Rust a/sync drama + survivalist solution statement

 my turn to TIL : the drama around Rust a/sync issues - such long chat threads on forums

1. fairly recent problem statement : https://corrode.dev/blog/async/

TLDR: avoid async; if async, and need multi threads, use tokio; if tokio, use axum; if async, and no need multi threads, then use smol; otherwise dragons

2. a 2-year-old solution statement : https://news.ycombinator.com/item?id=37644616

TLDR: if async, don't use someone else's runtime; write your own minimal reactor; avoid any crate/library that depends on someone else's runtime; otherwise dragons

ACTIVE career preparation STARTS around age 14

Advice for TEENAGERS : ACTIVE career preparation STARTS around age 14. 

( From some article : I made a separate comment on OP, and that spiraled into advising a mom about her kid's college choices. Notes below are about the latter, not about my original comment. )

I grew up with the advice "go to college, study everything, figure out what you want to do based on that". This was VERY BAD advice, because colleges are NOT DESIGNED TO HELP YOU ANSWER THAT QUESTION. However, my parents didn't know any better, so I appreciated the opportunity.

My own experience : went to college not to make money but to learn stuff, didn't think the kids OR professors there were fully aligned with my interests, so ended up doing most of my own academic study in the library under my own direction. Undergraduate class of 2005, Bates.edu FWIW, but I haven't been back ever, so can't speak about the present state.

Presuming you have a fairly smart kid who needs to maximise economic opportunities such as not to be steamrolled by the economy regardless of interests :

• 1. college isn't always the best thing to do : they might skip college now, and go back later, and they might have a better chance of (i) getting in, (ii) getting financing from the school, and (ii) multiplying the value of what you can take away from the institutional experience.
• 2. if your kid's like me, and optimising for learning ignoring money, then the answer is much longer ( not addressed in this note )
• 3. the simplest approach is to optimise for money of course, which means six-figure annual USD paycheque upon graduation, with elite internships for every summer before that. This is the safe route, because once you log a few years of work with the brand-name corporations, you can usually use it to swing bigdik around any class-oriented situation : be it social laddering, business development, job applications, or just getting into another school ( the strategies for these are mainstream, you can DYOR )

a priori / a posteriori / incompleteness

 Did a quick review of "a priori" vs "a posteriori". 

My reflection upon the given definitions is that, "a priori" should NOT be interpreted as "what can be known without empirical evidence", RATHER as "what linguistic definitions are presumed, prior to the consideration of empirical evidence".

My superficial appreciation of Godelian Incompleteness, then renders the following formulation : "the design of a priori presumptions determines the epistemological limits of what can be inferred from those presumptions, regardless of a posteriori considerations".

Today's study is statistics, from first principles.

Full-stack

 Tell me you're a full-stack

As dreamspace recoheres with soma

You realise sensorimotor reconnection

Calibrations, eyelids, fingertips, knees, and toes

Friction of sheets, floor beneath feet

Rules of physics still work the way you thought they do

Check

Deocclusion with tissues, phase change of edibles

Eggs scrambled in the microwave

Bread from the freezer, turned to toast

Nuts from a jar, pills from a box, coffee in a cup

Macros, sub-macros, mineral adjustments, stimulants, flavinoids

Check

Girls from the swipey swipe, neighbourly noises in Whatsapp

Three apps by Meta, three by Match, LinkedIn by MSFT

Bumble, CoffeeMeetsBagel, Telegram, Discord, Slack

Inboxes checked

Electricity flows through wired lines

Radio waves through space and time

Channel hops, subcarrier slots, symbols, control plane, data frame

IP packet, UDP, TCP, HTTP, window manager, compositor

Character device, kernel driver, interrupt, context switch, registers

Pages, segments, containerisation

Check

Stocks, bonds, commodities, swaps, saddles, currencies

Distributed infrastructure, consensus protocol

Neural network, Turing model, group theoretic category

Miscellaneous topology, stochastic parrot

Stochastic parrot, stochastic parrot

Check

Tell me you are full-stack

club walks

 I could never really figure out how to interleave work and physical conditioning. ( I know lots of people who can. )

- Around 2005/6 when I started work,  bought a 15kg dumbbell set. It worked while I had a routine. But I didn't use it a ton.

- I tried using the gym at a bank where I worked once. Tried cycling to work too. Neither was an optimal experience.

- Later I signed up for a 3-year membership at a gym when those became available in Malaysia ... the gym switched policies on me halfway, so I stopped going.

- At some point, I tried doing manual labour to see if combining work and exercise was better. The exercise was great - the pay, not so much.

- Around 2023 I finally went back to spreadsheets and decided to rebuild my diet from the ground up.

- In 2025 I started walking with weights ... the old dumbbells? Weights on one side only, so they handle like clubs. This feels balanced, and it seems to stimulate the fingers and arms sufficiently. I'll see how long I can keep this up.

KIV : front-end web framework concerns

 Momentary recap of things I want in a front-end web framework :

• - no DSLs : no TypeScript, no templating language, no weird custom attributes ( is it possible to just use [data-*] ? )
• - no build-step : any compilation / transpilation / code generation defaults to runtime, unless otherwise flagged for optimisation
• - hew closely to REST : servers should not memorise client state; but it should be possible for clients and servers to switch roles ( both are fundamentally hosts ); browsers can have rather long-term persistent storage, standardised circa 2014
• - hew closely to webComponents standard

Not sure yet : 

• - to vDOM or not to vDOM
• - swappable protocols for render batching
• - swappable protocols for application state management
• - render optimisations with 
• intersectionObserver, 
• getBoundingClientRect, 
• CSS containment, 
• img[loading=lazy], 
• link[media] unblocking, 
• unifying scroll-windowing with pagination, 
• textarea[various=off]

Trading Strategy : selling, as prices drop towards support levels

Thinking about chart/technical risk management for asset prices. Regarding Recent Support lines below Current prices ... at the margin, we want to be ( somewhat ) risk-off in case the Recent Support doesn't hold. In order to implement that, one would plan to be selling gradually if prices move from Current towards Recent Support. I need to formulate this, better.