orang kebencian

Jenis orang yang paling saya benci, bukan tertumpu kepada mana-mana negara, atau kaum, walaupun ada banyak jenis ini persekelilingan hidup saya. 

Iaitu sesiapa yang dahpun dorang diajar sesuatu oleh ibubapa atau gurunya, dorang ingat pula orang lain kena ikut cara itu. Bila tak ikut pula orang lain, jenis orang yang saya paling benci akan rasa gelabah. Dorang seakan-akan petu ... yang mat saleh panggil karen. 

"Kan, saya sudah buat apa yang disuruh, kenapa orang lain tidak menjunjung perbuatanku? Mana pengiktirafan? Aku mahu respek! Cara saya mulia! Saya seorang yang berjalan lurus! KORANG, tak beradab!" 

Karen puteh ada, karen cina ada, mana-mana jenis karen pun ada. Tapi ini kerangka mengenali penyakit itu, secara am.

Saya tidak sebut pasal kemahuan seseorang. Tapi pasal orang yang kemahuannya, diterima dari ajaran, dan bukan dikerjakan sendiri.

electrolytic muscle function, and disease

At rest : 

• [ neuron and muscle independently : sodium-potassium pump (slow) : pushes out 3Na+ for every 2K+ pulled in, maintaining negative voltage inside the cell ]

Firing : 

• [ neuron : ( due to upstream stimulation) opens sodium gate, Na+ inflow sends current to calcium gate, Ca2+ enters neuron, triggers acetylcholine release into synaptic cleft ] 
• -> [ muscle : acetylcholine from synaptic cleft, opens sodium gate, Na+ inflow sends current to sarcoplasmic reticulum, which releases calcium, which causes contraction of the cell : (fast) K+ leaves cell to restore negative voltage inside the cell ]

Muscle Diseases, and ordinary states which become diseases if not reversed within ordinary timeframes :

• - Ca2+ : 
• too much outside cells, "overshields" sodium gates of BOTH neurons and muscles, both become "understimulated" ... and subsequent to neuron stimulation, Ca2+ enters neuron slower, dumping acetylcholine slower, furthering muscle "understimulation"; 
• too little outside cells, "undershields" sodium gates of BOTH neurons and muscles, both become "overstimulated" ... and subsequent to neuron stimulation, Ca2+ enters neuron faster, dumping acetylcholine faster, furthering muscle "overstimulation"
• - Mg2+ :
• too much outside cells, block Ca2+ entry to neurons, reducing acetylcholine release, "understimulating" muscles
• too little outside cells, Ca2+ entry is less blocked, increasing acetylcholine release, "overstimulating" muscles
• - K+ : 
• too much outside cell, K+ cannot leave quickly, cell may become "overcontracted" - ditto "overexcitation" of neuron at early stages, but  at later stages, sodium gates fail, neurons and muscles both weaken; 
• too little outside cell, K+ leaves too fast, cell may become "hard to contract"
• - Na+ : 
• too much outside cell, Na+ enters quickly, cell may become "overexcited", ditto for neuron; 
• too little outside cell, Na+ cannot enter quickly, "underexciting" cell, ditto for neuron; nervous system may overcompensate with erratic voltages, causing "erratic excitation"

The Cartesian Product Bubble

I'm guilty of avoiding AI over the past decades because the techniques I've heard about still seem rather primitive. The last decade in particular has been characterised by VCs throwing obscene amounts of money at a fundamentally inefficient approach.

After noodling around in the current tech for a week or so, just to make sure I understand what's going on, I think it's safe to say that the the titular concern of this post stands firm.

An analogy. At the heart of trendy LLMs is a giant Y times Y list of known words in English ( also other languages, but nevermind those for now ), forming an enormous 2D table. It's Y times Y because both H and V axes have the same list of words - sure, half the table is redundant, so you can think of it as a trangular half-rectangle of unique pairs. If you lookup the junction between any two words, you find a WINDOW to a realm of many, about 10^(3 +/- 1 ), dimensions of information about each pair of words.

Based on the "best" available public information, this is the data structure for storing everything the LLM knows about the world - which it knows ONLY FROM READING TEXT ( except for MLLMs, which we'll also ignore for now ). Even in the case of MLLMs, sensory spatio-temporal data can be understood to be stored within the realm accessed by each WINDOW.

Of course, when you ask an LLM a question like "What is the meaning of life?" it doesn't need to peek through ALL the WINDOWS and all the realms above, rather it only has to peek through a subset of windows, W_n. However it's still fundamentally inefficient, because : it doesn't JUST look at W_n and answer your question. 

Oh no.

It looks at your question, goes to a set of windows, W_n1, takes a peek, grimaces in deep thought, spits out ONE word, then pats itself on the head and goes to W_n2, a COMPLETELY DIFFERENT set of windows, spits out a SECOND word, and goes on and on until W_nM to obtain M words. While there is a cache, this is exactly what it sounds like - quite a bit of work.

I'm not a very smart fellow, so I'm probably going to be wrong about this. But, it would seem that the thin red circle that I am drawing around the entire architecture that drives most of our current AI tech is going to have to collapse, and be replaced by more efficient methods soon enough.

When? Who knows. I've been waiting since 2003, and "they" haven't figured it out yet. Some of "them" are pretty close, I think. I like the JEPA and MLLM approaches, and am eager to see where they end up.

One thing's for sure - if you create a model of the world based on this Cartesian product approach to windows upon a realm of thousands of dimensions, and have to crawl through the whole library, to peek through a different SET of windows, once per answer word ... even if it gets you the correct answer to a question, it's bloody tiring.

And to BUILD that library of windows to that realm ... it still takes WAY more money than it takes to build a human brain.