June 25, 2026·RustProfit Team

What's New in Rust's "Friday Devblog 9" Update

May 2014, mid-reboot. Weapon world models returned, crafting finally got a cancel button, the procedural terrain went mountainous, and Facepunch started testing a game UI built entirely in HTML - while re-explaining why the experimental branch existed at all.

What's New in Rust's "Friday Devblog 9" Update

Friday Devblog 9 landed in May 2014, back when Rust was two games at once: the "default" build everyone had bought, and the from-scratch "experimental" branch that would eventually become the Rust we know. This entry was a grab-bag of reboot progress - weapon world models, sound, voice chat, terrain - written during a week where Garry admitted he'd cut back on the night shifts because it was his birthday.

The reboot, explained again

A big chunk of the blog was spent re-answering the same question. The default branch on Steam was not being updated; everything being written about in these devblogs lived on the experimental branch, which was getting daily updates. The plan for switching them over was refreshingly blunt: when more people were playing the reboot than the default, the reboot would become the default, and the old build would be shunted onto an archive branch so anyone who preferred it could keep playing.

That switch was explicitly "a while away." The reboot looked and performed better, but it was missing major features - AI and building among them - and Facepunch said outright that they didn't expect anyone to properly play it until those were in. As for why not just patch the old game: they'd spent a few months trying and failing to update at the rate players wanted, and the original code was limiting in "pretty much every way possible." Rust had been released expecting a few loyal players and gradual cleanup; 1.6 million sales later, that plan no longer fit. Everything on experimental had been built from scratch in the preceding few months.

Weapons, crafting and world models

Re-implemented weapon world models meant you could finally see what other players were killing you with. The team wasn't happy with how these interacted with animations, so Alex was re-rigging the player model to let weapons be treated more like clothing. The animations were working too.

Craft cancelling also arrived. Crafting itself had shipped the week before, but without any way to back out of a queued item - that gap was closed here. Resource stacking was brought across as well, working essentially the same as it did in the current version, though the blog cheerfully noted the context menu was broken and the gamma was wrong.

Terrain and rendering

Andre updated the procedural terrain generator, pushing it away from the cartoony look toward something more mountainous. Petur was working on new terrain textures alongside it, plus a pair of in-engine art tests.

The reason the gamma was off, and the reason screenshots suddenly looked different, was the switch to linear colour space that same week. It's the kind of change that touches every material in the game at once, which is why several other systems in this blog came with "the lighting looks wrong right now" caveats.

Sound and voice

Garry implemented a system for sound playback - foundational plumbing rather than a feature. More interesting was the footsteps experiment. The complaint was that footsteps in the current version had no relationship to the player model's animation, when they obviously should. The test drove footstep sounds and effects from animation events instead. It worked well, but with a catch: it meant animating players you can't see but can hear, so the team left it as an open question about whether the cost was worth it.

Voice chat came back too, this time built on Steam's voice system rather than Unity's. The honest framing was that they weren't sure it brought concrete benefits - it just meant the encoding and decoding happened in C++, probably threaded, and Facepunch didn't have to maintain any of it.

Clothing and character work

Tom and Xavier were sketching out improvised clothing, and the reasoning behind it says a lot about early Rust's design philosophy. The point wasn't protection stats. Strapping tires to yourself leaves you exposed everywhere else, and smearing mud on yourself does nothing mechanically - but you might want to, and that choice was the feature. You don't have real clothes, you don't have much stuff, so you improvise.

Dan finished the bota bag, and Tom was re-rigging the face so characters could close their eyes - small, but a prerequisite for basically all later facial work.

The HTML UI experiment

Players on the experimental server had spotted a webpage lying on the floor, and this was the explanation. After talking it over with Bill, the team decided to try building Rust's UI in HTML. Less mad than it sounds - Garry's Mod's main menu was already all HTML - and the upside was substantial. It would move UI logic and rendering out of Unity into its own process and therefore its own thread, so it might actually end up faster. Building interfaces in HTML/JS/CSS was described as "about a million times easier" than doing it in the Unity editor. And as a bonus, the stylesheet could be exposed for modding.

Read back now, Devblog 9 is a snapshot of a game being rebuilt in public, one subsystem a week, with the developers openly unsure which experiments would survive. Some didn't. The reboot itself very much did.