Introducing Unity 6, the next major release of Unity.

Source: https://twitter.com/unity/status/1725080342636192251



Unity 6 is slated for release towards the end of 2024. A specific date has not been announced yet.

Unity 6 beta feature highlights: https://forum.unity.com/threads/unity-6-beta-2023-3-feature-highlights.1532590/

 

No official thread about that?

 

Not that I can tell. Seems like socials people pulled the trigger a bit early. It was likely slated to be revealed at the Unite keynote in about 4 hours:

Or the youtube version is just a replay, and it has all already happened. I haven't paid attention to Unite schedule.

 

The keynote is long ended. On YouTube you will see the thing from recording. Hopefully not edited. Also no Roadmap talk for YouTube AFAIK. Only keynote.

Also people already are defending this performative BS. I guess Unity can't shoot themselves in the foot hard enough for some people.
https://forum.unity.com/threads/upd...or-software-terms.1513604/page-5#post-9474865

 

sounds to me like another desperate attempt to redemption from pricing changes.

Anyway, I'm confused more than ever. I guess Unity will have fewer versions of the engine to support.

 

I thought this was Unity 2024 but it's actually Unity 2023 LTS rebrand. Weird to say the least.

 

This explains the weird wording of all the revenue discussions that went "starting with the LTS release that comes in 2024" rather than just saying "2023.4 LTS". They were planning the rename.

This also seems to mean they can rejigger the release pipeline any way they want now. Rather than expecting a <year>.1 each year, they can just sit around with v6 for 18 months, then v7 for 29 months, then v8 for 43 months... just like they used to do.

 

100% the only reason they've done this is because they know they absolutely can't stick to the yearly naming convention anymore. What was the reason they gave for the switch to yearly anyway? Because I think they said that was to avoid confusion too.

 

I would be excited if I heard "We remove SRP and focus on one modular pipeline"....but atm nothing excites me there...we will see .

 

Its just 2023 LTS rebranded; nothing like the older 2, 3, 4 & 5 new versions which came with some big feature overhauls and additions and deprecations of some older stuff.

The skeptic in me says this is a marketing cover up of the out of sync release cycles (LTS shipping 6 months later than its named year) and also the runtime fee stuff (which is attached to this version). Also playing up dev-nostalgia for the old release cycles before JR became CEO.

I also hope Synty is getting a sizable cheque for the amount of times their demo projects are used by Unity.

 

Like I mentioned in the licensing thread that's precisely what this is, but like I also said it makes sense to abandon it too because it's a scheme that is intended to have releases come out the same year that they're labeled. It was just needlessly confusing.

 

I haven't touched Unity in over a year and seeing Unity return to the old naming scheme helps a bit. Now I hope they DO focus and improve the engine and improve the life of the developers who use it

 

If unreal wanted to be cheeky their next release could be Unreal 6.01

 

If the release cycle doesn't change, will we get Unity 7.1 before Unity 6 LTS?

 

This is great. Now we just need them to start going backwards with that number.

Unity 3.5 here we come!.

 

I think it will change. The .1 will be public alpha, the .2 will be public beta of the next major release. So I think there won't be 6.1 or 7.1. There will be 7 alpha (what would be 2024.1) and 7 beta (what would be 2024.2) and 7 LTS (what would be 2024 LTS).

 

Nah, they went from 2023 to 6, which means the next version will be -2011

 

Well that's still a step in the right direction..

The release notes:

• Removed Unity Hub.
• Removed Unity Package Manager.
• Made dark mode even darker.

 

When Unity would do it, this would be the original plan but the results:
- we removed the editor from around the Unity Package Manager, it was too bloated, the new one will enter public beta in 666.6 tech release.
- you can purchase assets directly from the UPM
- you can subscribe to our AI services directly from UPM
- since you won't have an editor, you will have free time to check out our constant sale on the asset store and subscriptions...

 

I wonder if this will mean anything or if it's simply just that - a naming change. Part of me would assume a change like this signifies a wish make considerable changes and implement major new features without having to worry about compatibility with previous versions. But at the same time, considering how everything seems to move into and revolve around packages nowadays, it feels like it's too late to pivot into such a strategy.

 

It's a really clunky update. You've got some rendering performance enhancements that are nice, but hardly revolutionary and could have been added 10 years ago in a minor version update without much fanfare.

You've got all this investment and hope in these new AI tools, that you want to put out there and see if maybe it's a lifeline to sustained revenue and the next big thing... but the more you understand the tools, and the massive headache of monetizing them, you realize they're a massive gamble and you're just opening pandora's box.

Outside AI tools are going to be better, cheaper, and easier to use. There is no future here for Unity, but they've likely invested tens of millions, and sunk cost fallacy plays out.

This reads as "ok, what do we throw in a minor update before we cut our losses on all the noise going on internally?"

Unity 6 is the barebones cut their losses state they want to build out from (hopefully).

 

unity 6 better go back to what made the engine good, unity have 1 chance to maybe win back some developers and if they mess this up, the engine is absolutely on a path to annihilation in the face of better competition either feature wise (unreal) or open source wise (Godot)

hell, many of us have found building our own engines more fun and less painful than using unity recently...

 

Don't know what all the bitching is about. I watched the keynote. The only thing I was underwhelmed by was the hand holding to make the AI "safe and inclusive". Every time I see that terminology bandied about it is all about lowest common denominator and restricting comedy and creativity and herding into approved opinion spaces. Will play with Muse and see if it can hold a style I generate. The Stable Diffusion apps I have used so far can generate some impressive one-offs but can't sustain a style or character's features to be able to construct a catalog of usable, coherent assets. We have been digging into URP and ShaderGraph and getting super impressive high frame rate results. I hadn't touched URP before this project because I did not have time to wrestle with the pink prior. It seems pretty mature now and some items I wanted for our VFX shaders have recently been added. I will close out by saying the rumours of Unity's coming demise are severely over-exaggerated.

 

So the turmoil the engine and community are in, is written off because you are enjoying using URP? Forgive me if ill need a little more than that after the last decade of using the engine being a downhill experience as far as usability and reliability are concerned

 

WTF - I just saw this announcement on Twitter and initially thought the poster had mocked up a fake post, then when I went to the source I thought someone was either spoofing Unity account or it had got hack! Then I realized it was real and had to double check the date was April 1st!

Well I guess I get to tell Unity 'I told you so' since when they first informed us ( via private channels) of the change to a yearly versioning system I said it was crap and only a matter of time before they ditched it and went back to pure numbers.

But lets not forget that these changes are only done for the benefit of Unity, not their customers. The change to yearly versioning was a blatant move to support the new subscription model, because it didn't work when major releases were 18-24 months apart. Why pay a sub for two years and get nothing new? Now we get a new license payment model and they change it back to number version just to suit them and the marketing!

This is what I said to Unity in an email from 13.12.2016

So they force everyone from perpetual licenses to subscriptions, claim using a yearly versioning number benefits developers ( it doesn't and didn't ) and now when they have to keep subscriptions AND charge a royalty fee they decide to change it back again, for their own benefit again!

I think I'm done with Unity.

My trust is gone. Its no longer a relationship based around me evangelizing Unity to people I know, no longer about me exploring new features and coming up with cool stuff, no longer about me investing my free time helping others in the forums. Going forward its purely a business relationship now. If i need Unity to do a project or work with my legacy client projects, fine. If I think I can create a new concept or project quickly in Unity, fine, but if it can be done elsewhere I'll probably do that.

I think at the heart of it, I just want to make cool S*** and not waste my time and effort dealing with stupid license changes, royalty payments, versioning schemes that change every half a decade - all that S*** is just dumb noise that gets in the way and makes me hate Unity as a product/company a little bit more every day.

The sad thing is I just realized since the debacle of the new payment/licensing scheme I've not even opened Unity. Well I think i opened it a couple of times to check something out, but just didn't have the interest to actually do anything so closed it again and I wasn't going to take on any new client work while things were so up in the air. Even now I just feel the shine has gone, and now with this blatant marketing switch, completely designed around the issue of the pricing model, well I guess my interest is at an all time low. Its not like I don't have other interests to keep me busy and those have just had far much more draw recently.

I haven't watched the whole keynote yet, so maybe there will be something in that to fire my interest back up, but a quick browse of the blog had zero points of interest for me, something that has been an ongoing issue with Unity for the last few years. I can see some stuff in the past being of interest to companies, but nothing has really got me excited on the developer level for a long time.

Such a sad end.

Edit:
Oh and can I just say how stupid this now is as someone with legacy projects dating back to Unity 3.5. That I now have to waste my time thinking about versioning instead of working. Having to have 3.5, 4.2, 5.2, 5.4, 5.6, 2018.4, 2019.3, 2022.3 and now 6.1 all on my machine. Plus how stupid I will sound to clients when i've talked them through going from Unity 5 to 2017 and now from 2022 to 6, and the extra time I will have to take to explain why this numbering scheme sounds like I made it up just to charge them more money.

 

Oh gawd, what happens with Unity defines now?

2015 - Unity 5.6 UNITY_5_6_OR_NEWER
2017 - Unity 2017 UNITY_2017_1_OR_NEWER
2023 - Unity 2023 UNITY_2023_1_OR_NEWER
2024 - Unity 6 UNITY_6_1_OR_NEWER

Which means UNITY_6_1_OR_NEWER is newer than UNITY_2017_1_OR_NEWER!

Put another way when I read UNITY_2017_1_OR_NEWER its actually going exclude Unity 5.6 but include Unity 6.x and up!

What a bloody mess.

 

i am slightly disappointed they picked version 6 as the number, as there was a 5 and a whole bunch of releases between, but a name is just a name, we can get over it, i just wish they had release more of a timeline about the things they talked of, and as someone in the webgl line, there was little for me - and URP hates me, light leaks everywhere..

 

If you can know a tool that generates HDRP materials in a easier way than what they showed, please tell me.

Regards to Unity 6, yes it's mostly a rebrand and a signal to the older better days which was needed. But at least it seems everyone is working one upcoming version, instead of an alpha, beta, and tech stream.

 

Ah yes, them sweet generic photorealistic materials there are a billion sites for. Writing a material importer for any of them is not hard, albeit a bit harder than chatting with an LLM. And you don't have to deal with Valve banning any use of AI for which you can't prove the ownership of the training dataset.

We don't know what they're doing. Unity 6 is just 2023.4LTS rebrand. It could very well be that the current development cycle remains, and it's just a rebrand in name only.

 

I DESPISED the package manager and I hope I can quickly make packages for distribution on the asset store. And now they have added Subsurface Scattering to URP, which is okay, but I wish it was just ONE UNIFIED RENDERING PIPELINE.

My current shaders are hdrp and I hope we can switch pipelines on the fly or at build time.

 

So, just to sum up the things:

- now Unity got more expensive
- now Unity is planning to deliver less under the same time period

And to cover it up, they change the versioning back to major.minor.

 

O Unity 6, Unity 6, wherefore art thou Unity 6?

‘Tis but thy name that is my enemy: O be some other name.
What’s in a name? That which we call a rose
By any other name would smell as sweet and have as many bugs;
So Unity 6 would, were it Unity 2024 call’d,
Retain that dear imperfection which it owes
Without that title. Unity 6, doff thy name,
And for that name, which is no part of thee... yeah whatever.

 

the 2023.3 package manager is 40% more evil than the 2022 one, which was pretty darn evil..

you forgot an adding a $30 a month sub for using an AI doodad

 

Unity's keynote in a nutshell:

 

I agree. Unity's keynote spent too much time chasing the current AI hype, and not enough time planning for future hype. Obviously 2023 was all about AI hype. Next year it will be something else. A keynote towards the end of the year should probably try to figure out what the next year's hype is going to be. Some of the AI stuff looked neat, but it was nothing amazing and cutting edge.

 

We saw a fraction of a fractal of the technical hurdles of AI content live onstage during this demonstration, and bear in mind this scenario was chosen because it was as simple as it gets. Generate a 2d texture, and STILL they run into technical hurdles.

When he generates the texture for the wall, it's tileable, but he quickly dismisses adjusting the tiling which would take 2 seconds in an effort to mask what might have been a texture that isn't large enough to scale.

They want to highlight the ease of use, but anyone who's made games understands that at some point you need to get these textures at the proper size, otherwise you sacrifice quality or memory. Beyond this, truly usable materials are complex shaders that use a variety of tricks with multiple textures of multiple sizes to deliver the most optimized and complete visuals. Newcomers using these tools will not be learning the fundamentals that will allow them to make games in the future, it's fool's gold all around.

Every texture, every material, every animation or arrangement of characters is unique to the scale of your scene, to the existing assets, and unique attributes of the object you're applying it to with infinite edge cases regarding positioning, collision, AI mechanics, etc. These are all things that current generation AI cannot account for, and will not for many, many years if ever.

The core problem is this AI does not make gamedev easier, it makes it infinitely more complex as you still have the original problem of making good content, but now you need to make tools that fix that problem which is so absurd and makes the already complex problem of making games exponentially more complex.

If they ever get these tools semi usable, they will only be able to produce a very cookie cutter experience limited to a small fraction of scenarios of what would be possible currently. It's a disaster in the making.

I was open to seeing some grand new paradigm in gamedev, something fresh and new with vision powered by AI, but nothing they showed shows a spark of innovation, or technical marvel. It's a low effort cash in that uses existing AI tech exactly how a layman would have expected it to be used. It's not impressive.

In 5 years, if they're lucky, Unity will be able to produce remedial games with no unique properties with unique AI generated art, there will be waves of clones of simple pong type games where a small group of enthusiasts will be able to swap out the art and a few animations and maybe adjust the timings of a few things, but the sheer number of edge cases is going to restrict this entire system to something painfully simple.

This AI venture by it's mere existence is a massive red flag that Unity is doomed for perpetual decline as its "potential" drains endless ambitions and funds year after year. It's very easy to see the amazing art generated by AI and wonder, "wow, what kind of games could AI make?" if you have a fundamental misunderstanding of AI and gamedev.

It took years to figure out how to get fingers to render properly, how long until AI can track down an elusive bug where zombie collision pops on and off randomly because of some exterior system misbehaving? That's the problem, it's still artificial intelligence, it can't iterate and track down problems, or approach complex situations with a strategy or plan.

AI can execute a defined task, and gamedev is the antithesis of that, gamedev is the process of creating new and exciting systems that cannot be brute forced with CPU power. These two paradigms are intrinsically at odds with one another.

I'm not trying to say AI isn't or won't be a powerful tool. AI will generate incredible art assets for 3d, and animation as well as 2d in the very near future, as well as generate stunning scenes and organic arrangements and interesting interactive dialogue. Content creation is going to go to the machines, but systems and design that push the envelope will still take the long man hours and iteration.

Ironically, it would be more ergonomic and organic coding tools that would allow AI to thrive in more areas of development, the exact same thing that would enable humans to better interact with game design. But for some reason Unity has been averse to improving ease of use as of late.

There are many things AI does INCREDIBLY well, but it's not everything, and persuing a prompt that you type into thinking you're going to get AI to make a full game shows a fundamental lack of understanding of the technology.

Game creation by prompt is a gimmick, it's a symptom of the marketing team having 100% control of this company for the past several years. It's where greed and lack of technical understanding intersect so hard that they disregard basic reality, shutting out any voices of sanity about what's feasible in our technical space.

 

Bogus interpretation. I lost a contract from Wolfram over the former CEO debacle. I have no love for Ricotello, nor did from the start. If your whole paradigm is holding a grudge then mine is just moving on and living my best life. I can't change others nor will I jump in the gutter for no good reason for a toolset that has put bacon and butter on my table for over a decade ( and frustrated the heck out of me a times). I just use any tool to the best of it and my combined abilities and workaround any snags..just like in every other career path I have taken.

 

It looked super barebones, like someones lunchbreak project. I'm not faulting the engineers, but exactly how much time was put into it?

Signal is all we been given, what we need is prof - not PR gibberish. I was sorta hoping someone would jump on stage as say - right, we know things are a little out of whack, but we got this sh*t - here's to plan...

 

Come to the dark side. We have MIDI support.

https://docs.unrealengine.com/5.3/en-US/midi-in-unreal-engine/

Joking aside I keep finding things in their docs that we've been asking for over here for years.

 

Substance Designer, Armorpaint, and any other number of procedural PBR generators will get you actually high quality results instead of the slop that Unity's AI offerings generate.

...what? No, not in the least. They've literally just renamed the engine. Tech streams will still exist. Betas will still exist. Alphas will still exist.

 

I mean, there probably doesn't exist another tool that gives as bad results as what they showed on the keynote.

 

My previous experience with Unreal was not optimal. My intricately hierarchied Harley chopper came in exploded into parts, Z pointed upwards and dragging any part into a scene crashed the application. I also did not like the deep invasion of my OS by a Chinese company. 42GB to get MIDI? At this point I can write an API of my own. The other rub is often my clients are not making games but applications using interactive 3D..medical diagnosis media viewers, Revit import and translation, voice driven AI command and control, automobile HUD and virtual dashboards, court case collaterals. Unreal is a set of sharp kitchen knives, whereas Unity is a Swiss army knife. One is great for prepping the big feast in a stocked kitchen whilst the other is what you would rather have in the woods.

 

Oh, it's worse than that now. I have just one release installed and it's sitting at almost 120GB. Most of that is for C++ but it's not like I'm going to try to use nothing but Blueprint.

 

I had an ok time trying unreal. But then i learnt c++ long before .net truly was a thing and without internet (cos i am old) so i am a little rusty but its still in there as i prefer c# by choice.
I had a hard time however with unreal. I found a tutorial and while i could have fumbled about i thought it would be a start. And it said go get this starter pack. So i did. Like you say the whole thing was a lot of gb. Anyway. So. Import. Poke about. A little blue print stuff. Never was into visual scripting. But ok. Ah. Make your own class. Ok. Edit type type. Oh you need to recompile outside unity now. Um. Oh ok. Chonk chonk c++ is a very slow compile. We have moaned about unity. But we have it good. Barf barf barf. Seems something in the starter park doesnt work right. As a like 1hr experience in of unreal. Really barf? Oh and now it wont start my project because its not compiled. So. It broke it. Marvellous. So. Using my c++ i fixed it (well actually i found the bits it didnt like and moved them to a land farfar away and tried again) result. Back in now. But it just seemed such a lot of downtime to compile when only 1 class not yet used by anything and of course as a newb experience very off putting. Sure unity can fart at you but you generally get to stay in the ide it might say im starting in safe mode but it starts.

i want my efforts on my code. Not feeling like unreal.exe is made specific for each project

 

For those interested in the AI gen crap, there exist free models, trained with CC0 textures. Find a stable diffusion client you like (on a mac I like diffusion bee) and install something like this model: https://huggingface.co/dream-textures/texture-diffusion (there exists an add on that plugs in blender that uses this). It is probably not too far off from what Unity did anyway, although It looks better than what Unity shows and it probably has less ethical and legal issues.

Or pay $30 to Unity per month for really sub par results.

Personally, I would invest in learning how to use something like armorpaint.

 

You only have to recompile outside of the editor in certain circumstances. For the most part that's just when you're making changes to an existing header file. Aside from that you can just activate the live coding recompile (Ctrl + Alt + F11) and be fine. I've been using it almost exclusively while learning how to work with UE5.

 

Those tools aren't AI driven which was point I was talking about, you can obviously getter better results without AI.

My point about the versions was there is currently is only one pre-release version available which wasn't the case on the old version scheme.

 

interesting i tried in i think it was 5.2 and it made me compile outside, i havent tried 5.3

 

My whole paradigm is not "holding a grudge mate", its making technical decisions based on facts presented to me and my own experiences since the 2014 shift in the company.

Right now the engine is where it is and the company is in the state it is in and trying to see the rainbows isnt going to change that. Im glad you have an optimistic outlook and nobody is trying to take that from you.

But your experiences with the engine != my experience or my teams, and for us it was a nightmare. Im still here because I am a diehard fanboy and I believe unity have the potential to bring people like me back by making the correct decisions going forwards. I am not trying to "hold a grudge" that makes it sound like everything has just been fixed and all is fine again. Its not. The firing of the CEO and changing of a version number, do not negate any of the last 9 going 10 years of BS.

I am sorry that does not align with your own version of events or your own experiences. We are in different countries working on different projects for different organisations under different terms - learn to see others views mate instead of shooting them down assuming yours is the only possible version that can exist. They exist alongside each other, yours does not negate mine.

 

You dont like punching yourself in the face?