@PaulineHansonOz SHAME
Australia needs radical solutions
Net negative migration or piss off. Let the economy contract, it already is per capita anyway - migration is just a bandaid
using web tech is one of the easy ways to write UI code that is portable across many environments, including:
- browsers
- desktop
- mobile
- VSCode extensions
- other "embedded" UI environments
I guess it always *feels* like the easiest way to write reusable and portable UI while avoiding having to roll your own
@hollowearthterf@icefire99 incorrect
maternal infanticide occurs in multiple species and it can be evolutionarily advantageous for various reasons
it is already normalised in humans via abortion, which is not really that evolutionarily different from killing a postpartum child
@Mikerav6@UsingLyft if she divorces you then that's an excuse to just get even more women pregnant
she can be stuck raising your children, sucks to suck
abundance mindset versus scarcity mindset
@ErbunnNinja@gfodor Why do people trust smartphones even if they don't understand most of the technology behind it?
People will see that it ultimately works and that's why they will trust it.
@ErbunnNinja@gfodor Yes
Many proofs today are already beyond the understanding of most humans anyway
Math experts had already been using formal verification to reduce the chance of human error in proof checking - even before AI. Machine proof checking is not a new idea.
We already have the problem of proof transparency with humans anyway.
Some human-written proofs are so complex that only a very small set of humans have the expertise to be able to check that it is valid - and it's always possible they make a mistake.
Machine proof checking is strictly an improvement on this. It's much easier for me to convince myself that the core of a proof checker like Lean is correct so that I can trust a Lean proof of a complex theorem than it is for me to check the original proof myself - which may be beyond my understanding.
Automated proof checkers like Lean are based on a small core called the "kernel". This is the critical part of the proof checker that needs to be correct for the proof checker outputs to be correct. The kernel is designed to be as compact as possible so that it can ideally be checked by hand.
In theory as long as the kernel of the proof checker is correct, and the hardware executes the proof checker code faithfully, then the proof checked should be able to verify arbitrarily complex proofs.
The full proof might be totally incomprehensible by any one human, but ultimately it can be decomposed into individual steps that can all be checked by the proof checker.
These systems (like Lean) allow theorems or definitions to be reused. I.e., you start with a core set of axioms, then prove some theorems, then those theorems essentially become new more complex axioms that you can use to prove more theorems. Each step along the way is checked by the proof checker.
A proof consists of applying a sequence of allowed transformations that convert a set of sentences already known to be true into the sentence that states the theorem.
Since the formal language is inherently precise and discrete, following a fixed set of rules, you can write algorithms that only accept valid sequences of transformations, which can be used to verify that a proof is correct.
Note that these algorithms require the proof to be stated in a way that the algorithm can clearly see is correct. In other words, valid verification algorithms will always be conservative - they'll reject some correct proofs that aren't "obvious" enough. It's up to the human or AI to write the proof precisely enough that it can be verified.
Think of it like algebra, you are simply applying a limited set of allowed rules to transform one equation into another, allowing you to "prove" things like, "x = 2" from premises like "2x + 1 = 5".
@ErbunnNinja@gfodor Not really. The ceiling is just that formal verification can only be applied to formal reasoning - but mathematics is formal reasoning.
In theory any formal mathematical proof should be machine checkable almost by definition.
@ErbunnNinja@gfodor The proofs to these theorems are machine verified (by a proof checking algorithm - Lean in this case - not by another AI).
Humans don't need to be able to understand the full proof to be able to understand the verification process that proves that a statement is correct.
@JFGariepy it doesn't work if there's not a cost imposed on defectors, which is what patriarchy did
you can't get women to en masse opt in to this, not until they get older and their sexual bargaining power goes down
@aruvinchan ok but does "romantically attracted" actually mean anything on an emotional level or is it just a cold calculation of self benefit (if the latter, then the term is misleading)
@CaudilloNuclear they're hideous, or they're cowards, or just don't actually want women that much (they don't have that dawg in them)
has nothing to do with knowledge