🧙 CEO @DeepWorlds_SA 👤ex-military 🪄 Game Director @BeautifulLight

Black Mesa
Joined May 2012
Pinned Tweet
Guys, She said yes 🎉❤️
68
8
3
429
41,808
Probably the only thing that would reconciliate me with blueprint
zip-ties prototype
7
2
64
5,385
Scream if you love isometric rendering
25
39
2
1,178
23,903
Aherys retweeted
I’m genuinely thrilled to work with a team that’s so passionate about the game they’re making that they’re playing it during the week, late into the evenings and over the weekends. We’re all in right now, gathering feedback, experiencing things for ourselves and figuring out what we can improve. It’s incredibly exciting to see so many of you engaging with Hunger, and fantastic to be building it alongside people who care so much about what we’re making. Stay tuned for more comms in the coming days.
21
4
132
3,059
- How much space do you require for your telemetry server ? - Yes.
5
20
1,279
Ironically, i think the safest job with AI on long term is artist.
85
48
32
1,483
133,907
Quick thoughts, because there’s a lot of copium out there, both from people who use AI and from people who don’t. An AI learns from its training data. Everything it produces comes from patterns it found there. Ask an AI for a random number between 0 and 10 (try it): there’s a very high chance it says 7. Humans do exactly the same when they want to sound random, so the bias is already in the data. Then the fine tuning that makes models helpful and pleasant pushes them even harder toward the most expected answer. That’s also why AI art from a basic prompt so often looks the same. Part of it is a product choice: models are tuned to please as many people as possible. Work on the prompt and you get something different... but the same patterns keep showing up. To be fair, AI can produce genuinely new things. AlphaGo’s move 37 surprised every Go master, and AI has found faster algorithms than the ones humans wrote. But that works where there’s a clear score to optimize and millions of attempts to test. In art, stories or music, there’s no “right answer” to aim for, and that’s exactly where AI struggles to step outside what it has already seen. And let’s be honest, humans recombine existing ideas too. The difference is that we choose what to break, with intent, taste and lived experience. Right now, creative jobs are actually among the first to be hit: illustration, background music, voice work. So no, simply being “a creative” won’t protect anyone. What will stay hard to replace, unless there’s a real breakthrough in how AI works (Hello @ylecun 🇫🇷🥖, and dont hesitate to correct me if i say something wrong), is the ability to bring a genuinely fresh vision. The writers, artists and musicians who do that will probably be the ones who stand out. Of course, making predictions about something that evolves this fast is risky. But for now, if you follow the trends, that’s where we’re heading (for better and for worse).
4
2
55
9,202
"come well dressed please" Us :
3
1
49
1,333
POV ur game is planned for Q1 2027
Rockstar reportedly plans to release GTA 6 on PC during the first quarter of 2027. bit.ly/4ilf8FB
6
78
3,189
When you are good on network and code 😎
2
15
1,075
(sadly, claude was better than me on this one... and it's scary because yeah... i'm pretty good with network and code).
1
221
Game dev fun fact : More you optimise player performance, more it cost for your server. Do you know why ?😌
9
46
15,143
Answer : it's not on EVERY network system, but most of moderne game netcode work like that. You run your server at 30hz. A client run at 240hz. You send your data at 30hz. But client run many event, so for example, you do many thing between two net update, so you combine them in one, but still need to send the order, what happen, etc etc... the server receive them, combined, then execute them. It's necessary to have a good simulation on FPS, so someone at 240hz cost for example on average 4x someone at 30hz. It's totally normal, excepted, and something you want to simulate properly and at finest level your player on your server. On CS2 i believe this is not true anymore and cost are the same no matter who because they subtick on client, but that another story.
3
1
18
1,268
We've added staggers to some of the AI. Fighting them feels much more unpredictable, as you can get suprised when one suddenly gets back up!
9
25
4
463
14,486
The audacity.
Can someone point to where joe explains what the exploit was that got me banned in wardogs in this post? I'm trying to improve / not do it again, I am confused.
Readers added context they thought people might want to know
Katie's ban was for a cash exploit via coordinated FOB/recon structure builds and intentional captures/destruction with viewers for farming rewards, as her stream and devs confirm (not mere "too much money"). x.com/Brammflakes/st… x.com/onepeg/status/… x.com/GHBSmith/statu… wardogshub.gg/faq/exploit-ba… twitch.tv/videos/2894079…
12
814
Aherys retweeted
One of the things I don’t think players necessarily see much of is how the type of work you’re doing changes as you get closer to release. You might have a serious bug that only affects 1 in every 1000 players. During development and internal testing that’s something you might hardly ever see. But once you have 50k people playing that’s potentially 50 people hitting it. At 100k it’s 100 and at 1m players it’s 1k people. Obviously you don’t just have one of those issues. You might have dozens of different things that only happen in very specific circumstances or on particular hardware or when a strange combination of things happens. Individually they’re all “rare” but at scale they start to become a meaningful part of how people experience the game. The complexity of hunger is why we’re doing larger and larger testing waves. Every time we increase the number of players it exposes another layer of stuff that was basically invisible at smaller numbers. It’s also part of the fear of each new testing wave. You’re basically waiting to see if something suddenly appears that nobody on the team has ever seen before. Then when it does you have to work out how serious it actually is. How many people are hitting it. What the knock-on effects are. ultimately - whether it’s serious enough that you need to find the time to fix it before release. It means that as you get closer to release a lot of development naturally shifts away from the big interesting headline gameplay stuff and towards fixing hundreds of these smaller issues. Different people on the team are obviously doing different things but a huge amount of work goes into things that most players will hopefully never even know existed. It’s not particularly exciting or visible work but it’s necessary. Too many 1 in 1k problems left in the game and at scale the end result is one that feels buggy and unpolished.
9
5
1
86
3,586
Meanwhile, on steam.
1
18
1,095
Weekend sur le @SRGfestival 🥰
3
18
1,123
OZARK 1.0 OUT NOW ON STEAM 🔻 2.5D tactical action horror shooter 🔻 Finish it in a sitting or two 🔻 One copy = 2-player co-op via Steam Remote Play 🔻 Costs about the same as a box of donuts 🍩 🔻 Optimized for Steam Deck Play solo or grab a friend. It's time to descend!
17
81
12
300
35,370
Our friends at @ALTERBOYDOTCA are releasing OZΛRK! To celebrate the launch, we're giving away 3 keys right now, with 7 more dropping randomly Good luck, and don't forget to share clips if you got one! 👀 RL68E-93MBG-IAEIP 0E45H-EVACE-E0P6B MKA2L-2VEDI-DATXE
10
14
1
102
9,205