"Get rid of the AI and it might be good."

This was my initial review, but after Facepunch decided to break my game less than 24 hours after launch, I felt compelled to update my review and elaborate on why this sucks. For reference, almost all of my time in-game has been spent working on my game, not playing.

In their infinite wisdom, at some point close to launch they put a fullbright button on F2/~, with no way to turn it off. My game relies on darkness to hide things and fullbright is equivalent to cheating, so I had to implement a system to detect when mat_toolsvis wasn't the default string "Normal". If this system triggered, it'd cause an instant loss.

The day after launch they release a patch to lock it behind sv_cheats, but in the process ♥♥♥♥♥♥ everyone who had their own system in place. Without sv_cheats on, running Sandbox.ConsoleSystem.GetValue("mat_toolsvis") returns null instead of returning the value. They've made it so sv_cheats doesn't just stop the client from reading/setting values, but not even the games can check. Because my system wasn't designed to handle the convar suddenly not ♥♥♥♥♥♥♥ existing anymore, it caused a bunch of false positives and ruined the experience of a bunch of people before I caught on and fixed it.

This is one of what must be thousands of examples of Facepunch making random breaking changes daily; changes made with no thought behind them (definitely not AI, for sure) and shipped without any testing. If you decide to fork over the ridiculous $26 CAD for this circus and decide that you want to make a game in it, you need to prepare yourself to babysit it every single hour of every single day just in case Garry decides to introduce a silent game breaking bug.

I remember about a year ago now Garry said he'd stop breaking games (https://sbox.game/news/march-update-8c52a424#release-cycle or https://sbox.game/news/june-update-d91005cc#dont-break-games), but in the last year or so 4/5 of my updates have just been cleaning up after their mess to keep my game operational.

Everybody has already said what I'd say about the bro-culture of the Discord, or the horrible documentation, or the horrible performance (Facepunch calling 3000 series cards low-end right now is laughable), or the super-duper-secret $100 USD fee you end up paying to export your games, or having to dual fist hammer and their bastardized scene editor for the best lighting, or the forced client-authorative server model, or the forced arbitration, or the AI-ridden games list, or the FOMO-slop, but I have a take about the games list I don't see many people talk about: the sorting.

A long time ago, the in-game browser used to be able to sort by thumbs and favorites, which should be obvious sorts to have. The things with the most thumbs are probably the best representations of your platform. Doesn't take a genius to figure that one out, but Garry woke up on the wrong side of the bed that day and decided that these sorts are too sophisticated for us and removed them. Now the only sorts left are the volatile sorts like updated and trending, which have already proven to just push junk because their algorithm sucks. Bring back the thumbs and favorites sorts, if a game is good it'll make it there.

Further, Garry dismissing the "this isn't GMod 2" complaints isn't entirely valid. He spent the better half of this projects existence hailing it as the sequel, promising refuge for the migrating developers, yet not even his sandbox mode is anywhere close to feature parity. He only changed gears to abandon GMod developers in late 2023, I guess he wants you to forget about that.

Lastly, I don't see people iterate enough that this isn't Source 2, because it's really not anymore. S&box is a great analogy to the Ship of Theseus; walks, talks and acts like it, but in reality it's made like an estranged nephew to Source 2.

If this were $5 CAD or USD, then maybe. If it ever gets to the state they claim it'll get to, then maybe the $26 CAD will be worth it, but promises should mean nothing from Facepunch, proven by the litany of promises made in blogs during development which were never kept.

After 8 or so years in development, I can say that it was worth the wait 3 years ago, but not so much now.

I bet Garry wishes people given the game for free couldn't review the game like he thought.

the seconds i finish installing it and launch the game i see my stupid sausage character all naked so i went to go customize and i so happen to open the shop and for the love of GOD why does it 20$ game have PAID CHARACTER CUSTOMIZATION??? in what world doesn't the "successor" to Garry's mod have any microtransactions in game? this hot steaming pile of ♥♥♥♥ will die within the year

Successor to Gmod? No, far from it. This is a ♥♥♥♥♥♥♥ mess. I was looking forward to dive into sandbox, but it is bare bones and no clear mod/workshop system. This is disgraceful to it's predecessor. It's like they didn't even try and just needed an excuse to monetize dumbass cosmetic items.

2 years of waiting for a stinking pile of ♥♥♥♥♥♥♥.

Wasnt gonna leave a review, but the devs decided to make a new announcement titled "Post Release". In this announcement they talk about the AI slop problem with the game except they dont actually believe it is a problem.

You can read the announcement for yourself if you want the exact details but in short it seems like they see no issue with these low effort games being all over the platform. This game/game platform had protentional but if its just gonna be plagued with soulless games it wont last long just go play roblox if you enjoy that stuff its even free.

maybe in a couple years this game will actually be playable but in it's current state this ♥♥♥♥ is ass from a butt and i blame rust players for giving garry newman more money than God

As of right now its fairly empty and doesnt run exceptionally well, however they are open and transparent about issues and have already shown work towards improving the quality of the game. Is it a 10/10 right now, no. More of a 4/10, its a good start, just let them do their thing and check back in a month or two, but disliking it into oblivion isnt going to help anybody, its $20, not $60 so dont expect $60 worth of content right off the bat. I can see the vision, it will be a 10/10, its just gonna take some time.

I paid $20 for this so I'm not biased. It's genuinely really good.

I haven't noticed many performance issues, most gamemodes work fine, the few that have had issues are very much on the developer of that mode, and not the game itself.

The actual modes are surprisingly good too, I was expecting AI slop and low effort stuff (of which there is) but there's actually a large amount of well crafted stuff here too. Just for example there's a COD zombies clone with a unique cartoon artstyle that's really well made (Ghoul Grounds), a plane arena game (Minifly) that's surprisingly addictive, a physics based combat dungeon-crawler (Dark Descent) and a few others I'll leave down below. My point being yes there is low effort AI slop here, but there's a lot of really cool stuff too.

For $20, I think it's pretty worth it, you're essentially getting a bunch of indie games bundled together, and your library will only keep expanding. They're not the best games ever of course, but they're good fun. To me this is a great game for those days when work has tired you out and you're not sure what you want to play, I felt like that today and this game just hit the perfect spot.

That being said, this is the type of game that lives or dies by it's community, if it's not popular enough talented developers probably won't waste time making cool stuff for it, so the future is very unknown, but as I said as it stands right now I think it's in a good place.

Other cool gamemodes:
Strike Force
Shoothouse
Blocks & Bullets
Creepy Donner: Night Shift
Terry's Granny
S&Coin
RagRoll Skater
Jumpy
Cut Them Down

The double down on AI + the terrible performance made me refund this game. Really big disappointment too since I was such a big fan of Gmod growing up. Please consider your next move face punch and maybe I'll repurchase it.

I think it's important for people to remember that S&Box is ultimately a platform for creating games. You are given an incredibly flexible toolset, pretty much the entirety of Source 2 toolkit that has been modified to be expansive to shape to your own creative will. Of course, in the first few days of the release, you're not going to find many games out there. Garry's stance on publishing games without taking royalties is incredibly generous too. You can publish your games on steam without paying any fees to them (aside from the obvious standard steam fee). Facepunch is pretty receptive to feedback too which is neat.

Some people are concerned with people polluting the market with "AI Slop" games, I understand and agree entirely. I wish it could be filtered and moderated, but ultimately that's up to us users to leave bad reviews on these published games. Unless they somehow figure out how to get this automatically curated but I can't see it happening - such automation could lead to draconian flaggings.

Ultimately, right now... Yep. The confusing UI, the layouts being a bit sketchy, the rough optimisation, the void of lack of games is going to be really coarse for a long while - maybe for the most part of this year. A rocky start is to be expected, much like how it was with their other titles.

All in all. Think of this as an expansive Source 2 SDK - not gmod2. Let them cook.

TL;DR: A disgrace to the name “Source 2,” standing behind it and barely resembling that amazing piece of technology.

A lot of generative AI content is seeded all throughout the platform, which is actively endorsed by some of the Facepunch developers on the s&box Discord. No biggie because you can ignore all of that, but just something to take note of.

Performance-wise, even on a good rig with the stock sandbox mode, it feels kind of sloppy compared to what you see in other Source 2 titles made by Valve, which run amazingly. The editor feels smoother, but that was just with an empty scene/minimal assets.

It’s kind of deceptively perceived as a “Garry’s Mod 2” by a lot of people in the community, which it’s clearly not. It’s not a replacement, and it’s not a direct successor, but it follows the basic principles of Garry’s Mod - it’s more like something, and sorry for the comparison, like Roblox. From what I’ve seen, you’re put into the C# .NET hugbox that’s wrapped around and above Source 2 here but see no actual Source 2 tooling except for Hammer.

In its current state, I’d stay away from it; give it some time to ripen. Either Facepunch drives this thing against a wall in the long run, or they actually use the potential it has and turn it into something amazing and polished. Facepunch appears to be pretty enthusiastic about the product, which is good, but it also really makes me hope Valve decides to give us an actual Source 2 SDK instead of relying entirely on this.