Unite 2023 Changed My Mind About Unity...
Вставка
- Опубліковано 7 чер 2024
- Unity brought myself and a number of other creators from the Insider Program out to Unite 2023 in Amsterdam. Here's what I thought of the event, and some of the most notable announcements they made.
Checkout the other creators who were also alongside me at Unite! See if you can spot them in the video! @iHeartGameDev, @CodeMonkeyUnity,@gamesplusjames, @gamefromscratch, @LanaLuxGames, @ChristinaCreatesGames, @BinaryImpact, @samyam, @TurboMakesGames, @DapperDinoCodingTutorials, @AIAdev, @MinionsArt, @acegikmo, @TheShelfman
💸Save 30% On @Blackthornprod's game development course with coupon code MATT30! www.gamedevrocket.com
Unity UI Toolkit Talk - • Create better Editor a...
--------------------------------------------------------------------------------
Want to support the channel?
▶️ Help fund new episodes by joining the Patreon - / gamedevguide
🎁Grab the latest Humble Bundles - www.humblebundle.com/?partner...
Use these links to grab some cool assets from the asset store:
Get the Must Have Assets! - assetstore.unity.com/top-asse...
Free Unity Assets! - assetstore.unity.com/top-asse...
New on the Asset Store! - assetstore.unity.com/top-asse...
Top Paid Asset Store Packages - assetstore.unity.com/top-asse...
Asset Store Partners - assetstore.unity.com/lists/as...
--------------------------------------------------------------------------------
Socials and Other Stuff:
• Subscribe - ua-cam.com/users/gamedevguide?...
• Join the Discord - / discord
• Twitter - / gamedevguideyt
• Facebook - / gamedevguideyt
• Instagram - / gamedevguideyt
Unity posted their UI Toolkit talk: ua-cam.com/video/HQ0TmO8ZA4o/v-deo.html
Writing good documentation ❌
Developing AI chat-bot that will hallucinate 20% of the time ✅✅✅✅
😂😂😂😂
more like 90% Generative AI is mostly a gimmick
If upcoming iterations of Unity become more performant than the current ones, then I hope they backtrack the runtime fees entirely.
A game engine shouldn't take upwards to 20 seconds to compile a single C# line.
I feel this. 😂
I thought I was the only one who suffers from a 20 seconds wait time after each script change lol
In UE You should Restart the engine whenever you change a line 🙂
Ong after trying godot I just cannot go back to unity. It is so so much faster. Why the hell do I have to wait 20seconds everytime I change a small thing and an additional 15s to enter play mode. Unity is extremely bloated
@@MellowKlug I'm not sure bloat is the neccesarily the main issue. Back around ~2017 things were pretty fast, and after that people started seeing the load/compile issue. After 2+ years of gaslighting the community saying the new visible load bars made things 'appear' slower but they were actually the same speed, they admited that it was actually a problem but because it had gone unrecognized for so long it was just impossible to find/fix.
One possible practical benefit of the improved version numbering is they might be more willing to let things cook, rather than constantly rushing things out the door. This is just a possibility, we'll see, but it occurs to me that naming after years may have pressured them to get more features released *that year*.
I’ve found myself feeling less and less impressed by generative AI as time goes by. Glad to hear someone else express the same feelings about it.
There is value in constraints for artistic endeavours and AI isn’t impressive in that regard.
I do see a useful future for AI as tools for artists if integrated into the artistic software instead of the current focus of replacing artists wholesale.
I’m cautiously hopeful that Unity will steer in the right direction under this leadership.
Yeah absolutely, there's just so many more useful things I can think of when it comes to AI assisted tooling. Generative AI is fine, and novel. The tech is interesting and useful for many different reasons BUT in my opinion it's such a waste of resources for a company like Unity to be spending their time
@@GameDevGuide LLMs are good at what they do, generating text, consistent auto-completion. Copilot works well for what it is, beats search most of the times and it reshapes your code to fit another situation, but one should never expect novel solutions from it.
How do you differentiate if it only assists or replaces you? The better it gets at assisting the sooner it will replace you.
Still waiting for Unity to update their terrain system.
You will be waiting forever. Why would they spend money on that when people buy 200euro terrain assets and make them bag
Glad you pointed out the problems with Muse. I was skeptical when I saw it's features but locking it behind a 30 dollar pay wall ensures it's failure
Yeah, it's just massively disappointing how Unity fails to strengthen the core value of their base subscription plan, especially with the runtime fee looming. Instead, spinning off all of these separate products and plans. It's absurd to me.
@@GameDevGuide IMO since they went public, their primary focus is pleasing board members and that's something that won't change. The best we can hope for is that the new changes work for both the board and the average user.
Absolutely reflects how I feel about it too
Generative AI as a Service cost is still hard to predict and difficult to make it profitable. A studio can spare 30$ a month for their lead artists (Photoshop Cloud license and other tools are already around the same price).
I'm a developer and the price ( in corporate context) isn't prohibitive
Ai can be a useful tool as i personally use it for helping make me understand but i see a may people use a script generator and not as a tool@@GameDevGuide
Where does that talk about UI Toolkit come from? I don't recall seeing that in any of the Unite videos
They gave me early access to it to include footage in the video, but it'll be up on Unity's UA-cam channel sometime in the near future - i think they just want to edit it and feed content out slowly rather than flood the channel with Unite videos - so I'll link it in the description here once it's live!
Putting Muse behind a paywall is a really stupid idea for how replacable it is. The Chat can be mostly replaced with ChatGPT or other chat LLMs.
As for texture generation you will get way better results with Stable Diffusion, which can be great for seamless textures, etc.
Good for Unity and all of us. I'm also positive about the future.
For me the challenge and the fun of coding/making a game is in all the roadblocks and finding solutions to overcome them. AI should be like a coding friend sitting by your side helping you along the way with some suggestions and ideas.
The ai sounds like it might come in handy for small indie developers who do not have a big team. For example if a developer consist 1 or 2 devs.
0:57 even he's not believing whatever bullshit written in his script
Good ole CEO non-speak talk. He said all those words which can be vaguely interpreted as "we love our community so much after our mistake, so community please love and forgive us for the sake of our shareholders". Not saying he isn't going to make a better CEO or that Unity isn't going to improve, but this speech in particular said nothing of significance in my opinion.
Yeah, I think having a skeptical attitude towards the situation is probably the right one. Having met the guy, and also having worked alongside JR in the past, I can definitely tell the difference between the two of them, hence my somewhat optimistic attitude but without that I can absolutely see this kind of attitude. As I said in the video, I think actions will speak louder than words here...
@@GameDevGuide Well, I hope to be proven wrong. My optimizing went out the window when they went public. They will try to squeeze money wherever they can, Runtime Fee and subscription based AI services is proof of that.
Feeding documentation to an LLM is a great idea. But most of Unity's documentation is already ingested into ChatGPT. It's crazy how useful GPT4 is. Anyone building a competitor platform or even a wrapper has quite the fight on their hands... Anyway, thanks for the video. I would love a deeper dive into your UE5 experimentation. Many Unit devs are interested in this topic.
I'll likely cover it in my next Cabs of Chaos update :)
Unreal is unreal. Even with Jetbrains Rider. Unity is really making all the efforts to make games development available for humans. No matter that I have 6+ years of experience with UE - everyday is a fight with compiler. And so the speed of development is discouraging. But the good thing is that stuff is there and the architecture is ready. UE has a fine dependency injection with it's types repflaction and assets/memory management is truly easy. But there will be enough drawback for everything speedup it gives.
Isn't the new CEO an interim? And that, the CEO may very well be changed in the future. If that's the case, I wonder if the promises he made really get carried over to the next person.
the game looks fun
"Unity will defend you in court if their AI tools get you into legal trouble"
Ok, Unity, and are you going to pay for all of my legal expenses and travel costs?
New CEO seems like the right choice and i hope he stays. I agree that these are just "words" and we will have to see.
I think with Unity 6 it could be a huge comeback if they focus on improving editor speed and the URP/HDRP coexistance that shouldve been a thing in 2018 lol.
I am REALLY hoping to see Dynamic APV at some point as a replacement for current unity Global Illumination
Totally agree about the AI stuff, it's a crazy distraction from them finishing other features. They should add it to the search on their manual/api website.
Saying the right things is a good start.
No, the pricing is about "Bankruptcy on succeed". I'd come back only if this model changes.
Unite 2023
Muse is meh, for animations will be worth it for one or few months subscription, but Sentis is a game changer, specially for mid-long term.
It's not about the person, it's about the culture; the word salad you chose to highlight as you noted are words... okay... but more than that they are meaningless words. I do not understand where he was starting or going with the words he strung together.
This will be good for people still watching this stuff and are interested in business to gauge how quickly a business like Unity can shift. I think it will take them another year and a half at best and that's only if they are singularly focused on going to a product/services company. Right now Unity is a acquisition and mergers company. As in their culture is no longer about getting developers tools to make games so they can make a profit but rather to gain revenue by any means possible.
The "humbled" interactions are simply corporate measuring the resilience of their clientele and given your sentiments it appears resilient enough, don't be shocked when in 2026 its like JR is back in spirit. =D I'd love to be wrong on this, but business is my background, and this is how things go.
5:17 and 5:19 were the highlights of this video! 😍
I can't trust a company that leaves critical features unfinished for so long. And so long as they continue with the package manager, nothing is going to change
i'm kinda disappointed that ai bot isn't free, but ok :(
In simple terms tools were supposed to empower, now everything is rip off pockets
The sheer fact that they keep "the fee" at all cost is reason #1 to not trust them anymore. This should have been wiped off the table to start fresh with the new CEO. Now it will always be that stupid leftover-thing "from those dark times". And once it has been established, it will bite people in the butt. It's still so vaguely defined, that this definition will be extended and exploited once there are enough people on the new version who will have a hard time backing out of that nonsense.
No one - ever - should use Unity 2023.3 / Unity 6 for as long as that stuff is in there.
Personally, the politics in this interest me less. If I have a project that fit Unity, I will use that. If Unreal is the one, then I use that, but if Godot fit my needs, I use that. Currently I use Godot for my current project but have used Unity before an I might use it again. If you think you will be a multi million dollar project, then the 4-5% you will pay of your revenue is fine. Of course, paying nothing is always appealing but if you compare Godot to Unity, it have it's quarks for sure. I see it more as a front-end library rather than an engine (and I like that) but you also loose out on a lot of stuff that comes build into both Unreal and Unity. So pick your poison. They all have good and bad things. As for Unity's Muse, ehm... that is 100% a tool pushed from cooperate with not enough funding. I doubt it will be anywhere mature enough to use for the next 5 years. Push more to the GPU is good but keep in mind, pushing stuff to the GPU means you need a GPU with a lot of memory. So it won't be great for people with anything less than a 4090 if you make semi complex scenes. Let's see how well it will handle out of core memory. It could end up a lot slower than running on the CPU. Just saying. All in all I think the Unite this year was disappointing. I worry about Unity's future.
As soon as I saw that they're still pushing generative AI garbage as a priority, it was obvious to me that nothing has changed. Same old trash, new coat of paint.
What limits Unity is bad architecture with a huge community stands on it. They cannot make dramatic changes, even insignificant ones comes with backwards incompatibility cost. They remind me Nokia. Just like them, their growth blinded them to the warning signs.
Generative AI is what now to NFT's?!?! Mate are you serious? Unity's implementation of it beyond chat...sure. But in general?! You're too smart to believe that.
There are so many more interesting and useful things AI algorithms can be used for, but instead companies like Unity are distracting themselves by chasing the latest trends instead of setting them.
In it's current form, the technology behind Generative AI is limited in its usefulness in general because it's so derivative (especially when it comes to Game Development) and when you consider there are tonnes of other alternatives out there all trying to do the same thing as well, it feels like a waste of resources to me.
As I mentioned here, I would love to see Unity trying to use AI to solve the more mundane and repetitive tasks devs have to do everyday rather than chasing the latest trend. Which is what it feels like Unity is doing here, because they fail to recognise ACTUALLY useful applications for AI to me.
Some people will find this stuff useful probably, but I don't see it being used by mid to large studios making games professionally whatsoever. Hence my comments.
he meant that its being used inauthentically as a marketing buzzword the same way NFTs were in a lot of spaces not long ago. Although generative tools like Muse and such may actually have some use (where NFTs hardly ever had any) the similarity when watching conferences preach their own shiny new techno-toys as The Future of the entire industry is quite obvious.
@@GameDevGuideYou completely lost me making that comparison. The AI that we have seen emerge the last year have changed how everything works. It can be an extremely competent tool for game developers and will only get better with time.
I'm not saying it's not useful, there are loads of really neat products out there doing cool things with Generative AI. What I'm saying is that because of this, Unity are chasing a trend where they shouldn't be, and that they could be using their resources for much more meaningful developments in the field of AI.
So, to me, it's disappointing to see them dedicate half of the keynote to it -- and combined with the additional costs (it's an entirely separate subscription price) I just don't see the mass adoption that unity are chasing here.
When they've outwardly expressed that they've been burning money the past few years, this to me is a prime example of why and how they're doing just that.
As I already said, and as I've said in the video, there are tonnes of really compelling AI tooling use-cases that I wish Unity would be dedicating their resources to instead of this Muse stuff which feels very lowest common denominator.
I did not like "new" numbering system. Programms novadays is a just a tool that evolve, year after year it is a ground truth. Having year naming is just a good convention, to understand what is new and what is old. It is robust and simple solution. Having number version is a just a marketing bulshit. Previous version are 5.6.1 new version is 6.0.0? Where did 2017 come from?
Unreal engine !!
lol he just deleted my comment
Guess he's afraid of constructive feedback 😬
Buddy how much did they give you to shill like this?
Did you watch the video?
Unreal engine !!
Unreal engine !!