spitchell

@spitchell@indieweb.social
48 Followers
138 Following
704 Posts
I changed the way I work a couple of months ago to follow this advice, it made a big difference and my productivity went up by miles.
You shouldn’t use an em dash because people think it might have been shit out by an LLM — over my dead body!

I see people talking about "reach" and "engagement" and "algorithms" and whatnot on X vs Bluesky vs Mastodon and I feel so lucky that I have no commercial needs whatsoever.

It's 100% a feature for me that unless I actively include a bunch of hashtags no one will ever see my vapid, drunken, digital utterances.

#VapidDrunkenDigitalUtterances

I "uninstall" copilot from my teams sidebar every day, and the next day it's back.

If I'm also actually using it on purpose, however tentatively, surely it's just a matter of time before my resistance begins to crumble and I just give in to the need to rely on this bullshit for one reason or another, like I did with self service checkouts.

I feel like a frog being boiled.

Recently I've used a "AI" to generate mock data in a specific json schema, and it worked great.

I've also used it to produce an exhaustive list of handwavy arguments for a particular software architecture I want to critique.

In both cases I'm leaning into LLMs' specific ability to come up with convincing bullshit, but even still I'm troubled.

I am proud to consider myself a bloody-minded, socialist refusenik. So I'm obviously staunchly opposed to this ubiquitous commercial shoehorning of LLMs into every aspect of my experience as a web user and a frontend developer.

But I was bloody-mindedly opposed to self service checkouts for several years, until they became tolerably usable and the shops became intolerably understaffed, and now. I use them every single day.

Look, I know AI is controversial, but just for a moment, let's set aside our preconceived notions, our biases, the environmental impact, the massive cost to train and run models, the labor exploitation, the intellectual property theft, the inaccuracies, the mania it causes in users, the destruction of search, the deskilling of professionals, the devaluation of creative work, job losses, and lack of economic value from enterprise implementations.

Wait, what were we talking about?
Wait I think I withdraw my above statement because "demand chain" / "software demand chain" is just too good
At work we have started a new project called "Hot Chocolate Leather Queen" with sub projects like "transfemme" and "black justice". It has nothing to with any of that. We named it that to keep the prudish AI bots from stealing our work.

It takes a lot of energy to believe things that you know really aren’t true.

#ActuallyAutistic #EnergyManagement #AuDHD #BurnoutRecovery #AutisticBurnout

×
I changed the way I work a couple of months ago to follow this advice, it made a big difference and my productivity went up by miles.

@kenney

yeah I highly agree. For instance there is a saying in programming "first make it work, then make it fast"

@farooqkz @kenney where would „writing the docs” be in this approach?
@metaphil @farooqkz @kenney you would be surprised...

@uffke @metaphil @kenney

Some recommend writing docs before you write the program.

@farooqkz @uffke it is surprisingly helpful to start with a little readme file describing the features and how to use them, in order to get meaningful answers form future users @metaphil @kenney

@farooqkz @uffke @metaphil @kenney

I find that writing the docs helps me find both bugs and potential usability improvements, because it helps me step out of my own mind into an imaginary newcomer's, and that newcomer uses the code quite differently from the way I do. The newcomer is also less forgiving than I am. Thus careful documentation is important for the output it produces, but also for the effects it has on the developer and the code.

@CppGuy @farooqkz @uffke @metaphil @kenney Even doc comments help here. If you find yourself describing a lot of quirky edge cases, it's a good opportunity to reconsider whether the logic can be simplified or cleaned up.
@metaphil @farooqkz @kenney or making it secure ✌️
@luceos @farooqkz @kenney 𝗈̶𝗋̶ and accessible ✌️ 😇
and yes, I'm well aware of the inaccessibility of unicode strikethrough „font“ generators, thanks for not pointing that out, unfunny-part of fedi ;-)

@metaphil @farooqkz @kenney pick two

Fast, secure, documented, cheap, performant

😬

@farooqkz @kenney
"Premature optimisation is the root of all evil."

@farooqkz @kenney

first make it work* then make it fast

*with enough logging you can figure it out when it does break 😆

@kenney my daily struggle 😅

@kenney #AltText4You

A hastily-drawn circle that isn’t quite closed, titled “Just make it exist first”.

Beneath it, a perfectly closed circle, titled “You can make it good later”.

@jdeisenberg Thank you, I've added it
@kenney Besides productivity, I'm sure you feel much more at peace now.
@kenney this is also called "technical debt" btw. :)

@tomtrottel @kenney
Yes and no.
Something existing gives more concrete details, and at that new point you see more and clearer.
It is debt only if you leave it there for a long time. But initially it was a step that supported you in your climb of unknown and complexity.

Those do no mistakes, who start nothing.

@kenney This is difficult to learn!
@kenney beautiful wisdom. thanks for posting.
@kenney This is definitely Lisa Murkowski’s philosophy. 😑
@kenney I was about to doomscroll and procrastinate, I'm going to give it a go right now
@kenney my ADHD perfectionism need this reminder on the daily.
@kenney good that this work for code, but not for governance. The system has deteriorated, not gotten better.
@kenney That's a really good principle that I adopted years ago. Especially when I translate, which I do a lot. First: translation almost as stream of consciousness, superfast, then two or three correctional evaluations.

@kenney

make a mockup

management decides it is good enough
now mockup is in production and never touched again

@Joe_von_Saporski

I can't tell you how often this happened to me. I learned pretty quickly not to show my work until I was absolutely forced to. And when that finally happened, which it always did, I made sure to plaster warnings ("test system" or "WIP - output not verified" or "demo only") all over it and all its output. That still didn't stop them from trying to put unvetted systems that were still a WIP into production but at least I could say 'I told you so' later.

@kenney

@Joe_von_Saporski @kenney Unfortunately, that's true. Not only management is the problem also marketing.
If marketing guys are in your area, cover your monitor or show boring stuff!
@kenney Most people forget "later". :/
@kenney
I didn't need to be told that. I figured it out when quite young, probably in my twenties when I started writing things, possibly a teen.

@iclast @kenney

looks like you got the first part down, could use some work on the second judging from your website 😹😹😹

@kenney I have been yelling about it to my gamedev and gameart students for 5 years and a single one listened
@kenney just as long as 'release it' is step 3, not between steps 1 and 2.
@kenney I like the first one better.
@wcbdata @kenney The first is far more expressive.

@kenney mvp

:

small good first prototype

make it great ( give it enaugh space to breath )

scale it

:

minimum viable product

:

@kenney just make it work, then keep the temporary state forever.

@kenney

Sir R.A. Watson-Watt would be proud of you.

@kenney

Distance from couch now exists.

@kenney Which can be OK as long as the "later" bit happens, and it does depend on what "IT" is so many companies do the first bit but release the thing too early in it's lifecycle.

@kenney

This is the only way I know. Sometimes even the draft is good enough.

Very likely that the entire creation has been done that way. And we still wait that ends will meet. And that shit people become angels....
@kenney this is a) excellent advice and b) the advice I have the most trouble taking

@kenney There is however a fine art to making it exist first without digging any enormous holes that will haunt you for the rest of the product lifetime. That's one of the signs of a really good systems architect

For open source it's also often the case that make it work will pull in "make it work nicely" and "I've written an install document" people.

Sometimes even "I can't stand the random formatting in your repository can I please fix it" people!

@kenney
I use, "don't let perfection get in the way of good enough "
@kenney Gonna take this to heart
@kenney that's true man.

@kenney this can be great advice for individuals like me who procrastinate.

Not so great when adopted by for-profit corporations. Case in point: "AI" companies.

@kenney not sure how good of an idea this is 
@kenney I definitely agree. Just not sure why it is so hard for me :)
@kenney i like the top one too

@kenney

There are no great writers. Only great re-writers.

@kenney You're lucky you work somewhere that you get to make it good later… for a lot of us, that opportunity never arises.