in a higher local-maximum than a non-random starting state.
There is always a starting state; using a random one only means you don't know what it is.
Slaveholders were a minority of the population who managed to structure all of society to suit their wishes to the detriment of everyone else. As you might have gathered, I’m opposed to this sort of thing.
As you might have gathered, I’m opposed to this sort of thing.
In that case perhaps we could compare the number of drivers in the United States with the global population and then talk about the environment? The US is a major contributor to global climate change and the US is exceptionally bad compared to other developed nations in GHG contributions arising directly from motor vehicle use. I'm not interested in contrived arguments about slavery but if you want to take a principled stand against a small minority forcing their preferences and the adverse consequences of those preferences on a much larger majority then unfortunately you are on the wrong side of this issue.
If your team lead can protect you they will but the employer at large is rather indifferent. Also further in your career those relationships to your team mean a fair bit while who you worked for as a company might not make that much difference.
I’ve gone back to work for a manger who laid me off. There were literally no hard feelings.
When he laid me off it was clear that he had to hit a hard headcount number, and I knew the project I was working on was “discretionary”. The HR meeting was “this is a headcount reduction and not a reflection on your work. Have a lawyer look over your severance and please accept or decline it within a week.” Really quite professional.
Others have pretty much covered the matter, you will always bump into former team mates, or it will be thanks to them that you will get some gig.
Employers themselves usually look into spreadsheets with a bunch of KPIs deciding who to lay off, without any regards for the effort you have actually placed into the job.
Nothing interesting will be reliably funded over the next 5-ish years given the current macroeconomics.
If we’re talking about where the opportunities (jobs) are going to be, then you’re probably looking at tech roles within non-tech companies. These companies have been dying to modernize but haven’t been able to hire engineers due to the tech bubble.
After that, tooling that enables non-technical companies to build software - whatever that looks like.
The really high paying jobs of the future (well, and really now to be honest) are going to be some sort of combination of Data Science + Speciality Science. Think biomedical engineering, material scientist, chemist, any engineering discipline because the thing we need the most right now are better medicines and antibiotics against the rising threat of resistant bacteria, COVID showed us we still don't have a shot against a really bad virus, we need better batteries, better power generation, better cars and modes of transportation, etc.
What we don't need any more of is web cruft and CRUD apps, social networks, and people figuring out more ways to mine our data and shove ads in our face.
> What we don't need any more of is web cruft and CRUD apps, social networks, and people figuring out more ways to mine our data and shove ads in our face
(No snark intended, my background is in science...)
The fields that would benefit society the most are not typically the fields where the most money is to be made.
My guess about "can be trapped/filtered" is in theory and possibly, with a lot of effort, in practice. But then of course -- how much easier it is (and how much more financially-incentivized it is) to let the pollution out rather than to filter it.
We need good incentives with a system that self-corrects; we need to stop playing whack-a-mole with easily-known-ahead-of-time problems and implement good systems.
Modern island nations (Japan being one of the first) have been doing clean incineration through Arc Reactors / Plasma gasification for energy for decades.
I was curious as to how effective scrubbing was, turns out Packed Tower Wet Scrubbers can filter out 99%+ of VOCs from incinerated plastic. You still get carbon emissions however.
At the scale humanity is using plastics, 99% is lacking after-comma 9s in orders of magnitude. We're at 400 million tons of plastic waste a year [1], so even 1% inefficiency means that the VOCs of 4 million tons end up in the environment.
Yes, it's better than burning it outright in open fire pits (as it's done in an awful lot of "third world" countries to concentrate metals out of the waste), but it's way worse than a dump yard. And most of the places that have a lot of plastics waste don't have the resources to install decent scrubbers, much less perfect scrubbers.
Yes and no. If the energy is recycled (for heating buildings and/or water) you'll reduce carbon emissions from coal/gas/oil. Waste incineration (instead of landfills) is a thing in various European countries and the heat can (and will) be used to heat buildings or the steam drives turbines to generate electricity.
