Code Department

Fan of HTML, CSS, Datastar, Lua, R, Julia, Odin, Zig, Go, Kotlin, Elixir, Gleam and Roc who doesn't yet have time to use most of them.

[Code Department @code_department]

According to https://t.co/adR2QOVwRT 🤓 sites using @htmx_org 🗿or @elixirphoenix 🔥 perform better on CWV, INP & Lighthouse score than sites using React, Svelte, Vue, Angular and even Rails Hotwire Turbo, Laravel Livewire😎 probably due in large part to the lower page weight 🗿

[Code Department @code_department]

@ThePrimeagen Probably because it doesn't understand anything and is just predictive text that is more inclined to write the next line of the program than an assert.

[Code Department @code_department]

Unless you actually mean second best after some other option, in which case it's the least worse after that better option but in that case just say second best.

[Code Department @code_department]

btw it's "least bad" not "least worse"

[Code Department @code_department]

@ankkala Technically, there probably weren't, because the other side of our galaxy was empty, because the galaxy was probably in a different position, but it's impossible to know because all position is relative.

[Code Department @code_department]

@s_ntafdvy @BKitten16 @ankkala A past event isn't "current' just because you only just got the message about it.

[Code Department @code_department]

htmx vs Datastar fundamental differences in design:

Datastar: let's make the core as small as possible so that every feature is a plugin, and use typescript to achieve this & reduce errors & use SSE & CQRS as the default

htmx: let's not do that

[Code Department @code_department]

htmx vs Datastar fundamental differences in objective:

htmx: let's complete HTML by generalising what it can already do, but not add many new features that aren't generalisations of what's already there.

Datastar: let's complete HTML so it can do everything that the browser can

[Code Department @code_department]

@DougChampion @ThePrimeagen Another thing that is hyped a lot is something like the singularity where machines can replace not only programmers but the vast majority of current intellectual work. I think they could replace a lot of the mechanistic non-creative work but not the actual design for many years.

[Code Department @code_department]

@DougChampion @ThePrimeagen Of course it's also used to mean what I'd call GAI or "General Artificial Intelligence", like just multi modal LLMs with image and sound processing, which is not 50 years away obviously, so Prime is clearly not talking about that.

[Code Department @code_department]

@DougChampion @ThePrimeagen AGI could be 50 years away or more and is being claimed as around next Tuesday.

[Code Department @code_department]

@BenjDicken Didn't this change come after the success?

[Code Department @code_department]

@ThePrimeagen Bro I'd get so much anxiety if I were generating that much code for my apps.

[Code Department @code_department]

@SIRHAMY It'd be nice to have more programming languages that are "X but with breaking changes".

I don't care if it's called "Go 2" or "Julia 2" or if it's called "Gone" and "Julietta". Neither stops you from continuing to work on improving the stable versions Go 1 and Julia 1.

[Code Department @code_department]

@SIRHAMY It'd be nice to have more programming languages that are "X but with breaking changes".

I don't care if it's called "Go 2" or "Julia 2" or if it's called "Gone" and "Julietta". Neither stops you from continuing to work on improving the stable versions Go 1 and Julia 1.

[Hamilton Greene 🐷🦀 @SIRHAMY]

Programming languages strive for stability after reaching mass adoption to avoid breaking existing usecases. But this stability comes at the cost of improvement.

• C# struggled to add unions
• Python can't enforce types
• Java has unsafe nulls

[Code Department @code_department]

@ai_ops_lead @mmvlad @ThePrimeagen @AnthropicAI It least that wouldn't be inversely associated with code quality like LOC is.

[Code Department @code_department]

@ai_ops_lead @ThePrimeagen @AnthropicAI You may have already failed if you're requiring it to be fully objective, depending on what your definition is of fully objective. Improvements in user experience (subjective) is what matters. That can be measured "objectively" though: user surveys, bug reports, load times, etc.

[Code Department @code_department]

@ZH1YGD @TheGingerBill Don't use an LLM for stuff like this without checking the sources. Half the time it's summarising someone making something up on reddit or hacker news.