Yosuke Neuroscience
@yosuke_neuro
Followers
331
Following
1K
Media
53
Statuses
269
Found a mistake on Chapter 47 Maclaurin series of Schaum's Calculus supplementary questions 12b n terms are supposed to be k for this solution to be coherent. I love you authors because of the free practise books you provide but with the amount of self doubt I already had
0
0
0
@leothecurious deserves a trillion followers honestly
the person on this website who has disagreed with me in the most interesting ways & has forced me to think harder about my research intuitions is probably @leothecurious
1
1
35
It's crazy impressive how short term plasticity implemented by @timos_m yielded significant energy efficiency improvements without any specific instruction. I know it's meant to be this way in theory but seeing it in practise is beautiful
0
1
4
This is complete bullshit people peddle around to get attention and inflated private market valuations. Let me demonstrate the mice and one small little concept called one shot intelligence. One shot intelligence otherwise known as behavioral timing dependent plasticity is
Move over AI-2027! Here's my forecast of the AGI timeline. Based on the model size of the human brain, I predict that human-level AI is 15 YEARS away. My reasoning in 🧵below👇
1
0
0
I absolutely hate Math textbooks. 1. Ridiculous notations with zero explanations 2. Zero intuition involved 3. Practising for the sake of practising. Will be demonstrating how I combatted this with @GeminiApp to understand math through intuition rather than bs practise
0
0
0
Has anyone considered that a linear plane in the manifold theory could be missing the point? Has anyone thought in penile structures? What if the idea has been staring at us all along?
0
0
0
Noted. Table from https://t.co/2mLv0jzWsy updated accordingly.
@leothecurious @timos_m Backprop is gay
2
1
2
Yeah i hyped this one a bit too much but its still impressive
0
0
0
Literally reading a paper thats about to change the trajectory of mankind And it has 4 citations lmao
1
0
0
Do you want to be a moonshot or do you want to be another incremental improvement to society
0
0
0
insightful thread that i think applies to all parties involved in the current LLM "rat race" scene. gdm was until recently an outlier that maintained wider exploratory branches but at some point also pivoted most resources to making gemini a leading "rat" for better or worse.
1/n Monopolies create platypuses. I know how that sounds. But there's something here worth examining, something that cuts across evolutionary biology and market dynamics in ways most people miss. Let me walk you through it.
2
1
7
Man im not at the level to understand this paper yet, but i will be back https://t.co/jb33cHlC9v
0
0
0
Best read for me in a while. Isnt that related to neuro but it touches on something that constantly piques my curiosity. Why mental illnesses persist even after millions of years of evolution https://t.co/ol5LdVGgB3
0
0
0
gemini-3-pro just oneshotted a tricky coding problem that i've been throwing sonnet-4.5 and gpt-5-codex at for hours today without much to show for it. i'm impressed. i think i'll probably retire within a few months. there's no wall. we're still riding the exponential. sam and
10
6
193
cool thing about being 99th percentile in trait openness is the natural aversion to the slippery slope of tunnel-vision expertise. i've probably visited a thousand rabbit holes and a thousand "this is the coolest shit ever" and a thousand "everyone should learn about X" or "this
this also applies to physicists: EBMs, diffusion, world models theoretical cs: kolmogorov complexity, compression low level cs: pufferlib, low sample efficiency RL neuroscientists: that one sakuna project, liquid neurons I'm calling you all out, go outside🫵
2
1
26