Sasha [They/Them]

Yes, that Sasha 🍉

Transfemby 🏳️‍⚧️⬛🟪⬜🟨🏳️‍⚧
They/them

Anarchist/your local idiot with a guitar

If you’re occupying land in so-called “Australia”

If you eat food

And if you live on Earth

Introducing Trans Action Network Naarm! 🏳️‍⚧️
(Part of a wider solidarity network too!)

  • 0 Posts
  • 10 Comments
Joined 2 years ago
cake
Cake day: December 12th, 2023

help-circle

  • I’m pretty sure they touch on those points in the paper, they knew they were overloading it and were looking at how it handled that in particular. My understanding is that they’re testing failure modes to try and probe the inner workings to some degree; they discuss the impact of filling up the context in the abstract, mention it’s designed to stress test and are particularly interested in memory limits, so I’m pretty sure they’ve deliberately chosen to not cater to an LLMs ideal conditions. It’s not really a real world use case of LLMs running a business (even if that’s the framing given initially), it’s not just a test to demonstrate capabilities, it’s an experiment meant to break them in a simulated environment. The last line of the abstract kind highlights this, they’re hoping to find flaws to improve the models generally.

    Either way, I just meant to point out that they can absolutely just output junk as a failure mode.




  • AC’s use electrical energy to take some heat energy, and move it outside. You can kinda reverse the process in certain types of heat pumps to generate power, but it’s not even close to worth it, the efficiency is horrible.

    You need a temperature gradient to capture heat energy, basically a cold thing and a hot thing, you harvest the energy as it moves from hot to cold. You’ve cooled your house down, and want to use the waste heat to create power, you’ll either have to find a very cold place somewhere nearby (unlikely to be cold outside if you’re using AC) or you can use the fact that your house is cold. So now you’ve both lost energy and heated up your house, because that lost energy has been added to the heat you originally tried to remove.

    The simplified but always true rule of thumb is that whenever you use energy to create something from which you can harvest energy, you’ll never be able to harvest more than what you spent. In reality you’ll pretty much always lose energy trying to do this, I’m not aware of anything that’s 100% efficient in both directions (or even one honestly).




  • Sorry to nitpick a bit, dark matter isn’t connected to the expansion of space (as far as we’re aware) but dark energy is probably what you meant. My answer to your question is at the end.

    Full disclosure: While I have studied this, my expertise was in a tangentially related field. However a buddy of mine has a PhD in measuring this stuff so I’ve got some second hand knowledge.

    It’s a confusing hurdle for any student of physics to understand that spacetime doesn’t exist inside another bigger thing into which it can expand, it just kinda exists on its own. Mathematically we don’t even treat the expansion quite like growth, it’s a bit easier to understand it as our rulers getting shorter, the labels we give to distances changes over time. Personally I like the analogy of a sheet of grid ruled paper.

    If you choose two points and count the number of squares between them, divide that grid into a smaller one and then count them again, the “distance” has gone up. Those squares look smaller to us so it seems like the true distance is the same, but the universe doesn’t have an external view to make such comparisons from, all we have are the squares and physics obeys them. The point is you can cut squares up forever without running out of squares to cut up, nothing runs out this way.

    In spacetime maths (general relativity aka GR) we usually start by defining distances, and when it comes to the expansion of the universe we literally just have a number in that definition that changes over time.

    This kind of “our rulers and clocks are dodgy and unreliable” is unfortunately the backbone of this sort of physics. It’s a huge pain in the ass, but it’s cool af if you’re a huge maths nerd.


    How does it expand?

    🤷

    Anyone who can tell you how dark energy works beyond “it has a negative pressure” is full of it. It’s a theoretical idea and has never been observed, we just know that if something with negative pressure existed everywhere then it would cause space to expand. Don’t quote me, but it’s kinda like the opposite of how a black hole squishes spacetime down into a singularity, dark energy pushes out on everything everywhere all at once. (Couldn’t help myself it’s a great movie go watch it)

    There are a bunch of possible things that fit the bill, it could just be a number in Einstein’s field equation, it could be a specific type of quantum field that has a constant value everywhere, hell I’ve even seen models where it’s just caused by black holes existing. It’s also possible that Einstein got some stuff wrong and that expansion is just what space do. Either way, I don’t think these things require more stuff to be created, it’s just stuff that’s already there.

    If I had to make a mostly uneducated guess, I’d say it’s probably just a feature of quantum gravity, for which we have no proven theories. Loop quantum gravity just demands it exists for the theory to even be useful, I’m sure string theory has it’s own crazy nonsense to explain it too. If we ever do work this out, I fully expect it’s just going to be a thing we have to accept exists without an obvious cause, much like how the universe exists but we have no idea why or why the rules it follows are those specific rules and not some others.