Reading this recent blog post on a and research project that answers code questions:

and I think I can better understand why "fast intro to machine learning" or "all the math you need for machine learning" articles are perennially popular.

Apparently lots of devs read such posts and think "wow that's so cool I'd be so cool if I could do that".

Meanwhile I'm like "99% odds it's rickety research piece of shit vaporware moving on".

I don't think this is because I'm huge curmudgeon, but rather a combination of

(1) several instances of me excitedly asking authors about their papers and being told "o ya it took a lot of parameter tuning to get those examples to work", and

(2) myself writing papers that required a lot of parameter tuning to make a few examples work with no clue how to make it work in the cases anyone really cared about/made money.

A tiny number of research results matter. Most are funders' vanity projects.

Show thread

For anyone who doesn't know: & is one of the more boring corners of enterprise computing where you try to predict things from large reams of data.

All three of those elements—predicting what, predicting how, and what data—are tedious, thankless tasks with uncertain payoff. To specialists (people into math/Hadoop/marketing) these are worthy tasks, but no more interesting to most people (or most devs) than any other niche in the enterprise like devops or pentesting.

Show thread

I saw a headline, "Data isn't the new gold, it's the new uranium". Nuclear power & weapons of course changed the world but the vast majority of people, of technical people, of physicists/engineers, were content to not get involved & let the specialists do their thing.

Same with the aero industry. Planes changed the world but it'd be a little ridiculous to imagine the orange site/code blogs devoting so much space enticing everyone to become an airline mechanic or a nuclear physicist.

Show thread

I live in a relatively boring corner of the tech scene: I make websites for people who have bigger problems than websites. I find it very interesting & satisfying, but it's obviously only one niche in the huge tech ecosystem.

I lived in the machine learning niche for a while—it was also interesting and fulfilling work but again, just another specialization (if hard to access without post-grad math).

The blitz of data science articles peeves me. I should probably let people have their fun fad…

Show thread

It’s definitely ok to do things because everyone else is doing it too, since even if it’s not that fun, gelling with everyone else who’s doing it makes it super-fun. Examples from the mid-2010s that come to mind—Pokémon Go and Taylor Swift.

And if you think this is where I say “but you have to find what makes you truly happy”—surprise!—people who eke out happiness following the herd are no less happy than the ruthlessly idiosyncratic robots who get satisfaction of loving what they love.

Show thread

Because finding something you love for itself (& not just because it's popular) isn't what makes people happy.

Nor is finding something everyone loves and loving it too.

Data science won't make you happy. Finding your true calling (JavaScript in my case) won't make you happy. Discovering Lizzo won't make you happy.

All these things just fill time in your days.

If anything, being happy (being at peace, ataraxia—this has many names) comes from knowing nothing in & of itself will make you happy

Show thread

@22 I agree of course in principle: happiness comes from within, a peaceful mind etc.
But don't you think that only quite works if you have near-Buddhist-monk-level mind control? I think that for most people, perceived happiness is tied to hope or expectation or some sense of purpose.

@wim_v12e This is what I've been thinking about all day (not social contagion, which is what I've been replying to all day 🙃). Let's see—a few ideas.

① I don't think so partly because there've been so very many instances of things I've desired, sought, and acquired—objects, experiences, relationships, skills, lifestyles—to find myself right back to where I started. These certainly make my life more comfortable & enjoyable, but spending my time on more of the same seems pointless level-grinding.

@wim_v12e ② You mentioned monks—for thousands of years I think monastic life was helpful to get the observations described above, because regular life in India, Central Asia, China, Southeast Asia, et al. was hard and the sangha offered the "leisure" & "freedom" that today in developed nations is afforded by a normal middle-class lifestyle. Today, monastic vows (even Seom/Zen) I think would distract from these realizations—surprisingly, normal life seems a better vehicle for such life progress.

@wim_v12e ③ I guess that's why I propagate Buddhadasa Bhikkhu and Stephen Batchelor et al.: because people seeking things and overestimating the impact of their acquisition on their personal satisfaction (in this case, devs and data science knowhow) underlies many neuroses.

Even when these things are things typically considered very positive: most readily see new Nikes as hedonic, but I posit pursuing skills, friendships, mentorships, exercise, democracy, all have same pitfalls as sneakers.

@wim_v12e I rush to say that I am a lover of all the above—I support democracy, friendship, more Haskell, and new Nikes.

But they've all fallen into the same pot for me—things to fill time with, the pigments with which I paint my existence, while I get on with the real task at hand: dealing with the fact that pursuing all these things leads to nothing but more pursuit, more time filled, more pigments gathered.

Pursuing them doesn't make me better able to handle my feelings, confusion, sadness.

@22 Thanks for the very extensive reply. I agree, having things or experiences does not per se make you happy. And yet, the observation is that people in general strive to have things and/or experiences. I am sorry but I am somewhat obsessive and precise, and you did not address my main point. To rephrase it, do you think people do not need hope or a sense of purpose?

@wim_v12e @22 Wow. This is the most awkward and depressing thread I've read on the fediverse ever. It reminds me of sad, long-lost acquaintances from engineering school.

@alice @22
Hi Alice, can I ask which aspects you found awkward and/or depressing? It may sound like a strange question, and I guess it just shows I must be weird, but I didn't experience it like that at all.








Sign in to participate in the conversation

The social network of the future: No ads, no corporate surveillance, ethical design, and decentralization! Own your data with Mastodon!