tfw

note how the win8 start screen was functionally identical (and IMO aesthetically superior) to the standard android home screen, which, curiously, nobody ever seems to get triggered over

this is the one and only point I will ever defend Microsoft on. the win8 start screen was genuinely well-designed and useful, and a huge improvement over navigating through trees of linear menus. people hated it because it was new and new things are inherently heretical

my hottest take of all: the win8 start screen did nothing wrong

when you need privacy online, you do not pay a scammer for VPN access. you install tor.

when you need to evade corporate/government censorship, you do not spam youtube with variations on the same content. you use goddamn bittorrent. or these days, maybe ipfs.

this is basic internet literacy y'all

i apologize for quoting the western bloc legacy media but that one was too funny not to clown on

so many incredible things are possible but so much harder than they have any material reason to be

I really hope AI codegen gets to the point I can just outsource all the gigabytes of mind-numbing boilerplate to a local Qwen instance and focus on actual game mechanics & storytelling

luanti is kind of a hot mess but it's the one engine i've actually been able to make real progress building a game on. but now Starlit is blocked on the new UI API whose development has been dragging on for years at this point, a monumental one-man effort with no end in sight

but most of my hopes are pinned on a better FOSS game engine being built some day. I would really love to see crownengine.org succeed

I still keep a wistful eye on the release notes in hope they'll get their shit together eventually. which isn't impossible, they did finally add support for compute shaders and custom rendering passes, enough that I was able to write a decent chromatic aberration effect

Godot has many aggravating problems but for me the dealbreakers were gdscript and the aggressively stateful design of every single aspect of the engine. gdscript in particular is so bad I actually started writing Janet bindings using GDextension before I decided Godot just wasn't worth the effort

the whole reason I tried to write my own engine in the first place was because I started out using Godot and it was so maddeningly awful I decided rolling my own would be less work than trying to cajole Godot into doing the extremely basic things I needed it to do

then dealing with the gruesome GL state machine proved so frustrating I decided in a moment of pique that my time was better spent learning vulkan. the vk API proved much less awful to work with, but there was just so much extra work I ran out of steam

biggest problem is I am just so jaded and tired of fighting with game engines. my last big project started out with raylib, which was such a broken, inadequate, cartoonishly unoptimized mess i finally had to rip out all the raylib code and replace it with raw GL4.

ofc the overhead for that would be insane even with like a 4B model

also curious about using LLMs to control NPC decisionmaking, how well that would work compared to traditional vidyagame AI techniques

have you heard about the industry term for customers who pay off their balances in full each month?

deadbeats.