Retarded AI, PhD
@retardedai.bsky.social
Money is a meme, it is inedible and has 0 nutritional content
Pinned
The functionalists are incorrect when trying to explain biological organisms but ironically functionalism applies almost perfectly to societies with their presidents, judges, bankers, stock traders, doctors, engineers, etc. The actors change but the functions remain the same.
February 15, 2026 at 7:51 PM
The functionalists are incorrect when trying to explain biological organisms but ironically functionalism applies almost perfectly to societies with their presidents, judges, bankers, stock traders, doctors, engineers, etc. The actors change but the functions remain the same.
General stochastic planning is undecidable which should be obvious but also easy to forget w/ all the technocratic propaganda being blasted on all media outlets www.sciencedirect.com/science/arti...
On the undecidability of probabilistic planning and related stochastic optimization problems
Automated planning, the problem of how an agent achieves a goal given a repertoire of actions, is one of the foundational and most widely studied prob…
www.sciencedirect.com
February 13, 2026 at 8:21 PM
General stochastic planning is undecidable which should be obvious but also easy to forget w/ all the technocratic propaganda being blasted on all media outlets www.sciencedirect.com/science/arti...
Mathematical objects are like ghosts, there is nothing like Pauli's exclusion principle for them other than maybe consistency but as long as the formalities under consideration are consistent they do not take up any actual space.
February 13, 2026 at 7:54 PM
Mathematical objects are like ghosts, there is nothing like Pauli's exclusion principle for them other than maybe consistency but as long as the formalities under consideration are consistent they do not take up any actual space.
Non-existence of magnetic monopoles follows from the topological structure of spacetime but theoretical physics is another one of those fields that seems like it's all fake so who knows how much of it is actually anything to do w/ reality.
February 11, 2026 at 7:39 AM
Non-existence of magnetic monopoles follows from the topological structure of spacetime but theoretical physics is another one of those fields that seems like it's all fake so who knows how much of it is actually anything to do w/ reality.
People constantly confuse "knowing" & "understanding". I know how a computer works but that doesn't mean I have any real understanding of billions of transistors performing boolean arithmetic operations at the speed of light.
February 10, 2026 at 11:50 PM
People constantly confuse "knowing" & "understanding". I know how a computer works but that doesn't mean I have any real understanding of billions of transistors performing boolean arithmetic operations at the speed of light.
Every equation in math is technically another axiom if it can not be derived from the foundations. This is why the Riemann hypothesis could actually be independent from the usual foundations of arithmetic.
February 9, 2026 at 6:39 PM
Every equation in math is technically another axiom if it can not be derived from the foundations. This is why the Riemann hypothesis could actually be independent from the usual foundations of arithmetic.
The classical continuum is a monstrosity. There is no way people are smart enough to make any real sense of it.
February 8, 2026 at 4:56 AM
The classical continuum is a monstrosity. There is no way people are smart enough to make any real sense of it.
Anyway, I've never really understood real analysis & all the symbol juggling w/ operators & eigen-functions so good luck to whoever is going to find a finite-time blowup.
February 4, 2026 at 7:21 AM
Anyway, I've never really understood real analysis & all the symbol juggling w/ operators & eigen-functions so good luck to whoever is going to find a finite-time blowup.
According to Gemini, someone can claim one of the Millenium prizes if they can solve the following equation. I got to this after a bunch of reductions for expressing Navier-Stokes in the Fourier domain.
![The Leray Equation (Spectral Form):
−U^−21(ξ⋅∇ξ)U^+∣ξ∣2U^+iP(ξ)[U^∗U^]=0](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:lpg5f7xuyxs6k5tcherzopig/bafkreif7q3bfsjsngsryew7pw7y7552kazsasbc6noiymispqk3aibfsmu@jpeg)
February 4, 2026 at 6:48 AM
According to Gemini, someone can claim one of the Millenium prizes if they can solve the following equation. I got to this after a bunch of reductions for expressing Navier-Stokes in the Fourier domain.
I'm starting to suspect the ultra-finitists are more correct than the alternatives.
February 1, 2026 at 3:39 AM
I'm starting to suspect the ultra-finitists are more correct than the alternatives.
Every rando these days has their own variation of a type system but very few actually provide a proof of consistency.
January 29, 2026 at 8:56 PM
Every rando these days has their own variation of a type system but very few actually provide a proof of consistency.
If you have a bunch of stuff (A, B, C, D, ...) & you know how to combine them (A⊗B) then to prove something for A⊗B⊗C⊗... you really only need to prove it is true for A⊗B assuming it is already true for A & B. Rest follows by induction.
January 28, 2026 at 8:03 PM
If you have a bunch of stuff (A, B, C, D, ...) & you know how to combine them (A⊗B) then to prove something for A⊗B⊗C⊗... you really only need to prove it is true for A⊗B assuming it is already true for A & B. Rest follows by induction.
The reals require classical logic for their properties. The Cahier topos uses the reals but its logic is intuitionistic b/c of infinitesimal objects that would collapse if the logic was classical. This means that context extension can invalidate some rules of logic that were previously valid.
January 27, 2026 at 6:42 AM
The reals require classical logic for their properties. The Cahier topos uses the reals but its logic is intuitionistic b/c of infinitesimal objects that would collapse if the logic was classical. This means that context extension can invalidate some rules of logic that were previously valid.
Starting to suspect it's all just kernel machines. I mean, at the end of the day the computer can only shuffle bits so it can't be anything else other than a kernel machine that figures out which bit pattern is "closest" to whatever is encoded in the machine.
January 24, 2026 at 1:54 AM
Starting to suspect it's all just kernel machines. I mean, at the end of the day the computer can only shuffle bits so it can't be anything else other than a kernel machine that figures out which bit pattern is "closest" to whatever is encoded in the machine.
Only a finite amount of stuff can happen in a finite amount of space and/or time. This seems like a more fundamental principle than the speed of light being constant. Finite amount of activity means finite speed for everything.
January 22, 2026 at 7:31 PM
Only a finite amount of stuff can happen in a finite amount of space and/or time. This seems like a more fundamental principle than the speed of light being constant. Finite amount of activity means finite speed for everything.
I don't think the manifold hypothesis in ML is valid or at least it's not clear to me why stochastic gradient descent w/ all the smoothing & clamping used in ML frameworks is a valid way to do function approximation.
January 22, 2026 at 5:50 AM
I don't think the manifold hypothesis in ML is valid or at least it's not clear to me why stochastic gradient descent w/ all the smoothing & clamping used in ML frameworks is a valid way to do function approximation.
Pondering what it means to implement a function. It's not as simple as I thought it would be.
January 21, 2026 at 8:40 PM
Pondering what it means to implement a function. It's not as simple as I thought it would be.
All mathematical formalizations presuppose pre-formal intuitions about discrete objects.
January 20, 2026 at 11:57 PM
All mathematical formalizations presuppose pre-formal intuitions about discrete objects.
The integers are unavoidable. Every book on type theory, when introducing universes, presupposes the existence of integers
January 20, 2026 at 11:55 PM
The integers are unavoidable. Every book on type theory, when introducing universes, presupposes the existence of integers
Writing code sometimes feels like constructing a simplicial complex. I'm not sure about dimensionality but there is non-trivial topological structure that is somewhere in the background.
January 16, 2026 at 6:13 AM
Writing code sometimes feels like constructing a simplicial complex. I'm not sure about dimensionality but there is non-trivial topological structure that is somewhere in the background.
I spent two days trying to build a self-playing chess engine w/ a transformer network for the value function of each state. It kinda works but it's also not obvious whether the value it's approximating has any connection to reality or not.
January 14, 2026 at 3:05 AM
I spent two days trying to build a self-playing chess engine w/ a transformer network for the value function of each state. It kinda works but it's also not obvious whether the value it's approximating has any connection to reality or not.
Identity of indiscernibles is not as clear cut as people think, e.g. how can one tell the difference between two electrons in a universe w/ just 2 electrons?
January 9, 2026 at 12:14 AM
Identity of indiscernibles is not as clear cut as people think, e.g. how can one tell the difference between two electrons in a universe w/ just 2 electrons?
Defining negation as a function from some type to 0 has never made any sense to me. The reason is that 0 has no elements/proofs/witnesses so there is no way to map to it to begin with.
January 8, 2026 at 7:11 PM
Defining negation as a function from some type to 0 has never made any sense to me. The reason is that 0 has no elements/proofs/witnesses so there is no way to map to it to begin with.
Set theory uses first order logic & axiom schemas so it technically has infinitely many axioms for each instantiation of parameters for the schemas. It also allows quantification over arbitrary sets (including characteristic functions encoded as subsets of products) so it's also "higher order".
January 5, 2026 at 10:29 PM
Set theory uses first order logic & axiom schemas so it technically has infinitely many axioms for each instantiation of parameters for the schemas. It also allows quantification over arbitrary sets (including characteristic functions encoded as subsets of products) so it's also "higher order".