larsnystrom

I assume that’s why Altman wants the government as a co-owner.

I’ve only used Opus in GitHub copilot and was hugely underwhelmed. It was barely usable. Are you saying it’s better with the official Anthropic tools?

Sure, from a theoretical perspective, but in practice there’s got to be some sort of order at some point even when storing timespans.

You could also make the price column nullable and just insert a row with price null and the date from which there should be no price.

Why are they storing a time period (start and end date) in the first example? Why not just store the date when the price comes into effect? That would make both overlaps and time travel impossible without using any constraints.

By that logic VPNs and many other technical solutions are also not solving technical problems, since it is theoretically possible to achieve the same results by other means.

Just use array_values() and suddenly you can use int indicies again.

Yeah, at the moment we use arrays as anonymous objects and phpdoc+phpstan to verify the types, but I want it in the language. PHP already supports intersection and union types, it really feels like just skipping the naming part and going all in on structural typing is not that far fetched by now.

I've been writing PHP for 20 years now. It's my bread and butter.

The one thing I really wish PHP would add is structurally typed objects. I really miss it when moving back and forth between PHP and TypeScript.

They could call them anonymous objects if they want to (that would be a more culturally correct analogue to anonymous classes).

Like, I wish it was possible to do

  {
    string $mystring = $myvar,
  }

  new class($myvar) {
    public function __construct(
      readonly public string $mystring,
    ) {}
  }

  function ({ string $mystring } $myobj) {
    echo $myobj->mystring;
  }

  type myobj_type = { string $mystring };

I don’t think AI demand is driven by consumers.

What a great blog post! I have wanted to do rootless docker with subuids, but putting it all together like you have is not easy. Thank you for writing it down!

My comment was a bit tongue in cheek. Obviously it is a hard problem. But in a profession where we work with machines that literally were made to crunch numbers, and where abstraction is something we deal with daily, why can’t we have a performant abstraction for doing arbitrary calculations? The answer is that to be performant it must be solved in hardware, which would cost more than the hardware we have.

So in fact it is not just telling me it’s a hard problem, it’s telling me that the cost-benefit is still not there. It’s like it’s just not a very important problem (in an economic sense). And that is what surprises me, given that computers were made to do arbitrary calculations.

I don’t understand how dealing with numbers correctly is not a solved problem in computer engineering by now.

I remember now what the problem was. I was instantiating a new ArrayList in a loop. The solution to the performance issue was to use a Vector instead. I was used to just writing PHP arrays when I wanted a list of something, and since they’re dynamically sized I thought the analogue in Java was ArrayList, which is also dynamically sized. But somehow that was extremely unperformant in Java.

I remember writing Java for our introductory programming course at university around 2010. I was already familiar with object oriented programming in PHP at the time, so I just wrote the Java code like I would write PHP. I was absolutely astounded at the poor performance of the Java app. I asked one of our tutors and I can still remember him looking at the code and saying something along the lines of ”oh, you’re instantiating objects in a loop, that’s obviously going to be slow”. Like, what? If I can do this performantly in freakin PHP, how can Java, the flagship of OOP, not have fast instantiation of objects? I’m still shaking my head thinking about it.

Maybe I’m just dense, but it says the fs module is fully supported, so what happens when I try to read a file from disk if the app is fully sandboxed?

I can really relate to this. At the same time I’m not convinced cycle time always trumps throughput. Context switching is bad, and one solution to it is time boxing, which basically means there will be some wait time until the next box of time where the work is picked up. Doing time boxing properly lowers context switching, increases throughput but also increases latency (cycle time). It’s a trade-off. But of course maybe time boxing isn’t the best solution to the problem of context switching, maybe it’s possible to figure out a way to have the cookie and eat it. And maybe different circumstances require a different balance between latency and throughput.

WASM is kind of like Java bytecode, right? A compilation target which requires a runtime? But better sandboxed?

It’s still possible to order new and original parts for SAAB models, almost 20 years after they went under. The spare parts are made by a separate company which is still going.

Human eyes are much better than cameras at dealing with dynamic range. They’re also attached to a super-computer which has been continuously trained for many years to determine distances and classify objects.

There seems to be a huge business opportunity in Europe right now, to sell support and customization of open source software to government players. Has anyone heard about a European company that’s been successful in this area?

On the other hand, many things attributed to chance are actually the aggregate effect of other people's choices. If we make choices based on not just what's best for ourselves but what's best for all of us, we will all suddenly become more "lucky". And vice versa, if we only think about ourselves that luck will diminish.

And AFAIK there is still no upload progress with fetch.

Nice to only have to push the layers that changed. For me it's been enough to just do "docker save my-image | ssh host 'docker load'" but I don't push images very often so for me it's fine to push all layers every time.

Yes, it is because of NATO. They've been thinking about it for some time, but NATO tipped the scales.

Please do share some links to these events, because this is the first I hear of it.

Sure, but TOTP still defends against password leakage. So it’s still more secure than only using a password.

Allowing third party payment systems does not mean they have to be less strict in the rest of their app vetting process.

But this is from a non-Apple app. You’re already on a page outside of Apples control.

Why are they using any language at all? Why not just let the app maker open the payment screen directly?