Prescriptions Are a Dead End | Hacker News Reader

Prescriptions Are a Dead End

garybasin.com

44 pointsgbasin30 comments

C’mon man! A blog post just to link to a twitter thread? Should be the other way around ;p

For those like me, averse to twitter threads, a more readable form.

https://threadreaderapp.com/thread/1274859897826488322.html?...

I’ll get over it. I am interested in the concepts, it’s something I’d like to think more about.

When/ what areas do prescriptions (or recipes, programs, protocols, and what have you) prove to be useful and successful, and where do they break down?

Prescriptions seem to work well in "static" systems. The ones that don't involve competition or comparison with other humans. Single player games, basically. Fixing a car. Building a house. Open heart surgery.

There are no static systems. Every prescription or practice should come with a use-by date.

Otherwise some hapless new guy will take outdated practice for good money. It is also why people straight out of school are not that great unless they learn on their own - they generally know outdated practices.

Instead of prescription, a guideline is better, especially a clear statement of goals to strive for. Examples or exact methods are secondary.

Even something as bound to physics as say civil engineering changes all the time. New materials, new analysis tools, methods to save cost... Imagine trying to follow a few year old guides on programming practices. Even one year old stackoverflow post can be hopelessly outdated, not to mention low quality. I've hit current linters recommending old, now wrong practice. (Java: replacing concatenation with format strings or string builder. Concatenation used to be slow before Java 1.7, now it's fastest.)

Sorry, it wasn't my intention. Typically I think out loud on Twitter, and then try to summarize it succinctly in a blog post

> summarize it succinctly in a blog post

as far as I can tell, that part was a failure. There's no content in the blog post.

Just messing around, couldn’t help myself.

I’ve ranted in the past about this, and had folks tell me they prefer a Twitter thread. So there’s that, and once I learned about thread reader, my life was changed!

I’ll check the podcast out!

Indeed. Thanks for posting the link!

I love this and yet it brings up a conundrum for me: explorative learning seems to be more effective, and yet people seem to want prescriptive learning.

I see a lot of people, myself included, clicking on the "how to" links, even though they almost never help me. I've seen this struggle with self-help books even more so.

Any suggestions on how to make explorative learning more sexy in marketing? Oh no, I just asked for a prescription! Perhaps I should just explore and figure out what works for me lol.

Again, I find when I explore and play with different solutions, I learn MUCH faster and more appropriately.

Anyway, grateful to have read this tonight to remind myself to stop reading so much stuff. Reminds me of reading a book by David Deida, I think The Way of the Superior Man, and in one chapter he said that we have all the answers we need within ourselves and then I paused and asked myself, "So why am I reading this book again?"

Thank you for this spark tonight :-)

Consider that “here’s how to do explorative learning” is also prescriptive. I tend to think of exploration as just exploring everyone else’s prescriptions, testing them, and seeing what appeals to me and whatever my goals may be. In that sense, I don’t think prescriptions are bad. I still think explorative learning is that way to go, but you need to have something to explore.

haha well said :-)

Some things are better learned prescriptively. Like what's legal in a given area: exploratory learning by trying things and getting arrested repeatedly is a bad plan. Other things are better learned through exploration. There's no one-size-fits-all strategy.

Ah, fair point! I really like that example, especially for legal aspects. I think I perceive the world to be so focused on prescriptive learning that I lean too far in the other direction to try to re-balance, without saying that it's about bringing things back into balance.

Thank you for bringing this up :-D

Glad to hear it. Enjoy your journey!

I think we spend too much time online. I have read this same idea over and over again in the last few weeks. The idea that we study too much and do too little. The idea that reading a walkthrough is easier than doing the thing, but it feels about as good.

With the internet there are a billions of How To articles, and lessons, and lectures, and advice articles. It is easier to read about how to do something now than ever before. Often its the same entry level info repackaged over and over. While the internet symbolizes limitless knowledge, it has started to replace real knowledge gained through practice.

oh I thought this was going to say doctors are bad at prescribing drugs and so we should just 'tantum quantum' it

but author is actually saying that copying napoleon will not make you king of france

a claim that has not been proven

Same here. Makes me curious to see an article/study about exactly that, perhaps with info on alternative ways you might be able to deliver drugs to patients without the use of a physical note or pharmacy.

It will not make you king of france -

It will make you l'empereur

l'etat c'est mois

I'd love to see this tested rigorously

I see it as a form of cargo culting and yes, see it all the time. It’s sadly a fundamental part of the educational systems, everywhere. For a current example: in school, science “experiments” are performed (make this acid, stick litmus paper into it, write down pH) for which there is a right and wrong answer. So people grow up thinking that science is about certainty when it’s the other way around. As a result they are confused by the ambiguity and unknowns around, say, COVID-19. They have an introduction to the process but not the basic point.

It’s like asking a photographer what kind of camera they use. Learning the syntax doesn’t mean you have learned a language.

I'm interested when I see this critique, because I definitely agree with you, but I also see a strong bias against new pedagogical approaches like Common Core math that focus more on teaching concepts like numerical composability and less on memorization and recitation. Personally, as someone whose mental map of how math works lines up much more with the CC standards, I wish I had been taught that way; the memorization & recitation approach fell apart for me when I reached higher level math classes.

Same problem happened when it was attempted in the 1960s under the name “new math”. And that’s what I mean when I say it’s a fundamental part of the educational system.

At the primary school level in particular, teachers aren’t necessarily (in fact are unlikely to be) that conversant with mathematics anyway; it’s just another subject to be taught. So as long as the results are correct, all is well.

"We start focusing on the How To, and too easily lose sight of the actual end goal. The prescription ⏤ the training regimen ⏤ becomes the object of our desire. To execute it perfectly. And just like that, we become robot instead of artist."

A counterpoint to that is the idea of "trusting the process". Work out the path to the goal and then focus on the path without distraction. Don't be mislead by hunches that you incorrectly think will get you there, or bad habits come naturally. There'd be issues though if the goal moves or if a significant part of it all was about enjoyment.

I am a big fun of learning best practices but also of iterative “projects”. You know why most business projects fail? Because they are too big and to reduce risks the tendency is to follow _only_ best practices. This doesn’t work because each project is different with its own set of challenges and unknowns. Prescriptions are a good guidance but only sticking to prescriptions is path to failure. These days I rather iterate to a solution the same way startups iterate to a product. In fact I started to call our team’s project managers product managers instead.

That's the dangerous way of thinking. As far as I'm concerned and seen projects fail because their either build the wrong thing, or don't deliver (on time). The practices can slow you down, but they can also speed you up.

Sometimes the failure to deliver is a managerial failure or the problem is too complex for the team tasked with it to deal on deadline. It's rarely because of "we keep code clean and production running" or "instrument everything".

What I've seen instead much more often is that good practice is replaced with common practice or deviation (problem) is normalized.

The latter leads to some rather interesting failures - bugs, and failing to deliver. Sometimes explosions if you're dealing with rockets.

I've also seen bad workflows cause serious slowdowns and issues, while not being replaced because things are "done" on paper. I find this on par with lying to your boss, which it is. Also in case of bigger multiple teams, tardiness to implement improvements because someone might have to change something about their setup, instead of engaging them with a discussion, leaving it for a flag day. (When everything invariably goes wrong.)

AKA a recipe. You can follow a Michelin starred chef's recipe all you want, but you're not getting your own star that way. You actually have to learn to cook. But recipes can still be useful anyway.

But the statement is extremely context dependent and depends on if the reward requires novelty.

Novelty: If the person you're copying isn't doing something novel (like playing Bach on piano) then in all likelihood you will get the same reward did. If you copy someone who built a CRUD app for their company and do the same for yours you'll probably get the same reward.

Context: If you make a delicious meal for yourself, you and your SO, your coworkers, your friends, or the public the rewards are all different and context can sometimes be complicated so it's really hard to know if captured it.

Exactly. It's obvious when you think about. And it's obvious when you start down the path. But then many forget.

Yeah, we see a lot of these here. At one time, I thought they must be valuable, because successful people were telling everyone else, for free, what it was that made them successful. If we just followed their advice, we too could be successful.

I now see it differently, and though I'm quite certain I was wrong before, I'm less certain that I'm right now. But here are my two thoughts on this anyway:

First, no successful person, that I can recall, has ever credited their success entirely to following some other successful person's habits. No great inventor has said, "yeah, I just looked up Ben Franklin's habits, and followed them exactly, and here I am." People have credited other people for inspiration, for guidance, for mentorship, but if there's any evidence whatsoever that a person's success is entirely due to their habits -- that is to say, that the same success can be replicated by anyone who follows exactly the same habits -- I haven't seen it. Likewise, no successful person has ever turned out to be a clone of a previously successful person. Each one has made their own name, developing their own habits according to their own innate strengths and weaknesses. At YC, Sam Altman did not become Paul Graham; sama continued to become the most sama he could be, while pg was the most pg he could be.

Second, I've reluctantly decided to believe -- careful choice of words there -- that these sorts of "prescriptions" are not only not helpful, they are unhealthy. The natural conclusion for a shlub like me, having read so many of these, is, "if that's all it took for that person to be successful, then why am I not?" Am I lazy? Perhaps. Stupid? Too arrogant? Not arrogant enough? Am I not sacrificing enough of the other things I love? Am I sacrificing the wrong things? Too many things? Have I read the wrong books? Do I have the wrong priorities? Am I broken?

A lot of people are fond of saying that certain fields of study aren't science because they're not testable or repeatable enough. I submit that the nature of success has never been tested in the scientific sense. It's all, mostly, cargo-cult flim-flam. There is no way to know whether one successful individual could be born into entirely different circumstances and still come out as successful; it's unknown to what extent biology, early childhood, developmental psychology, and dumb circumstantial luck each factor in to the result. There is no equation for it, no universal constants.

25 million copies of The 7 Habits of Highly Effective People have been sold. 25 million! Has it made 25 million successful people? 10 million? A hundred-thousand?

Maybe I could have made different decisions and found some great success years ago. Maybe it is yet to come. (See? I've still kept my youthful naivete.) Or, perhaps, my equation long ago determined that I could only be a decent uncle, or mentor for some better programmer, or maybe just... not a shitty person. That one's still a work in progress.

So there it is. I've come to see those essays "On How I Became So Successful" as approximately indistinguishable from heavily-altered Instagram photos of pretty people on beaches.

I'm sorry, but I don't think this is true at all. Like not even remotely. The closest thing I could say is that prescriptions are almost never sufficient to achieve real innovation, but that's a far cry from "dead end." It's like saying that flu vaccines are a "dead end" because they don't cure AIDS. Sure, they may not be not sufficient, but no one would want to get the flu without an immune system.

We should find it shocking how many "new" things aren't really that new, but just looking at some existing problem slightly askew and seeing a cleaner arrangement. Pure accident is likely an even bigger cause of the new. This is true of the invention of calculus to the discovery of penicillin. (Does anyone seriously find it mysterious that Leibniz and Newton independently invented calculus at roughly the same time?)

"Prescriptions" themselves often take the form of a procedure or algorithm that is used to improve some instrumental or operational aspect of achieving the new. If I follow the Pomodoro technique, for instance, that's not new, but it might help me actually stay a bit more focused and even creative. Maybe it won't. It's just a prescription, it's not going to change my life by itself, but it might move me in the right direction.

Finally, prescriptions can often help us do "dimensionality reduction" on the present and help filter out a lot of the extra _signal_ (not just the noise) from the present, which is simply too vivid for a human to process. Our brains do this normally with most sensory-level information, but we can use prescriptions as sort of a higher-order, more conscious variant of this. Since most genuinely new things actually involve tearing down a complex thing and rearranging it just a bit better, filtering out everything except for what is most likely to help you solve a problem (and thus create something new) can make you much more productive and see the problem more clearly.

I'd also observe that most great artists find some set of constraints under which to work. Picasso was excellent at producing classical-style art (most great artists actually are, contrary to popular "my-five-year-old-paints-better" bullshit), but constrained himself to imagining the objects of his paintings as cubes. Monet, Manet and Van Gogh constrained themselves to seeing a very short (possibly infinitesimal) glimpse. I have had Hollywood screenwriters say to me that they can get nothing done if they don't constrain themselves in some way, even if it's something like "I'm writing sci-fi today."

TL;DR: Prescriptions are all they're cracked up to be, and nothing more; (almost) never sufficient but often necessary.

Reply on news.ycombinator.com