thekiptxt

> When caught doing something they’re ashamed of, people say things like, “That’s not who I am” or “That wasn’t really me.” They are implying that reputation-damaging acts are not reflective of their true self. This doesn’t mean they are lying: Many people really believe their shameful acts are not reflective of who they are deep down.

To be charitable, TFA is loosely using truth in the sense of “not intentionally lying,” which is somewhat common. But yeah, just because actions align with my mental model of self isn’t an indication that the model is accurate from the perspective of a third party, which is obviously the more important consideration.

I’m kind of sick of these shock value statements which then recover to something sensible by bending words beyond the breaking point.

One time I needed to sort some data arbitrarily — the resulting order did not matter, it only mattered that it was the same for the same data in different orders.

My senior engineer advised me against using Java .sort() because “we didn’t write it so we couldn’t be sure it would do the same thing every time.”

This reminds me of how chess tactics feel so obvious when I’m on a tactics trainer, yet I can’t identify them in games where they actually happen.

Or code that “obviously” has a bug only after it caused an issue in production.

Would be nice if someone warned me before I do a PR: “ok, fyi there is a subtle but devastating bug in the new_feature.cpp”

I didn’t question the math, but what made me intuitively believe the Monty Hall problem was this: Me: I am thinking of a number between 1 and a million. If you guess correctly, you get a prize.

you make your guess Me: Ok, I will reveal that the number is either $YOUR_GUESS or 452,871. You may keep your original guess or switch to 452,871.

Of course, you would switch, because the chances you guessed correctly originally were 1 in a million.

> YOLO releases are common with: Side projects. Early-stage startups with no customers. Mid-sized companies with poor engineering practices. When resolving urgent incidents at places without well-defined incident handling practices.

Whoops, the author missed “massive companies when no one is looking.”

I’m (cautiously) optimistic about this. I’ve found that magic of writing code evaporated for me when I joined the industry and spent most of my time fiddling with environments. Your workstation crashed? Be prepared to spend an entire day doing nothing but getting another one initialized.

After a few weeks on the job this dies down, but I can’t count how many times I’ve been curious to try something, only to dismiss it because I might have to mess with my environment for a good chunk of time.

I’m so sick of the sensation of getting a ticket, knowing immediately the 5 minutes of coding I need to do, and then spend the rest of the day with a flaky environments until my tests successfully run. This is at a major, FAANG-size tech company.

I too have been wondering if this is mostly placebo. For familiar words, don’t our brains automatically read entire words as a single fixation point? I.e. we don’t read in terms of letters, but see the word as an atomic unit?

I would guess that my brain, by seeing certain words hundreds of thousands to millions of times, has found a more efficient way to identify it than the same word with a few letters bolded. Of course, this is based on no data, I’m genuinely curious if this helps.

And on iOS, _every single time_ I share a link from HN, that website persists in screen time.

E.g. if I were to share the link to TFA in WhatsApp, my screen time tomorrow would say “21 hours on apple.com”

Only clearing browser data fixes this. It’s absurd.

I’m guessing it doesn’t need to be all that accurate for certain use cases.

Even just knowing the length of the password, estimating which keys in the sequence are capitalized (if Shift behavior is fairly easy to pin down) and being able to pin each key down to 5 possibilities would make a 20 character password trivial to crack. Right?

The leading number is one thousand. In many cultures, the period `.` delimits thousands, the comma `,` delimits fractions.

E.g. the price of the salad is €12,70.

Yes, you’re right.

I was presuming you were referencing users who know how to work with files/folders, but couldn’t describe them precisely when put on the spot.

For example, I’d consider “Well, folders are containers for files and other folders. And a file is, well I don’t know, but an example is a Word document, or a jpeg.” to be a good enough description. That person is ready to roll.

And I think we agree that systems designers have a bias of being able to describe and reason about their abstractions precisely, and sometimes make the fatal flaw of assuming that the end user will have this too.

I can relate to this, and it gets worse.

Sign-off would take 24 hrs to process, and the mysterious entity that signed off would have no context of our product or the changes, and no way to assess the risk.

Then, due to our lack of trusted regression testing, every damn change would take ~6 developers an entire work day to manually confirm that everything was working.

Why? We were measured on “number of tests.” The tech debt was too high to write quality tests before the next review, so we opted for quantity if we wanted promotion.

This was the hot path in a major (Fortune 10) financial company.

I don’t think a definition-first approach is all that relevant for most users.

People also can’t easily _describe_ the difference between a dog and a cat, but most can distinguish between a dog and a cat and have an intuitive understanding of the implications (e.g. they expect the right behavior/traits when they see a four-legged animal).

I think one problem is that the people who design computers and software must be perfectly precise when building them, and easily forget that this is inherently not how most people interact with most things.

> Google Docs uses a "Markov Chain" to predict the next word for autocompletion purposes. In this case, of course, since we've already written "And" 5 times, the only logical next word would be "And", as showed in Djikstra's 1989 paper on the subject. Therefore, the Markov Chain never terminates and hence the memory chain overflows with infinite ands.

Does anyone know why this bug doesn’t repro for some words other than And if this is the case?

Step 1: find one Step 2: emulate them

I’ve never gotten past step 1.

I’m trying to figure out how to leave the industry without taking a 50%+ pay cut, since I’ve found out that a) I am bad at software engineering and b) I don’t particularly enjoy it (in the professional setting)

This is so cool. I’m a few years out of college at a major tech company and have taken to reading some books about this era. I find myself envious of the seeming purity of the time when it was all getting started, and notice how scrum boards, discussions of negotiating TC are absent in these stories.

I don’t remember an Internet without Facebook.

To me, Ritchie is a legend, Ullman was someone who wrote some of my textbooks. To hear an account of an intern working side by side with some of these people is amusingly bizarre.

What a fun idea! Didn’t get today’s just because I didn’t know the song and don’t really listen to that genre. But I DID guess that artist choosing an arbitrary song. Could the yellow tile of wordle mean something like “correct genre or artist”? Alternatively is an easy mode possible that accepts the artist only as a correct answer?

I’ve always been really curious about how other developers visualize our system. I know that we tend to use architecture diagrams as a common language for major components, but I don’t see the system in boxes and ovals, each of the components are very different. Sometimes they become characters, with personalities, quirks, and flaws. Sometimes they’re looming and intimidating, sometimes they seem childish. When I’m onboarding to a new codebase, the better I get to know these characters, the better I know the system.

The relationship of wordle to cheating is underrated. The fact that it’s so astoundingly easy to cheat (the daily word is global), and falsify your game board (just text) makes dishonesty completely undesirable. If you post 1/1, people will assume you hit the once-a-year lucky first guess, or that you’re cheating. There are no high scores, no leaderboards, no reason other than personal satisfaction to work through the puzzle.

> Let other people say no for you.

Wow. Upvoting because this is such a succinct and powerful way to say this, and I’ve never heard it before. i. e. It’s not our job, domain, or in our interest to reject ourselves from a job we’re interested in before applying

I had the same thought. It only recently hit me how much of the world around me is sneakily driven by individuals wanting to strengthen their resume.

I had underestimated how often the frameworks/languages chosen was influenced by a developer simply wanting to built up skills in that area, rather than it being the best tool for the job.

But, interestingly enough, it seems to me that this can often lead to great results! The developer is motivated both to prove that their decision could solve the problem and also to learn the technology for reasons beyond that project alone.

It seems that might be what’s going on here: the graphic designer is flexing, and a page about a font is trending on HN.

Sure, but criticism comes in different flavors. “This software is difficult for me to use because of X” is a valid and fair take. “This software/development team sucks” is uninformed criticism if someone doesn’t know the full context. For all I know, the developer may very well be aware of my situation, but there’s some reason that X must exist.

I worked for a very large financial services company, and when I told people this, they would commonly bring up something to the tune of “your app/website sucks because of X!” And the response was very commonly: “we know, but we’re legally required to do this.”

I also worked on film projects, and occasionally someone would catch some goof (e.g. continuity error) and say “I can’t believe you missed this!” My (internal) response would be: “I didn’t overlook this. I have seen this film hundreds of time in slow motion during editing. We just didn’t have the budget to correct it, and it drives me nuts every time I see it. I could point out 70 other minor errors that YOU missed.”

It always reminded me how often we (myself included) mistake our own mental models and experiences for a comprehensive-enough understanding that we can pass judgement.

Lots of good advice here, here's how I kept sane:

Focus on your strenghts for the two components of getting an entry-level job: (1) landing the interview and (2) landing the job.

Landing the interview is some percentage of the number of real leads you have, based on the strength of your resume.

To get real leads (i.e. resume won't get trashed): connect with fellow alums, ask everyone you know if they know someone who can put you in the pipeleine. Shotgunning applications are not real leads. If you have a small network, then focus on strengthening your resume as much as possible. If you have a small network, find creative ways to expand it.

Landing the job (given that you have an interview) is a function of luck, likeability, and performance. You should intuitively know if your weakness is your charisma or how well you can answer technical questions. Hammer on the weakness.

The benefit of thinking this way is that I could survive rejections without thinking that I was a bad candidate: I just had to tweak some component of how I came across. From there it becomes a problem to solve, rather than a depressing result to accept.

Best of luck! Like others said, the first one is the hardest, but it can be done.

Would you say this is the case if one speaks German (or French) already?

I've always thought that recommendation services would be better of this logic were layered in. E.g. I'm on YouTube and 1 in 10 recommendations are popular/trending videos that the algorithm doesn't think I'll enjoy.

Aside from the obvious necessity of bubble-breaking, it has the nice side effect of retraining the recommendation engine without rampant confirmation bias.