1,116 karma · joined April 22, 2020
And while 2019 might have been third-highest year for investment in 2020, according to this it's been surpassed in 2021, 2022, and 2024
https://kpmg.com/xx/en/media/press-releases/2025/01/2024-glo...
So why have graduate hires continued to decline since 2023? It seems funds have been diverted from junior hiring into AI investments.
However, as others have remarked, this might be a case of "AI is not taking your jobs, AI investment is taking your jobs"
Junior hiring might pick up again once the spending spree is over
He also has this series of talks beginning with the question "What is probability, what is randomness?"
https://www.youtube.com/watch?v=aJAQVletzdY&list=PLx5f8IelFR...
https://www.signalfire.com/blog/signalfire-state-of-talent-r...
https://en.wikipedia.org/wiki/Casualties_of_the_Iraq_War#Tab...
Still, "over 100,000" is technically correct if it's more than 100,000. Since this subject isn't the main point of the article and the Iraq war is generally acknowledged to have been disastrous, I suppose he chose a safe figure so as not to derail the article with disputed estimates at the outset.
I think it was this one, unfortunately archived now. I don't know the new one
https://ocw.mit.edu/courses/6-00-introduction-to-computer-sc...
But mathematicians can talk to each other about arbitrarily abstract concepts, as long as they have enough shared background, and they don't (always) need a blackboard to do it.
Conversely, you can have conceptually very simple things that are basically impossible to follow just by listening, like multiplying two nine-digit numbers or following one of Euclid's proofs in plane geometry. The difficulty isn't about abstraction, but how many things you have to hold in working memory
Richard Behiel, Physics Explained, Dr Jorge Diaz
For mathematics, I really appreciate the simpl but effective format of Richard Borcherds (Fields medalist)
And for videos specifically, it can obviously help understanding in many cases to have animated visualizations
The Meta paper, as well as other studies which have been interpreted as rudimentary "mind-reading" have measured activity in sensory cortex correlated with direct sensory inputs. There's a fairly close mapping between the initial layers of sensory cortex and patterns of activation in the sense organs. e.g. the optic nerve from each eye projects onto the initial layer of visual cortex in a way that closely preserves the geometry of the retinal image, so it's not that difficult to correlate information in the stimulus and these parts of cortex. Making sense of activity in deeper areas of cortex which isn't as closely correlated with immediate sensory stimulation is a much harder task.
Secondly, the idea seems to be that the brain could make use of a "lossy image" of its own overall functioning. This part seemed very handwavy to me. The brain already contains the information about its own functioning, by definition, so it's not clear to me what functions would be served by the brain's being able to "sense its own magnetic field". It's known that the brain integrates information from distant regions through the patterns of neuronal connectivity. It's not clear that something similar can be done with magnetic fields, because these would mostly affect very nearby areas of the brain, and long-distance effects would be scrambled with all the other activity going on in other parts of the brain.
The idea to look at the effects of the electromagnetic fields in brain functioning is interesting though. The general idea has been around for a long time[0]. The dificult part is making a detailed proposal for how it would actually work and finding experimental evidence for that.
[0] Burr, Northrop (1935) The Electro-Dynamic Theory of Life
https://leanprover.zulipchat.com/#narrow/channel/284757-job-...
https://en.wikipedia.org/wiki/The_Hitchhiker%27s_Guide_to_th...
About tests of quantum gravity, there have been proposals for feasible tests using gravitationally-induced entanglement protocols:
It's an interesting hint at the deeper evolutionary origins of language in the ability to plan complex actions, providing a neural basis for the observation that language and action planning have this common structure of an overall goal that can be decomposed into a structure of subgoals, which we see formalized in computer programs too.
This is an older reference (1991) where I first heard about it. there are more recent studies reinforcing various aspects of it but I didn't find one that was as comprehensive
What about provoked attacks? And who's going out there provoking sharks?
I also like the implication that the provoked attacks are understandable and don't count
"The shock produced a stupor similar to that which seems to be felt by a mouse after the first shake of the cat. It caused a sort of dreaminess, in which there was no sense of pain nor feeling of terror, though quite conscious of all that was happening."
Even 1824 by Fourier
https://en.wikipedia.org/wiki/Greenhouse_effect#History_of_d...
Physical objects aren't like that because you can discover that they exist by empirical investigation.
In mathematics the discoveries are about the logical implications of sets of axioms. Some of those axioms contain assertions of existence, like a number 0 in Peano arithmetic or the empty set in set theory, and then you can prove statements about these objects based on the axioms. It's circular to infer from these conclusions that the axioms are true.
What's interesting is why certain axiom systems are so useful and fruitful. Personally I think it's because they evolved that way from our investigations of the physical world, but that's another matter
It's similar for the case of programs or algorithms. We can say that a sorting algorithm exists, or a chess-playing program or whatever, which means we know how to implement the logical process in some physical system, but it doesn't mean that they have some kind of existence which is independent of the physical systems. It's just a way of talking about patterns that can be common to many physical systems