jordancostello.bsky.social
143 posts
127 followers
82 following
Getting Started
Active Commenter
comment in response to
post
I could write some code that handles every possible scenario and maybe runs super fast. But if it's hard enough to integrate with, then it's less likely to be useful or cheap to use
comment in response to
post
Yep. Gotta write stuff at the ground level that's useful and intuitive to use so it can be quickly built upon by future you or a team.
Some skills apply before the code too: defining and communicating intuitive interfaces pays dividends
comment in response to
post
Think we can boolean algebra here to simplify:
Sorry
comment in response to
post
Closest thing I can think of was MGS2, but that was more about the general internet and centralized, automated information control for civilization's survival...and sort of about nukes too...I think 🤔
comment in response to
post
Math never lies. Good work!
comment in response to
post
reliable way to get a 10x programmer: let a weird cowboy coder design the system. everyone else will be slowed down by the unintuitive design, weird cowboy knows it by heart because he made it. one guy works 100% faster than baseline, everyone else works 500% slower than baseline.
comment in response to
post
I'll mark the calendar.
Excellent work you've been sharing! 👏
comment in response to
post
And I would add also: that if you can crush a population with stress to survive and navigate extra complexities (real or not), then conspiracy theories become an increasing necessity to cope.
comment in response to
post
Illuminating followup question:
"Do you want to understand?"
comment in response to
post
Realizing that people can hate writing unit tests because they're forced to do them when:
- constraints and usage are not defined, or the those are so volatile that many tests are invalidated and thrown out
- they already wrote a unit first, then must retroactively define its constraints and usage
comment in response to
post
I might be okay with feeding in the requirements to generate the tests. But if I have to review the tests anyway, I might just write them myself
comment in response to
post
If it is given the function to generate the tests, I don't think I'd want to risk the LLM getting the real intent/requirements of the tests wrong either. If that's wrong and the tests pass, then I just get a false sense of security
comment in response to
post
I'd write the tests myself, but that's because I'm a freak that enjoys that sort of thing
comment in response to
post
bullies need victims, victims do not need bullies, same as it ever was
comment in response to
post
Chef's kiss/10
comment in response to
post
Could try your luck with discogs 🤞
www.discogs.com/sell/list?ma...
comment in response to
post
Or someone else suffers the consequences of your choice and you learn absolutely nothing 😩
comment in response to
post
😭
comment in response to
post
That nice balance is definitely crucial. Your description of simple and flexible access to a complex solution is spot on.
When the balance is off, oh boy do you feel it.
comment in response to
post
FkTht 💀
comment in response to
post
Nominating Foolean for the 4 variant
comment in response to
post
Or maybe more simply: how often does the original author need to be contacted when there is a perceived problem with what they produced?
comment in response to
post
Maybe: if something unexpected happens, how costly is it for someone else to comprehend or resolve the situation?
That someone could be anyone in the chain from end user to developer.
I think your example fits within that.
comment in response to
post
I get the disproportionate sense too. I don't know if it's because combat is the simplest way to implement interactive conflict, or it's what most easily resonates with the creators or the audience for interactive conflict. Or maybe combat gets the biggest ad pushes because it's a known formula.
comment in response to
post
I keep hitting the like button and nothing changes
comment in response to
post
Plus, with teams, you can't share more than one user's screen at a time. It's tragic.
comment in response to
post
I'm just tired rn 💀
comment in response to
post
This is an old topic. It's all obviously subjective. People aren't wrong for comprehending how they do. It's more important that people who align in how they communicate and build do work together, I guess. Still good to try to meet people where they're at.
comment in response to
post
(I would not consider copy-pasted source to be reuse, I'd call it reimplementation)
comment in response to
post
I would rather have self-explanatory tools working simply together than having each of their complex inner workings sprawled over a long single page.
I would prefer tools that can be individually trusted and shared with other people over a massive task that can't be reused by anyone.
comment in response to
post
I would rather have complex processes turned into externally simple tools, so that I can use those tools to manage comprehending and solving even more elaborate problems.
comment in response to
post
I would much rather have patterns that are cropping up repeatedly to get decomposed into externally simple containers: a few pertinent inputs, and one or two pertinent outputs. Stuff that's easy to reuse or consume, and can basically be understood at a glance.
comment in response to
post
I especially can't manage balancing the above factors all occurring at the same time.
Notably, some people can.
comment in response to
post
I don't want to be scrolling up and down and left and right and need to memorize what was off the screen to know what the current set of 100 lines does- and especially why it does it.
comment in response to
post
And with a huge chunk of code that "does everything" I can't keep track of the 15 mandatory parameters, which ones are relevant for my task, and how they each intricately work.
I can't keep track of the 5 sets of nested parenthases in a single line sometimes.
comment in response to
post
I don't want to manage that big mental map either. I can't hold in my head the operations that happened 50 lines ago, or keep track of which deeply nested loop and nested conditional I'm in from 200 lines up. I can't keep track of the single sprawling line that is packed with 10 different calls.
comment in response to
post
I think I get that other people don't want to be hopping around a file or hopping between multiple files to figure out what's going on.
But I imagine if abstractions are accurately named to describe what they do, then you don't need to hop around to comprehend things.
comment in response to
post
Brutal. Hope you find something better soon
comment in response to
post
It's true that usage supersedes th books. Just kind of a shame when we decide to settle on stuff that's easily confusing, misleading, or contradicting. Not everything has to be literal - that's boring and not my point. At least accurate would be nice. We can do better.
comment in response to
post
I wonder why "virtue signaling" needed to exist when phrases like lip service, hollow words, hypocrisy, or insincerity were already there
comment in response to
post
Not upholding your verbal values with actions is probably a widely agreed upon criticism.
But the definition of this simple phrase seems to be wide open to disagreement, and I imagine that's intentional.
comment in response to
post
The literal reading of the phrase doesn't line up with the meaning of how it is generally used.
This opens anyone who reads or uses the phrase at face value to be in unnecessary disagreement and argument with someone who has unintuitive contextual knowledge.
comment in response to
post
"Virtue signaling doesn't mean literally signaling virtues" makes it seem like there's inherently an issue with the language of the discourse.
comment in response to
post
Rumor is that Valve is focusing on physics and interactivity for an upcoming title. Whether or not that gets released, I want that to be true because I completely agree with you: reactive systems are more important to push than textures and lighting at this point.