Joe's State of Unreal 2026
If you've been tracking gamedev news over the last few days, you might have heard that the Unreal Engine's next version will be leaning fully into dead trends such as the "metaverse" and some kind of zombie version of NFT-style shared game items. It's also going to discard Blueprints, the visual scripting system which made Unreal so dominant in gamedev over the majority of my career. Tim Sweeney seems to have fully lost the plot.
These issues are summarized much more completely by Joe Wintergreen, an expert Unreal user who has published his own tools for the platform. Joe's overview of the recent announcements has both the white-hot anger of someone who's built their career on this stuff... and the clarity of someone who's built their career on this stuff.
This is existential threat level stuff for the engine, certainly for Unreal as a brand. It’s propelled Unreal into the Unity-a-few-years-ago realm of almost total uncertainty. People don’t know if it’s even safe to use it or train in it anymore. Here is the Actual State of Unreal as declared by me, someone saner than Tim Sweeney.
Game development as it currently exists today rests on the foundation of these big, free-to-learn, well-shepherded game engines. As of 2026, both of the two big ones - Unreal and Unity - are in insane-brain territory. Years ago, most of my friends were using Unity. Most of those people now use Unreal. Three weeks ago, I would have told you that in five years Unreal would probably still be dominant. I wouldn't say that today.
Discarding visual scripting is to me the craziest thing in here. Joe points out that many people have learned to code with Blueprints, but there are plenty of games out there which were only possible to make because Blueprints allowed dev team members with poor-to-intermediate coding skills to implement content through it. Teams are built, today, with staffing and allocation assumptions based on Blueprints.
Unreal seems to be assuming that in the future you would not need a visual scripting system because you would just tell an LLM to generate you some code. They seem to assume that in the future, a desperate engineer will run around cleaning up after the vibecoding everyone else is doing, and that nobody will need or want visual scripting to help them understand what is going on. (And also everyone will be making outfits for Fortnite? And all their code will be public?) Dire!!
Many of us in gamedev assume that most of Sweeney's dreams simply can't be realized in a world where LLMs cost what they actually cost to operate. Unreal 6 might currently be built on the assumption that everyone will be using Claude, but a lot of us aren't sure Claude will even exist when Unreal 6 is done rolling out.
My current two projects use Unity and a custom engine... so I'm not personally affected by any of this. It's fucking crazy to watch, though! Because it's gonna impact most of my friends!