pdonis

51 ms·

631.
by pdonis 2y ago
A calculator can be trusted to give you the right answer if you punch in the numbers correctly. An LLM cannot. That's because a calculator is designed to give the right answer. An LLM is not. It's only designed to produce text t
632.
by pdonis 2y ago
> the speed of light is constant That's not a good way of describing the actual law in a curved spacetime, i.e., a spacetime that contains gravitating masses. In such a spacetime there is no single global definition of "speed
633.
by pdonis 2y ago
> Does the idea of a single cosmological time even make sense? Sure, it's just a convenient choice of coordinates. Even in a model that is not exactly homogeneous and isotropic, it can still be a convenient choice of coordinates t
634.
by pdonis 2y ago
> If they are empty, those billion years didn't happen. No, that's not correct. Here's a better way to look at it: In our cosmological models, we "slice up" the spacetime of the universe into slices of "sp
635.
by pdonis 2y ago
> there is a kind of relativistic Tsiolkovsky equation for mass-energy propulsion vehicles that carry their own fuel Yes, it's called the relativistic rocket equation. https://math.ucr.edu/home/baez/physi
636.
by pdonis 2y ago
> You haven't met anyone that has quality issues but doesn't realise it? Lucky you! We're not just talking about someone who has code quality issues. We're talking about someone who has code quality issues that have
637.
by pdonis 2y ago
> everyone is rejecting their code That's not what "rollbacks" means. "Rollbacks" means other people accepted their code, let it get into production, and then an issue surfaced that forced the change to be
638.
by pdonis 2y ago
> The main sentry app takes ~10 seconds to pip install. Which is much longer than the "couple of seconds" I gave for my use case. Yes, if it takes that long, I can see how you would want some alternative. > Also more pai
639.
by pdonis 2y ago
> Meanwhile, deployments keep getting rolled back? One could say that the scenario as described in the post I originally responded to, by localghost3000, doesn't make sense, yes. If there have already been multiple rollbacks, and
640.
by pdonis 2y ago
> the new manager is technical and familiar with the codebase, and can assess that the reason for the rollbacks is a genuine quality issue. This is useful information Only if the report doesn't already know it. But they probably d
641.
by pdonis 2y ago
> I’ve noticed quality issues in your code recently that’s resulted in some rollbacks I would tend to even leave out the first part of that phrase. Focus on the actual objective measure: the rollbacks. They happened, and the goal is to
642.
by pdonis 2y ago
Why? It works for me.
643.
by pdonis 2y ago
Orders of magnitude longer than what?
644.
by pdonis 2y ago
> It’s also a step not needed by most other ecosystems. From what I can gather, most other ecosystems don't even have the problem under discussion.
645.
by pdonis 2y ago
> it’s potentially very slow. Potentially, perhaps. But it's certainly not for the cases where I use it: a pure python package, whose dependencies are already installed and are not changing (only the package itself is). Under thos
646.
by pdonis 2y ago
A lot of the problem seems to be driven by a desire to have editable installs. I personally have never understood why having editable installs is such an important need. When I'm working on a Python package and need to test something,
647.
by pdonis 2y ago
Even with the smaller number, there are still only a tiny fraction that kill people.
648.
by pdonis 2y ago
> You're not supposed to run red lights Even if it's an empty intersection and I can clearly see that there is no traffic coming, and the stupid light is staying red for way too long simply because the local government can&#x
649.
by pdonis 2y ago
> I don't like that my kids can no longer cross the street we live on because people have stopped obeying traffic lights and cops no longer enforce the law. The best solution to living in a place with bad people in it, if you can&
650.
by pdonis 2y ago
> Speed limits and red lights being on the honor system isn't working. Speak for yourself. I don't need a nanny monitor in my car to drive safely. Nor do the vast majority of drivers. There are about 150 billion cars in the
651.
by pdonis 2y ago
It will happen if enough voters make it clear to their elected representatives that they want them to repeal the laws passed after 9/11 that mandate the security screening that TSA does. I agree that's highly unlikely, but it was
652.
by pdonis 2y ago
How well have the predictions made by reasoning from those axioms worked out? It's true that the most egregiously wrong prediction in the Federalist papers, that the US would not have political parties because it was too large, was the
653.
by pdonis 2y ago
> We absolutely have statistical and theoretical frameworks for evaluating collective behavior. Yes, we do, but their predictive power, if it exists at all, is very poor. Science is about building models with good predictive power. If
654.
by pdonis 2y ago
> The disdain engineers have toward the "soft sciences"[sociology, psychology] While I agree that disdain is not the right reaction, neither is misdescribing what these disciplines are. And the most important fact about them
655.
by pdonis 2y ago
> we know No, we don't know . Most physicists believe it, but that's not the same as knowing. We won't know unless and until we are able to actually do experiments in the relevant regime.
656.
by pdonis 2y ago
The issue in that paper isn't really limited to "singularities". The basic issue is that we use the continuum in our physical models, and "reality" might not actually be a continuum, so the continuum math we use is
657.
by pdonis 2y ago
> Every physical theory with singularities has has broken down in that regime. Can you give some examples? Note that GR has not even been tested anywhere close to the regime you are talking about. > It's not even clear what it
658.
by pdonis 2y ago
> GR has singularities. More precisely, GR allows spacetime solutions which are geodesically incomplete. > It's definitely wrong in those regimes. No, that's too strong a claim. Most physicists believe that the solutio
659.
by pdonis 2y ago
> for edge cases GR is wrong because it doesn't agree with quantum mechanics What "edge cases" are you talking about? I agree that GR is not a quantum theory, but it's not established that that has to be a problem,
660.
by pdonis 2y ago
> The ejection was months before he got the role at VMX-1. But he had already been a qualified test pilot--you have to have had that experience for a number of years before you will even be considered for a post like the CO of VMX-1. S

More ›