thanatropism

I grew up fascinated by these essays from maybe age 8 onward.

Now I realize the pervasive bad faith in them, e.g. dismissing Murray (The Bell Curve) and enthronizing the in-retrospect-irrelevant Jay Gould. I knew his novels were kind of silly, but he was supposed to be a Scientist educating us in the scientific view of the world, goddamn it.

Is it supposed to be flattering?

> new pandemic is killing millions

> yes, but viruses existed long before it

Maybe I wasn't receiving the full Fitness Message as someone proudly sedentary, but the way I experienced the world (seen through what I think I know now), it always seemed like people overemphasize weight loss and being thin as a marker of health.

People, people, I know many of you don't want to look like you lift, but skeletal muscle has moby advantages in minimizing the annoyances of old age. It also literally helps in weight loss. I'm also told it regulates insulin. Plus weight training also strengthens your tendons and bones. If you start deadlifting heavy at age 40, you will develop strong spinal erectors which will likely protect you from herniated discs when you're 60.

Losing weight is fine and will help preventing coronary / fatty liver etc diseases. But please make it priority number 2. Build muscle and try, as a bonus, to lower your body fat percentage. If you never do, it's better to be 35% bodyfat and yoked than to be 30% and made of blubber and chalk.

Ctrl-F doesn't seem to show anyone remembering the Ballad of Scott Alexander.

There's no reasonable narrative in which OpenAI are not villains, but NYT is notoriously one to shoot a man in Reno just to see him die.

AlphaGo passes Hubert Dreyfus's test -- it has a world -- in a way that LLMs don't.

"Trash Truck" and Pocoyo.

"1;

Ehh. Sometimes you don't want to have to have string builders to assemble REST calls.

You can always restrict the meaning of "causality".

E.g. Granger casusality means that A is typically detected before B and not the other way around (so not mere correlation). It's a moby useful concept.

This ain't it, chief.

> Jazz is music performed by an ensemble of players, each (very) specialized and proficient in their respective instrument.

So wait, a duet with Cecil Taylor and Eric Dolphy, the latter of which also plays the flute and many more wind instruments is...

NestedText + Pydantic.

Tihs probably refers to comon mispelllings an typo's.

Is any of those easily modifiable to return the arg-median (the index which has the median).

YouTube may have peaked as the outstanding source of impressively high quality content on a very wide range of topics about 3-5 weeks ago; possibly with Horses' guide on surviving in a desert island. It's still so, so good.

---

I remember a film of Salvador Dali talking to a cheering crowd and saying in full dictator voice "... in the future... toilet bowls will be made of gold..." (roaring crowd). That's what's happened to the internet at large.

What are the advantage of Cython over something like C++ with pybind11 or whatever the equivalent in Rustland?

I've been using ChatGPT to develop the concept of a novel or TV show in which Al-Andalus never fell to the Reconquista, but rather conquered chunks of Central Europe... and fast forward we're traveling the stars. It's pretty good at following the "story A in the present / story B in historical flashback / stories converge in major themes" pattern.

This is unintuitive, but modern fortran is a shorter jump from Matlab than Python.

The greatness and wretchedness of Python is "pythonicity". It's very hard to miss the mark with Python (all things considered: life is finite, most people are more like analysts than developers, analysts deliver generally more value per hour than developers, Rust is hard to learn) but you basically need to submit to the "Pythonic" brainwashing -- which includes "anything performance-critical should be written in $fast_lang and then used from Python; but almost nothing is perf critical, and most of what's conceivably needed already exists".

I wish the Python<->Rust interop story was a little better. I learned to write some C++ for an embedded thing (smart flashlight, story for another time) and immediately started writing the Python extensions I struggled to write with Rust.

(The average data science/ML-ish person encounters/figures out custom algorithms maybe three or four times a year, and three of these are fast enough with vectorization contortions. I've had two cases that were recursive and 25X faster in $fast_lang than I could possibly make in Python.)

I wish there were more serious studies on "Matlab risk". There's a lot on $old_lang risk already, and Excel risk is now a serious thread of research, but many in academia (and very possibly greybeards in industrial R&D) just won't budge from Matlab. So e.g. all the codes from my thesis are in Matlab because my advisor wouldn't Python.

Matlab in theory has OO, but it's very slow and who bothers to verify that? So practically everything is matrices and there's very little "semantic expressed in code itself" (to paint issues with a broad brush). Also matrix calculations can get botched numerically, particularly matrix exponentials, but the whole culture (that I can't even expunge from my own thinking) is that I have a theorem so tests schmests.

A better idea yet is to spend the time with Ralph Waldo Emerson.

My biggest issue with i18n is not the grammar. It's the sheer uncanny valley of it.

There are translations to "Hello" in Portuguese but I'd cringe at being greeted with them by a webmail client instead of the more formal Good Morning/Afternoon.

The formal/informal gradient is very culture-bound and even hard to pin to a scalar space of possibilities. In a work environment people will fluently code-switch too -- say, between ranks or in the middle of a tiresome meeting when everyone takes five minutes to kick back and comment on lighter matters. It's hard to situate a computer in this social context.

I used to be all in on the Blub Paradox discourse -- but I've been learning some C++ for the Arduino and the differences in power with, say, Python, are mainly related to how easy it is to think-as-you-code in a dynamic language.

Of course, my life story is such that I've used Python professionally to the extent that I've started to forget all other languages -- so I naturally find the Blub Paradox Discourse favorable. But all the footguns that come with passing around pointers to memory addresses also come with a whole different way of expressing problems; whole different range of thoughts that are thinkable.

(I valiantly await the Rust brigade in my replies...)

My manager called me for short chats on world events (semi-relevant to our larger work, not for current-assignment). Now I realize that was 20% to get my take on whatever (say, 2008 auto companies bailout) and 80% just to look me in the eyes and hear me talk and feel that I'm really here and committed and not silent quitting.

What. There are no proofs of convergence for many of the most popular NN optimization algos. IIRC Adam is known not to converge in some cases.

The bitter lesson is that Messy AI is better able to cope with Messy World Problems than Neat AI (by light-years at this point), not that it can hack Neat Problems.

I remember the hype around XGBoost and Kaggle contests asking to solve problems in prime number theory.

It's a very short article. It's not like it's full of substance and that one sentence is a faux-pas or overenthusiastic. The content is basically that one idea plus some padding.

I lived with my parents like three years into working a full time, decently paid job. We'd always lived in rentals and I wanted to help with the bills. After moving out I realized that, man, I was saving so much money!

This anecdote to say that I'd probably been terrible at a remote job while in the old structures. Now I'm married with a child, but I get to call the shots on daily routine, so to speak. I can even work late nights without mom coming to admonish.

If you can find intelligence about which CRE will crater to zero and which will just be half of what it's now, you can arrange some in-kind swaps. But the ethics of this feels wrong.

Then, isn't the long duree view of technology about replacing apprenticeship learning (sit with the master welder, clean the shop, watch him work -- this doesn't cut in software)? Sometimes with explicit learning (linear algebra), sometimes just with low-friction entry points into professions.