tugu77

75 karma · joined November 24, 2024

tugu77on Markov Keyboard: keyboard layout that changes by Markov frequency (2019)

There is a whole usability universe between shift and AltGr, the former being accesible with the other hand.

tugu77on Markov Keyboard: keyboard layout that changes by Markov frequency (2019)

German layouts with [ ] { } in such horrible positions are killing all joy when programming, especially in curly brace heavy languages.

Changed to US qwerty 30 years ago, never looking back.

tugu77on After 3 Years, I Failed. Here's All My Startup's Code

That catchy title is what brought it to the HN front page though.

tugu77on After 3 Years, I Failed. Here's All My Startup's Code

Not every little idea, as useful as it might be, needs to turn into a hyper-growth business. Just make it an open source project on the side. If it gets popular then you get famous and hopefully some big company finds it so useful that they hire you and let you work on it full time. Konfig seems like a perfect example for this.

tugu77on Ask HN: SWEs how do you future-proof your career in light of LLMs?

Companies like that deserve to die out since the quality they bring to the market is horrendous either way.

tugu77on Ask HN: SWEs how do you future-proof your career in light of LLMs?

> Moreover, common people cannot measure programmers quality as easy

So you can't fire the bad apples because you don't know who they are, but you feel confident that you can pick out whom to replace by an LLM and then fire?

It's hopefully obvious to everybody that this is a hopelessly naïve take on things and is never going to play out that way.

tugu77on Ask HN: SWEs how do you future-proof your career in light of LLMs?

That "need" is not as a helper but a cheap way to train the next gen senior which by the time they are senior know the ins and outs of your company's tech so well that they will be hard to replace by an external hire.

If your approach to juniors is that they are just cheap labour monkeys designed to churn out braindead crap and boost the ego of your seniors then you have a management problem and I'm glad I'm working somewhere else.

tugu77on Ask HN: SWEs how do you future-proof your career in light of LLMs?

"For the price of a cab you can take so many ubers it's not even funny". Yeah, until the price dumping stops and workers demand better conditions and you start comparing apples to apples.

Why do you have bad programmers on staff? Let them go, LLMs existing or not.

tugu77on Ask HN: SWEs how do you future-proof your career in light of LLMs?

> Large companies are still thinking about privacy of data etc but all of that will simply not matter in the long run.

That attitude is why our industry has such a bad rep and why things are going down the drain to dystopia. Devs without ethics. This world is doomed.

tugu77on Ask HN: SWEs how do you future-proof your career in light of LLMs?

> My prediction is that junior to mid level software engineering will disappear mostly,

That statement makes no sense. It's a skill progression. There are no senior levels of anything if there isn't the junior level as a staging ground for learning the trade and then feeding the senior level.

tugu77on The tale of two Shanghais

I'm quite disappointed by the article. It's not really about two Shanghais. Sure, it mentions an alleged divide, but it then doesn't actually explore the contrast but spends most of the time with skyscaper porn. I'm sure the average Shanghaian would be a little offended if their city is reduced to those.

tugu77on Gukesh becomes the youngest chess world champion in history

Ironically, in the match that Gukesh won and that made him world champion, he lost rating points, inceasing his distance to Magnus.

tugu77on Gukesh becomes the youngest chess world champion in history

Not when Hans or Hikaru are playing.

tugu77on Bitbanging 1D Reversible Automata

Automaton.

The singular is automaton. Automata is plural.

tugu77on Show HN: It took me 5() months to build a Plausible alternative

Have you considered open sourcing?

tugu77on The $5000 Compression Challenge (2001)

The rules should bar contestents from saving entropy outside the payload of the file. Whether that's in a file name or some file system data structure or in some timing side channel is insubstantial.

And once you ban that, it's impossible from an information theoretic point to win the challenge.

tugu77on Charset="WTF-8"

TIL

https://www.compart.com/en/unicode/block/U+20A0

Even Bitcoin is there. And "German Penny Sign"?

tugu77on Charset="WTF-8"

Type system ftw? As long as it's a blob (unnormalized), it should have a blob type which can do very little besides storing and retrieving, perhaps printing. Only the normalized version should be even comparable.

tugu77on The $5000 Compression Challenge (2001)

Easy.

Save the sha256 hash of original.dat in compressed.dat. The decompressor cats /dev/random until data of the right size comes out with the correct hash.

Now there are two cases.

1. The reconstructed data is actually equal to original.dat. Challenge won, cash in $5000.

2. The reconstructed data differs from original.dat. It has the same hash though, so you found a collision in sha256. World fame.

In either case, win!

tugu77on The $5000 Compression Challenge (2001)

Mike thanked random.org for the data. It's exactly what they are doing.