founder&developer @nooparmy Multiplayer, immersive AI, quality, philosophy, freedom, beauty, truth, family. With implicit ordering and out of order execution
Joined July 2010
- Tweets3.9K
- Following604
- Followers237
- Likes5.2K
Ashkan Saeidi Mazdeh retweeted
believing AI can't code or make games is a mistake
believing AI solved coding and game dev is also a mistake
It's weird how much biases people have against/for technologies. Game Engines are among many but still feels weird how wrong the reasons are that people hate/love Unreal/Godot. I don't see neither hate nor love toward Unity. I don't know if it is my timeline or what.
Ashkan Saeidi Mazdeh retweeted
An update on building games with Claude Code:
Last week some of you ran into Claude Code pushing back on game projects it should have just built. We'd added a short rule to the Opus 5.5 prompt to steer Claude away from recreating specific existing games, but it was also kicking in when you referenced one just to describe the mechanics you wanted.
We've updated that rule in 2.1.317. You can name a game to describe what you want and Claude should now build an original game in that genre. If you ask for a copy of the original, it'll offer to build something similar instead.
We want Claude to be as helpful as possible when you're building something original, but also to protect your work if someone else tries to copy it. We'll keep improving this, so let us know if Claude turns down work that's yours. Thanks to everyone who shared feedback!
We added two sentences to Claude Code's prompt on Opus 5.5 to steer it away from recreating someone else's specific game because using Claude to infringe IP is against our usage policy.
Original games, genre games, and games built from your own designs and assets are meant to keep working.
We'll continue to iterate to ensure we're striking the right balance. In the meantime, if Claude turns down something that's yours, send it my way.
Unity 7 should have done this too. Sometimes you have to start from scratch. Specially if invisible robots and matrix multiplications can do the time consuming part in much less time.
What we are trying to do: Build an engine that’s more powerful than UE5, easier to develop in than Unity, and is the most efficient effective way to build with AI acceleration.
The problem with having old systems and new systems side-by-side is the tendency to mix them, and then their interoperability becomes a giant tangle of blueprints and verse, actors and entities.
We thought that side-by-side was unavoidable, but AI assisted conversion is now clearly way more effective, and will produce a way more manageable UE6 project than a hybrid, while making it way easier for us to simplify and improve the engine.
We’ll be proving it out by converting Fortnite, the game with the largest base of blueprints in existence, to the new approach.
Ashkan Saeidi Mazdeh retweeted
me reading these news while working on @kaigen3d
It’s official: 5.9 is happening and it's coming in Preview next week.
It’ll include improved AI workflows, web as a target platform, rendering and animation updates, plus more.
We’re making a few changes in 5.9 branch over on GitHub, and more will come in over the next few days. More details coming soon!
It was so fast that it is unreal!! kudos to @TimSweeneyEpic for the leadership. I'm so happy I switched from Unity 3 years ago.
It’s official: 5.9 is happening and it's coming in Preview next week.
It’ll include improved AI workflows, web as a target platform, rendering and animation updates, plus more.
We’re making a few changes in 5.9 branch over on GitHub, and more will come in over the next few days. More details coming soon!
Ashkan Saeidi Mazdeh retweeted
Open-source "codes of conduct" are cancer and need to be retired; I deeply regret championing them. The inevitable outcome is an unproductive mob controlling projects through intimidation. Given enough time, it pushes away people who would contribute patches. Eventually, all that remains is an unproductive military-police class playing social status games.
We cleaned up a unity game for a client which among other issues, had performance issues due to the bad advice Jebrim disagrees with. Listen to him and save yourself money and time and headaches.
I disagree. Performance engineering is designing from day one around real-time performance guarantees such that you don’t ever have to do a deep dive in performance profiling later on to try to identify and hammer out each issue one at a time. Such an approach hits diminishing returns fast, dramatically increasing the amount of effort for ever small gains each time.
You ultimately run into an issue where your performance issues are death by a thousand little cuts and you need a full rewrite with an entirely new architecture and interface to achieve meaningful improvements. However, by then it’s too late as you’re locked into the existing interface that everyone has already built content upon and it’s your primary revenue stream.
All of which goes to say that being proactive about performance from day one is way more important than trying to be reactive to it after the fact, as you’re suggesting.
Also flame graphs only show detailed performance issues if you’ve architectured your codebase to be a bunch of many small functions, which is actually horrendously annoying and slow when trying to optimize an existing codebase. It also poorly handles capturing occasionally spikey performance.
Ashkan Saeidi Mazdeh retweeted
I disagree. Performance engineering is designing from day one around real-time performance guarantees such that you don’t ever have to do a deep dive in performance profiling later on to try to identify and hammer out each issue one at a time. Such an approach hits diminishing returns fast, dramatically increasing the amount of effort for ever small gains each time.
You ultimately run into an issue where your performance issues are death by a thousand little cuts and you need a full rewrite with an entirely new architecture and interface to achieve meaningful improvements. However, by then it’s too late as you’re locked into the existing interface that everyone has already built content upon and it’s your primary revenue stream.
All of which goes to say that being proactive about performance from day one is way more important than trying to be reactive to it after the fact, as you’re suggesting.
Also flame graphs only show detailed performance issues if you’ve architectured your codebase to be a bunch of many small functions, which is actually horrendously annoying and slow when trying to optimize an existing codebase. It also poorly handles capturing occasionally spikey performance.
i'm seeing so much performance slop in the timeline.
> "How can we use data-oriented design principles"
> "Can we turn this pointer into an index"
Perf eng is get CPU flamegraph, maybe some heap , disk , and network profiles.
Make biggest bar on these profiles smaller, if not move to the next biggest bar.
Ashkan Saeidi Mazdeh retweeted
I’ve never seen anyone be consistently as hateful and agressive as the anti AI people.
All of these comments are in response to me saying AI is just a tool. I’m literally being told to go kill myself.
Ashkan Saeidi Mazdeh retweeted
Join the club! It's a campaign of hate and intimidation, not aimed at convincing anyone of the merits of their position, but to bully and scare advocates into silence. They're spamming spammed with 1000's of "why doesn't someone kill this guy" memes.
Ashkan Saeidi Mazdeh retweeted
If you can just type and write a book, why aren’t we all established authors? The ease of making games doesn’t suddenly make us good at making them.
We modified several OSS libs with Ai and are starting to contribute back. #UnrealEngine BP graphs can get messed up by AI. We added MCP support to graph formatter and sent a PR to fab.com/listings/45218c70-1d… , my fork until he merges github.com/ashkan-saeedi-maz… 1/2
Buy the plugin or another plugin of his on Fab if you use it and like it. Also our stuff marketing.nooparmygames.com/… We got good AI plugins. AI makes everything better more quickly.
Previously having access to engine source code was directly useful to programmers, now it is valuable to everybody. Your Ai at minimum can learn from it, It can also fix it if need be.
It means potentially UE specific cheaper and local models have an easier time becoming a reality. Closed source engines will lose even more in the time of AI.
We’ve updated the Unreal Engine EULA so devs can now choose to train AI models on their UE and MetaHuman projects, and benefit from AI models that understand their code and content.
This change does not give Epic any rights to your content or code.
Check out the blog for the details: epic.gm/ue-eula-updates
Ashkan Saeidi Mazdeh retweeted
I published the first video, without audio, of Polysmith, just to show a subset of the available feature. If you are interested, I will add explanations and I will make more video. I have already used it in a product, but I wonder if there is a general interest for similar tools
Ashkan Saeidi Mazdeh retweeted
AI may well be able to write massive software projects “from scratch”, but that almost certainly isn’t the fastest / highest performance / highest security / most token efficient way to do things. Good building blocks matter.
In addition to RL’ing the agents on final task loss, we should try to get them to regularize the building blocks as well. What would the standard library look like that makes the agents most effective?
It is likely at the limit that it turns out to be some alien organization of code, but given the existing human-bias in training, I bet that in the near term it would be lovely code that old programmers would appreciate.
I suspect the same could be said for mathematics, but that is out of my lane.
Exactly. Probably file pilot with AI will have 100 times more features this year, even things which did not make sense to do before but were good. Those projects are abandoned and don't know what to do anyways.
I've been discovering new vibe coded file managers on a weekly basis lately and I've just gotten used to it.
Most are toy programs. A couple of things worth mentioning here:
- Vibe coding still won't give you top notch technical results. You can do better if you have the knowledge, even if you use AI to assist you.
- Programming != building a product as a whole.
- Designing a good UI and user experience is a whole different skill set.
- Building a good brand is another skill in itself.
- Creating and maintaining a good product website and presenting your product effectively.
- Vibe coding a toy project and putting it on github is very different from selling it to someone and taking responsibility for maintaining it.
- Marketing is again a whole different domain. Posting on social media, reddit, writing blog posts, creating showcase videos etc.
- Building and managing a community through discord, email or whatever platform you choose.
- Running your own company...
The list goes on.
Don't be discouraged by things like this. Put yourself in the user shoes. If you wanted a program that does job X really well, would you rather pay some money for a reputable product with an established community, reliable support, and a proven track record.
Or would you browse github for projects where you have to download the codebase and build it yourself?
And I'm talking about users who actually want to find something good, pay for it and get on with their work. Not those who complain on the internet about something costing money and not being open source. Those people would nag you even if you were on the toy project side of things.
So keep your head up and do what you're good at. Let others copy you. That's actually a sign you're doing something right.