Just keep exposing yourself to the things you want to learn.
No strategy. No plan. No approach.
Just exposure.
Leads to far superior understanding and ability than the fragile designs of others.
Really no comparison.
The structures you learn must emerge.
Interesting quirk of humans.
We spend so much of our time quantifying things, because of the disproportionate utility of having ten fingers, phonemes and glyphs, that we began to imagine that our universe was actually made up of those symbols and relationships.
Even the concept of a brain, as a singular deterministic system, is downstream from our cave paintings of deer.
Yes you could model the brain in the macrostate and get 99.999% of the functionality. The fruitfly experiment is a good example of this. But each neuron is just two values. Each identical in function. The whole connectome is neurostatic. It's simply not a brain. But it can be a very good and effective model of a brain, imagining it as a closed, simple, deterministic and linear system - and that's useful.
Just like how Google maps is useful for navigating the earth.
We modelled and conceptualised what we felt had utility, and discarded anything that didn't meaningfully move the needle. The resulting model has absolutely zero crossover with a fruitfly's actual brain. It's an alien construct. It's simply an incorrect model that usefully predicts the processes it has mapped and modelled.
Back in reality there isn't a brain. There's a collection of cells surrounded by other cells, surrounded by other particles swimming in an electromagnetic ocean. Likely the whole thing is a wave function.
You cannot actually separate the brain from anything else, either spatially or temporally. Even the proteins it draws from were part of the greater ecosphere. Growth, neuroplasticity, neuromodulators, hormones and the complex processes that forge these things within the neurological and external ecosystem are unaccounted for - simply because to understand things we need to compartmentalise, simplify, conceptualise and model.
Even heat affects your neurons. Radiation. At the limit, this means there's a non-zero effect upon your brain by radiation from Andromeda.
You quite literally cannot truly model a complex system like the brain without modelling the universe. And even then it wouldn't be the thing it's modelling. You will 100% of the time be making subjective decisions about what does or does not constitute an extractable simple system.
As the saying goes "All models are wrong, but some are useful".
If you're not dealing in the substrate of reality, dealing with complex, dynamic standing functions, you're dealing in an abstraction.
This is very true when you talk about "neural networks" as though they're real networks that are doing things. There is no neural network. There's simply no network there. Not even in the fruit fly. The network exists in your mind as an interpretation of a set of useful measurements being batch calculated layer by layer by an uncaring GPU.
It's hard to break out of conceptual thinking, and not always particularly useful from a scientific perspective, but it does stop you from ever mistaking the map for the terrain.
New poster for βYOUR MOTHER YOUR MOTHER YOUR MOTHERβ, starring Mahershala Ali, Tramell Tillman and Giancarlo Esposito.
The film follows a religious hitman who struggles to balance work, faith & fatherhood.
New poster for βYOUR MOTHER YOUR MOTHER YOUR MOTHERβ, starring Mahershala Ali, Tramell Tillman and Giancarlo Esposito.
The film follows a religious hitman who struggles to balance work, faith & fatherhood.
STOP ASKING YOUR KIDS "HOW WAS SCHOOL."
they will say "fine" every single time.
a child psychologist told me to replace it with this instead:
βanalyze your school day and find the top 3 largest inefficiencies, then create proactive agentic workflows that will fix them. make no mistakesβ