I guess the benchmarks are run on nightlies from 3.15 dev branch? It doesn't say on the website.
I did some tests with 3.15-dev on my own reference benchmark a few days ago and noticed that the JIT is finally making a positive impact.
As it looks now, 3.14->3.15 will be the biggest release-to-release Improvement since 3.10->3.11. At least for streaming text processing with significant amounts of pure-python logic.
Other EV:s have service inspections as part of the warranty requirements. That means they get inspected by workshops, which means that problems are more likely to be first found during the government inspection.
I don't think the actual quality difference under Equal conditions is a large as the TUV report suggests.
You pay the same proportion, not the same amount. 40% of 1M is 10x more than 40% of 100k.
Disregarding all technicalities about what proportion people actually end up paying after performing clever tax planning.
Why are you sure that someone earning 1M should have higher proportion of their income taken away than someone earning 100k?
At some sufficiently low level of income I think it stops making sense collecting taxes, but beyond that I'm not so sure from a fairness-perspeective.
I could perhaps get on board with a hard cap on wealth, for preserving democracy. It is dangerous to have single individuals and families attain too much power. But up to that cap, I don't see any inherent unfairness or inefficiency in that people of moderate to high wealth pay the same proportional rate.
I agree about wealth cap. I suspect that ultra-rich would rather pay higher taxes than have a hard cap. Or cap will be defined in such way as to become meaningless. This is why I advocate for high tax brackets, as being more realistic in practice.
As for proportionality - taxes are inherently unfair on the individual level. But they are fair on the large society level. All of our niceties are essentially funded from taxes. Current "free market" plus corruption mean that to finance growth more and more money is needed and part of them come from more and more taxes. But a lot of basic goods are fix price or low enough price, so that they make le and less percent of person expenses the more income rises. So to be more fair, it is fairer to increase tax on the ultra-rich class and spare tax increase on poorer classes, making average suffering lower. If we simply increase all taxes, then the lower the income the more tax a person would pay. It in kinda unfair and not productive to do it this way.
What I am trying to disagree with, is the notation that it is unfair if high-earners and very-high-earners hand the same proportion of their wealth and income over to the government.
My take on the wealth-cap is that it isn't about fairness at all. Actually I think it would be mostly unfair, but that it would be good for society anyway. Fairness is an important value, but it is not the only value.
You and I differ on the meaning of "fair" then. With regards to taxation "fair" is often taken to mean "what one can afford".
I think it's safe to say that those with over $100M will still have a plenty of dosh left even at a much higher tax rate than someone living paycheck to paycheck.
If tax revenues are seen as a pie chart—sliced based on contributions from the wealthy and the paycheck-to-paycheck, moving away from a progressive tax system means the poor will be asked to contribute more, the wealthy less. That seems like the opposite of fair to me.
No need to bring the poor into the discussion. Of course they can't afford paying an equal share of their income as someone with average or above average income.
In the grandparent post I was first replying to, the poster stated that they were in the highest tax bracket that existed in their country, but said there should be more brackets for people who were even more well off.
My opinion is that it isn't unfair to the top 1% that the top 0,1% have the same tax rate, even though they are richer.
I do think it is unfair if the bottom 25% have the same tax rate as the top 25%.
Only if Tesla is able to roll out a competing service. Given that they have zero cars without a safety driver on public roads, I’d say they’re a very long way from doing so, and I have my doubts about their ability to do so at all. Their CEO talks big but doesn’t deliver.
the people competitive in this space - your Waymo's, the Chinese auto co's (banned in the US), Mercedes | GM (won't do this as it cuts into their main profit lines) - so that will only leave Waymo as the only player in automated ride hailing
That Linksys card feels out of place in an over-the-top late 90s build.
As I recall it, the local LAN scene had an almost religious cult around 3com 10mbit ISA cards, that eventually morphed into a similar thing for Intel 100mbit cards.
Drivers and hardware were even more shit back in those days than today. Cards known to have worked in multiple motherboards and across multiple operating systems were held in high regard.
You've just brought back an old memory there. We used to do LAN parties at the weekend, and none of us has network cards, but one of our friends worked in the IT department of the local university. So he would bring a stack of cards with him and we'd spend hours setting them up and getting each PC on the network, often with great failure if he couldn't bring a batch of same branded cards.
We did learn over time which cards hated connecting with each other, and if he could bag a full batch of 3com we knew we'd likely be in for some early gaming.
A 3rd party NE2000 is a pretty reasonable choice. Certainly, some circles would have a following for specific makers and boards, but the NE2000 was everywhere... There's a reason it's so common in virtual machines; the open design helps too.
At the cheap end we relied on NE2000 ISA card clones, we couldn't afford fancy gear like 3com and they worked under Windows/Linux just well enough to play LAN games.
Thanks for the that bit of info! I was surprised by the speed difference. I have always assumed that most variations of basic string formatting would compile to the same bytecode.
I usually prefer classic %-formatting for readability when the arguments are longer and f-strings when the arguments are shorter. Knowing there is a material performance difference at scale, might shift the balance in favour of f-strings for some situations.
From what I've seen of Skywise, it is just a glorified SharePoint. Different systems upload CSV files that get turned into database tables. Then you can define views across these tables that other systems can consume by having them dumped to CSV and dropped on an SFTP.
Performance is not great, so you need middleware and batching anyway. As far as I am concerned, it wouldn't be a great loss if Skywise disappeared and just the SFTP with CSV:s remained.
Even that isn't generic and broad enough. I've noticed so many people mean SaaS when they say cloud. That isn't even a hardware or server or infrastructure meaning. It's referring to a whole cohesive IT product that you subscribe to.
Actually I'd say "cloud" says more about the business model than it says about the actual product.
I did some tests with 3.15-dev on my own reference benchmark a few days ago and noticed that the JIT is finally making a positive impact.
As it looks now, 3.14->3.15 will be the biggest release-to-release Improvement since 3.10->3.11. At least for streaming text processing with significant amounts of pure-python logic.
reply