What a bizarre comment. Take something like NumPy - has a hard dependency on BLAS implementations where numerical correctness are highly valued for accuracy and require deep thinking for correct implementation as well as for performance. Written in a different language again for performance so again an LLM would have to implement all of those things. What’s the utility in burning energy to regenerate this all the time when implementations already exist?
This is a great way of framing it that I'd never thought of before.
I worked in engineering software for a long time and because of who we sell to, there's always been a very hard cost-benefit analysis for customers of SaaS in that space. If customers didn't see a saving equal to more than the cost of the software in Y1 they could and would typically cancel.
This was the norm many years ago, I worked on a simulation software which existed long before Protobuf was even an apple in it's authors eyes. The whole thing was on a server architecture with a Java (later ported to Qt) GUI and a C++ core. The solver periodically sent data in a custom binary format over TCP for vector fields and things.
> As a HPC developer, it breaks my heart how worse academic software performance is compared to vendor libraries (from Intel or Nvidia). We need to start aiming much higher.
They're optimising for different things really.
Intel/Nvidia have the resources to (a) optimise across a wide range of hardware in their libraries (b) often use less well documented things (c) don't have to make their source code publicly accessible.
Take MKL for example - it's a great library, but implementing dynamic dispatch for all the different processor types is why it gets such good performance across x86-64 machines, it's not running the same code on each processor. No academic team can really compete with that.
I'm not asking an academic program first published 8 year ago (e3nn) to beat actively developed CuEquivariance library. An academic proposing new algorithms doesn't need to worry too much about performance. But any new work which focuses on performance, that includes this blog and a huge number of academic papers published every year, should absolutely use latest vendor libraries as baseline.
The big thing for me has always been (a) reliability of the hardware (b) good performance/battery trade off (c) nix-like environment.
In my prev. job I had a windows laptop with WSL2 though and I actually was super productive with that. But the laptop hardware offerings at the same price point are rubbish, just not very robust. Linux machines if you're in a corp and want one in the next 6 months are usually even more restrictive on hardware than they are on Windows.
It’s well known that living conditions went down for the average person as they moved into cities and industrialised. So the average living condition of someone living in London for e.g. being lower than that of a farmer in a country less far along on the industrialisation journey isn’t that surprising really.
If you had land yes. For a landless laborer in a rural area the conditions weren’t necessarily that great either. Of course population growth played a significant factor too.
I found with FB if you’re in some community that uses it then that’s it, you either use it or miss out.
For a long time my running group used it, and while it still does, the WhatsApp community is more used now. My (Catholic) church still uses Facebook for many announcements along with its own website.
You're probably being downvoted as most here are conflating WhatsApp and FB as both are owned by Meta so you're distinction is moot in the context of this discussion.
The body that made the recommendation, the "Safety and Advisory Board" met several times and changed their report multiple times. When it was finally released they redacted large parts of the decision making process including saying that:
* The police didn't want the match to go ahead (prior to any evidence for that)
* Two local Muslim councillors (Labour and Lib Dem) had been lobbying against it going ahead with one saying (quote) 'we are the voice of the people'
Additionally:
* They edited the report saying risk to local muslim residents went from Medium -> High
* They edited the report saying risk to fans travelling went from High -> Medium.
* They adjusted the number of police needed from 1200 -> 5000 in order to try and justify the decision.
When the full unredacted report was leaked, then they were put on the back foot and falsely threw out that they'd got the evidence (including of local muslim residents in Amsterdam being thrown in a river, which didn't happen) from a Dutch Police report, which wasn't true.
Anyone with a brain in the police should know that recommending cancellation or banning away fans from a Champion's League game is a major international news
story. The chief of police needs to be on top of the details and 100% sure that the evidence is there.
The Dutch police were a lot more fair-handed in giving intelligence to the WMP than you've been here.
They made clear that there were indiscriminate antisemitic attacks, and that WMP had made up claims that were not backed by what the Dutch police told them.
> a section of the Maccabi fan base was filmed engaging in violence in Amsterdam in 2024 and chanting anti-Palestinian racist abuse. That required WMP to contact their Dutch counterparts, who also informed them of the antisemitic violence by locals in Amsterdam, hunting down and kicking Maccabi supporters, leading to the only five convictions.
> Dutch police disputed the accuracy of how their Birmingham counterparts used information about the 2024 unrest in Amsterdam, with clear contradictions only able to be highlighted due to leaked WMP documents.
> WMP's intelligence assessment claimed that Maccabi fans apparently intentionally targeted Muslim communities in Amsterdam, but the Dutch force told me: "We did not see large groups of Maccabi's (fans) going into Muslim populated areas to target Muslims."
> Claims that Maccabi fans threw "innocent members of the public into the river" were also not endorsed by the Dutch.
There is quite possibly an argument that the match should have been cancelled/away fans not issued tickets, the issue is the cover up and dishonesty which is why he has to go.
