pnathan

What would you recommend as the canonical work to read for the first thing you mention - fuzzy set theory & fuzzy measure theory? I've run into fuzzy set theory before, but not the more sophisticated elements you mention.

oh yes.

there's enormous pressure to be normal. being weird is... like... someone's going to be uploading video of you and mocking you.

there's also the thing where its more expensive to have a normal life, financial pressures say "do financial things". like... the us world is shaping to be "high road" qua Stewart Brand whereas hackers thrive in "low road" environments.

there's also sort of this ... how do I say it ... hackers & that legendary online culture was sort of a product of a people without obligations. seeing your kids soccer game is much more important than debugging a program for fun. being monks for code & computers is not culturally sustainable.

Disagree. They are serving similar structural roles(although I highly doubt OpenAI is a scam). While it could be expanded in depth, it points the similarity out...

Higher dimensional algorithms get very messy - see X-trees vs R-trees.

E.g., KNN is very fast, if and only if you have a high performing R-tree query.

so while the 2d algos might "fall out" of the sophisticated cases, the sophisticated cases will need to go in a different optimization direction.

This confirms the widespread anecdotes online, where so and so's friend smoked way too much and had a mental break(more rarely, the sufferer speaks for themselves).

I'm glad to see research on this, but I hope it doesn't spawn another wave of reflexive & unthinking anti-drug legislation.

I've run Loki in production for over a year... as far as I know, Loki is not designed to be done all-in-one, that's kind of a dev/local design.

N.b., I've run this in both AWS/GCP in a k8s scenario, against S3/GCS respectively as a long-term store.

I've also run Mimir, and they both work _fantastic_ in this deployment scenario (as described).

The corner of the weightlifting world I keep an eye on has centralized around protein, veggies, and few carbs.

The other angle to look at is satiety; good food should also have a satiety so you don't go back and munch. :)

I'm experimenting with using Bard to generate Prolog code. I have this idea that if I can synthesize Prolog and a LLM, I can do some very clever planning work.

If I believe I have a lot of knowledge to transfer, I'll do a 2 week + X notice. I've done 1m notice, and that was excessive I believe. Otherwise, 2 weeks is standard in the US, and I have no reason to deviate. Nor would I conventionally just walk out that-day, as that leads to a bad odor.

If I was waiting for a bonus or cliff, of course I'd wait until after the date before formally giving notice. Because, of course, that would be a problem if the Management decided to terminate me immediately.

It's quite contextual, of course, but I prefer to follow local standards.

> Fritz Lang's Metropolis

(produced 1927)

One major aspect I want to touch on that bother me in the perspective you criticize.

Poverty and crime are not married.

My wife and I both grew up poor and working class. Living in trailers are in our life stories. You can be dog poor and still not be a junkie, still not mug others, still not assault people. This was understood widely in our upbringing and those we dealt with. People _did_ misbehave, but it was not "just what happens".

Presumption that poor areas must mean getting to deal with junkies, means dealing with violence, well, that is a morally bankrupt view. People don't have to do that. That is their choice. Improvement is possible.

Quite correct. Made an error. Carryover from a prior job where we mucked with weather data... I was thinking more along the lines of raster geo datasets like sat images, etc. Each pixel represents one geo location, with a carrying load of metadata, then a timeseries of that snapshot & md, so timeseries-raster-geomapped data basically.

I don't remember anymore what that general kind of datatype is called, sadly.

Figure every 15 minute someone looks through the door to make sure you didn't harm yourself. Mileage might vary by ward and illness.

Some wards are noisier than others.

I've heard of a camera in-room as well. Better for disturbance, worse for privacy.

Note that these people in acute inpatient wards are in a serious crisis: attendants _have_ to make sure they aren't hurting themselves.

I've had a very good experience with pg queuing. I didn't even know `skip locked` was a pg clause. That would have... made the experience even better!

I am afraid I've moved to a default three-way architecture:

- backend autoscaling stateless server

- postgres database for small data

- blobstore for large data

it's not that other systems are bad. its just that those 3 components get you off the ground flying, and if you're struggling to scale past that, you're already doing enormous volume or have some really interesting data patterns (geospatial or timeseries, perhaps).

> SF's favorite game is to send ambulances out to a guy sitting on the street, screaming insane things, covered in his own feces and trying to punch people. Ambulance shows, guy says "No!"

Roughly - without the violence - a friend watched this precise scenario last fall in Spokane, WA. If the person had been within hospital grounds, the paramedics & security could have restrained them under grounds of inability to make choice, given them medications, and maybe helped them some. But, since the person was a block away from the hospital, there was no ability to say, "buddy, you have feces all over you and are laying with flies all over. we're gonna take you in, hose you off, launder the clothes, then see what we can do for your mental state". That would have been the kindest thing to do.

The crux is that certain "out of your mind" states require external intervention, and there is no legal way to apply that in many jurisdictions today. Note that some jurisdictions require a trial-like apparatus for involuntary holds after some 24/36? hour period - I don't that's inappropriate at all. Being able to see a guy laying covered in filth and take him in to help should be legal; at least for 24 hours. The indigent will have their debt forgiven - hospitals know they don't get blood from a turnip.

It is true that some medications don't work for everyone, and some medications produce horrific side effects. So there is some % of people for whom the best medication regimes today do not work. That has to be considered and reflected through. It's also true that there's high correlation with addiction and homelessness. That all has to be considered and addressed.

Considering institutions. As I've had a good friend for a long time with mental illness who has been in wards, the quote from the article is very true.

> I disliked the hospital, but even with its heavy locked door, I knew it wasn’t a branch of the “carceral state” devised by a power‑mad society to torment him.

And this has to be understood by everyone who thinks through this. Mental wards in the United States today are not, conventionally, torture chambers for mad doctors. They are restrictive and often frustrating, for reasons that a little thought should reveal. Aberrant nurses and arrogant & high-handed docs exist there, but they are not, systemically, the norm. It is not a horror movie.

TBQH, I don't think the vast majority of mental health writings online by non-experts touch on severe mental illness in anything approximating a realistic way.

The status quo is really quite unacceptable. My druthers would be to re-institutionalize and keep a very high legal and regulatory scrutiny on the institutions. Other solutions might work well too. But a key cornerstone of improvement will have to come to changing the shape of the law for involuntary holds and medication for people in high crisis.

I've never really schemed, only Lisp'd in a common sort of way.

I've heard that some scheme things will compile out to C even.

- a helm 3 distro is 45M

- a terraform distro is 59M

- a kubectl distro is 44M

getting rust installed on a windows machine required not just a rust distro, but a visual c++ library set over 1G.

at a certain employer, the size of the teams golang tool was well over 100MB.

So - 50MB to start with a full Lisp image? yeah, sure, no worries.

My (older) code uses nickel, a sync web server that is, sadly, out of date.

I'll be migrating from that to one of:

* https://github.com/tomaka/rouille

* https://github.com/tiny-http/tiny-http

Eh, it was the original therapy chatbot. Focused on Rodgerian psychotherapy.

I'm not even sure ChatGPT existed two years ago. It seems like a very odd story. I wonder what the core facts are.

One thing I've been thinking about is getting a GCP certification to increase GCP recruiter reachout down the road, as I've really enjoyed working with GCP and would like to stick with it in future work. Guess I'll just do more keyword work in the LinkedIn. :)

re the edit:

> Nevertheless, if there were a world where amazing products suddenly snapped out of existence and were replaced with husks of their former selves, it would be this keyboard. Oh, right. That is this world.

Amazon review of the latest gen ergo keyboard.

If ms does a MS Natural 4001 or whatever, a genuine successor, I'll probably buy 3. It was _really_ nice.

The hardwood enclosure hasn't bothered me too much. The thumb cluster and layers bother me more.

I also miss the MS Natural 4K. There's a cheap knockoff on Amazon, but I didn't bother with it.

the 3 modern higher end keyboard companies that attend to ergonomic concerns are, afaik, - keyboard.io - kinesis - ergodox

and then there's maltron keyboards, a bit older-school.

I have a kinesis edge and a keyboard.io, I like the kinesis edge rgb variant. the keyboard.io is very hard for me.

> Instagram democratizes the mental illness that used to only affect child stars and the hollywood famous.

Astute. I have concluded there are social mental illnesses, and celebrity is absolutely a generator of one of those.

oh, _bother_. :-/

So this all justifies my general stance of "find a Unicode expert" when questions of unicode come up. But that's pretty wilfully blind (although practical).

I would like to understand why UTF-32 didn't catch on as The Standard Unicode for the modern world. it seems that - albeit memory wasteful - it would sidestep a lot of these issues.

the thinkpieces about how Millenials need to treat Gen Z are starting to flow already, ha. Retreads of what was written about millenials.

I remember.

I love the random conversations on the train. Really would love to see more investment in the train in the USA.

when you have work and personal email on your phone

and you have your own aws account

and the work finance department screws up

and you get an email from aws with a > 100K owed bit of information....

that was a moment of panic.

This looks cool!

I would, however, say that while I'm in the general target audience, I won't do crowdfunded hardware. If it isn't actually being produced, I won't buy it. The road between prototype and production is a long one for hardware.

(Still waiting for a very cool bit of hardware, 3+ years later - suspecting that project is just *dead*)