Now AI is DESIGNING my entire app in minutes... 

Thanks! - As in basically the same concept as this video but more of it? As I mentioned in the video I think I could do a better job by investing some more time in learning Midjourney and spending longer on the design, so maybe I could do a similar follow up video later

@@JoshuaMorony ☝️💭👨‍🏫 what are some use cases for using openai to figure out your app requirements for your MVP if you are having trouble identifying your requirements? 🤝🤝🤝

@@JoshuaMorony Yeah, like design an app like a cookbook or flight tracker, something that ends up as a working sellable use case. Also another idea would be to do the same with a website, I saw some videos of amazing site designs via midjourney too. thanks!

and then when you give them what they want its not what they want

No you have to draw everything by hand or it's not art! -some pretentious artist scared of losing his job

I have no idea about ui and art but here is my opinion. You are currently absolutely correct. However this ai art has been a resent development. Give it a few years. As it is now ai art is not much more than a concept. When it is actually developed for specific use cases it can do much more. I would compare it to os and programs. Yes the os can do much but without the correct applications its quite tedious. With programs it can do so much more

Lets say the coding of the app is done with a use specific ai and not an allrounder. An ai just for this one language. Then you have an ai for ui elements, and just ui elements. With the combination of those 2 you have an ai that can make an app. (although there is more to it than just ui and some code) another ai is there for the last tweaks for the client and bam.

Moodboard 100%! Apps are more than just one state of an image. They’re a flow of specific screens, interactions, states, and possible animations. In general, a mobile app should follow best practices and patterns of an OS (iOS/Android) in order for the app to have somewhat of a successful UX and pass store acceptance criteria. AI has a while to go before it starts replacing people entirely. Side note: The apps used in the video are native wrapper apps which, honestly, are great for quick shits but are the minority when it comes to mobile app creation. They have performance issues at scale and are challenging to hire for. Xcode and AS12 are mainly used.

But for how long...

@@LikesGamePlays AI Art is currently created from already copyrighted material Give it a few months and all AIs including Midjourney will be sued to hell and back for copyright infringement. Additionally, What 12XFactor said is correct, this only should be used as a concept. The problem is many people wont and will then continue to make money with stuff thats taken from copyrighted material. I am all for AI-art mind you. I like the concept of it being used as a tool. But some of you really ought to be thinking both ways instead of just one.

Totally agree this is usable right now - maybe not for high end apps, but for fun/demo things certainly and I think applications for a lot of small-medium sized business/startups where something lacking a bit of polish and a more carefully designed UX is fine (if I wanted to put up some hobby/side project on the app store I would totally use this for the design). Trying to not get too carried away lol but I've got a lot of comments to the tune of "so what it made a todo app?" etc - to me it's still incredibly impressive and seems to be a sign of where things are headed. To me it kind of feels like mocking image generating AIs for including extra limbs or fingers on people - the fact we are at that point at all is mind blowing, and it seems like it won't take long for the extra limbs and awkwardness of UI designs to disappear.

except the AI already did improve. ChatGPT that was used for the app is from 2021. OpenAI claims their not-yet released generation is 100 times better (whatever it means)... I always tought I could do my job for another 20 years. As of now you still need humans to blend it all together, lets see for how long

@josh did create the app from a midjourney image. But seeing the result. I would consider it can be improved by using the same images the "bot" produced and upscale with AI, if you feel the images became bad. The foliage looks nice but it is possible to get the 100% design from midjourney into a functional app. Just need to tweak those CSS styles, gradients, box-shadows etc. Therefore I ask, if you can share the images or the prompt producing these images, to give it a shoot, anyhow great work on the app!

@@somerandomchannel382 great idea, and I definitely agree this can be improved with a bit more effort - I've added the original images from Midjourney into the "assets" folder in the source code if you'd like to tinker with it!

@@kraldada6557 ai is using open source and free content from internet to learn, several big sources for that already banned chatgtp so it will stagnate as it will have limited resources to learn from. Even chatgtp creators said it's nowhere close of replacing anyone, it's just another tool to speedup development.

Yep!

100%

Programming is exactly the same. It’s easy to convince yourself as a programmer that the minute technical details (built on decades and decades of increasingly bloated tech), just have to be that way, and you’re special because you know how to work with it despite the complexity. The truth is those technical quirks and complexity you mastered are going to be made irrelevant by AI, and you should be happy you can focus on the bigger picture.

Well you're onto something I believe. Half a decade ago I was working with corporate design. They had a small book full of specifications, what colours to use, what type of font, the size, exact measures for the layout. It was soul crushing work. And already back then I thought, why the fuck isn't there a machine for that?

I really hope you are a relatively new graphic designer or a student to have come to this realization just now. Otherwise I’m not sure how much value you were able to create with your work beforehand, must’ve been stressful.

Atm, yeah. Next iteration, these bots will understand language, and how to place them, font use.etc. But you'll still need creatives. What AI does is gets them to their end result much, much faster.

Chat gpt can write code. I use to get stuck on some bash scripts but using chat gpt has led to less errors and more productivity

@@joeolso It can not write the code as such. It can understand some concepts and re-create solutions or patterns. It can't grasp new concepts, workaround the issues, and in general apart from skeleton work it actually slows down more experienced programmers than is helpful. It is good for simple quick scripts or repetitive code used everywhere online (to-do lists, simple base structured API). It won't replace programmers - but will increase productivity as you mentioned.

@@HCforLife1 It can be used to accomplish everything you listed above if you break the prompts into steps. I literally did it yesterday. You cant say make this app but you can ask it how it would complete each step and write it proficiently. Im not saying itll replace humans, im saying it makes the talented faster with less errors.

@@HCforLife1 I think Chat GPT is currently gathering a lot of intelligence from people solving its code by raising issues. At least, that would be a nice intelligence on top of intelligence. I agree, it's current way of coding is filled with errors and problems when you tackle more complex problems.

@@Master_of_Chess_Shorts it's truly not error prone in that sense when you consider that the only reason why it can't produce exactly what you want is because you didn't explain yourself correctly or withheld information. If you asked a developer to create a complex piece of software you'd need to specify a lot of stuff to get the product you wanted. Same thing with Chat GPT

It's the iOS Simulator you get with Xcode - and if you're using Ionic/Capacitor you can also just run "ionic cap run ios" from the command line to launch the app in the simulator

From scratch

scratch. Mid journey is better than Dall E for producing high resolution images.

Yes, i think with this advancements the designers and programmers can merge into a profession that can do these two types of jobs much faster and at unison.

@@will8826 so less people in teams/1 man?

On the side note, that's incorrect on all parts imo. It's obviously a bit subjective, but no hybrid/cross-platform apps do not have general performance issues at scale. Again, kind of subjective but hiring for cross-platform is generally going to be easier because it utilises things like general web tech or React which is generally going to be easier to hire for than Objective-C/Swift/iOS developers. And as far as I know the majority of applications on the app store are hybrid/cross-platform not native iOS. On your first point, I mostly agree. I think for this to be useful right now it still requires the person using it to have an idea about these sorts of things, and even still the end result is going to be less professional than a "proper" design, but I can still see this fulfilling certain roles.

Lol and then another AI to write perfect prompts for that prompt generating AI right??

@@zwenkwiel816 And an AI to generate the code for that AI

@@Spidertooth108 yep, seriously though once AI starts creating and using other AI's we might be fucked XD

I think I asked it to use a solid/transparent background with no shadows when I generated the icon sets but it still added shadows and the bg wasn't a solid colour - though I suspect maybe people who know what they are doing with Midjourney prompts might be able to get it to actually use a solid colour background to make the assets easier to extract.

@@JoshuaMorony Inpainting and outpainting features would help here. This is where the user draws a mask over parts of the image and asks the AI to regenerate it. (Inpainitng for stuff within the image. Outpainting for stuff beyond the image boundary.) Iteration and feedback loops are also valuable. Take the generated image and use it as an input.

Ehh look at it like a tool you can use. Somebody still has to type to the Ai. If you can't google search well, you don't get good results.

It could kill ui. Not ux though. Ux pays more anyways

Nope. It's a tool that will speed up the research/ideation process (And maybe kill graphic design & Illustration) and make devs less independent from designers. But a digital product is more than just a creative concept of 2 App screens... Just compare the concept and the prototype result and you will notice that the final product doesn't look as pleasant as the idea.

I'm not fully up to speed with the methodology for the prompts, but I'm pretty sure this is to get it to produce designs similar in style to those you would find showcased on the dribbble website

Lol. Just take a look at the concept and then at the prototype he build... Let's ignore UX for a moment but just based on the UI, does the prototype look worse or better than the concept he used as reference? ;)

Ai is dangerous but if you make for fun then its ok

You don’t have knowledge based job. Your Walmart job will definitely be took over. You can start worrying to

@@cabinmusikslaps as a developer, i can assure you, this ai is capable of spitting out new ideas (albeit primitive at the moment); in the future it is not going to replace ONLY knowledge based jobs ;)

I doubt it, you must have not seen the looting after George Floyd's death. Service jobs are safe for awhile

There's one major flaw in your logic. The customers of AI art are not customers of Human Art. They're different customer segments. It's the same way the Record Company Industry believes Piracy is Stealing. Which it isn't. It's copying. There's no revenue loss, because if Pirate Bay didn't exist, they wouldn't buy it in the first place. So while you may think AI is devaluing art. Human art and AI art will be two different products. Similar to digital art and physical art. Conclusion: AI is opening up more people to art. Thus making the world more beautiful and less evil. Which will benefit a world on the edge of a WW3 nuclear war. I'm not sure where you live in the world, but I bought Nuclear Radiation Pills this year, because our government in Europe was anticipating a nuclear explosion and it gives you 99% more chance to survive while eating radioactive food. The world is in chaos, and AI has given us generated art and automations to distract us from reality. That's huge W in my book 😄

When everything is automated, capitalism will obsolete. The next social reform, will not be pretty tho

@@iliaadamanthark8336 Doesn't it depend on the reform if its outcome is pretty or not? What do you expect to be not so pretty? Isn't not reforming anything going to lead to some not so pretty outcome? Right now, there is an (understandable) incentive to fight off improvements in productivity as incomes are linked to labour - the needed amount of which can be done by fewer and fewer people and leaves more people without income as a consequence. So we have to either find working ways to distribute that labour more equally or decouple labour and income to a degree.

@@midnattsol6207 I mean, the process towards the goal/reform, will not be pretty. It's power play, between the super rich vs everybody else. I think at the end of it, humanity will dwindle into a very small numbers of group. That will be good for the earth tbh, but won't be pretty for everyone affected i.e. The normal Joe like us.

@@iliaadamanthark8336 oh ok, i understand. That power struggle defnitly (already) is a thing, yes.

@@midnattsol6207 Yeah, and now it would be exponentially worsen, as the AI grows exponentially. Hope we will be allright tho. Here some AI generated melody for some inspiration ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-Emidxpkyk6o.html

I don't have any basis to think whether or not AI will replace these professions entirely eventually (say within our lifetimes), but to me it seems very reasonable to say that it has the potential (tools perhaps marginally better than what we have today) to replace a lot of peoples jobs. Maybe a small dev agency can now use AI + 2 devs to complete all their work, rather than having 5 or 10 devs - if the demand for software as a whole doesn't grow then that is a significant amount of programmers (or designers) that have been replaced.

Well unfortunately I can easily imagine AI, if it will be even slightly better than now, to be able to just spit out the entire code for a program/app you wanted to create. AI is starting to "think" like humans, so what is the thought process of a programmer, who wants to design and make an app? Well first you have to name what the app is going to do, then you list how it is going to do it. Then you need to figure out what code to use in order for the app to do every single thing that it is going to do. one by one, and piece it together. Last part is checking for errors, and checking if the intended funcionality works and looks just like you wanted, and if not, improve and repair. If the AI will be able to understand the intent behind the complicated prompt consisting of the app description, it's look and functionality, I can imagine it will be able to design and code the entire thing in seconds from A to Z, and it looks like it is already not far from that. Even if it won't be perfect, you will be able to order the AI to change the code in specific parts and thus perfect the code to your liking. I really feel like we are not that far from AI with such capabilities. You can also think obout it this way: Humans learn to speak a language, humans invented computers, which have their own language. Humans can't understand raw computers language, so they invented a language (programming) that both humans and computers can understand in order to communicate with computers. Computers do not understand humans language, so humans teach computers their language by using a language they both can understand - programming. Once computers will understand humans language fully, there will be no limitations on what computers can do after speaking with a human, because they will essentially "think" billions of times faster than us. It will no longer be a matter of designing complicated programs and models to make something (no matter easy or superhuman hard) with a computer, it will be simply like asking your all-knowing friend to do something for you, and AI will, like a human, find a way to solve your problem. It does not have to be AGI, it only needs to understand your intent behind your words, so, contextual thinking is the only step they need to teach it to archievie it.

Uhhh trust me, thousands of copywriters have already been replaced. You know how many people work with writing filler content for SEO? All of them. Gone.

I also think this color pallete is just not good at all. It is really ugly tbh.

@@sentionmusic so tell it to use a different one... You're the designer here aren't you?

​@@zwenkwiel816 Now that I'm zoomed in paying attention, the UX is terrible too. Hit me up in 5 years. This app is novice.

Another showcase how a dev is turning a creative and pleasant UI concep into a generic, boring and clunky prototype. He basically just used the colors and images, but ditched everything that made the original concept aesthetic. Midjourney as it currently is, is a great tool tho. Just like Wordpress, Webflow oder other tools. It may kill graphic design but there is a bit more tto UX/UI design that just images and colors^^

Yes on their bottom tier I think it is $10/mo - they also do have a free option I think where images are licensed under Creative Commons (would have to double check that though)

What if I told you that every human form of expression ever has been built on the totality of prior human expression and experience? Nothing is original. Everything is derivative. No one has ever done anything on their own.

@@ReesesCupOfficial that's a tight slap🤣

@@ReesesCupOfficial The main point I am making is that the potential for copyright infringement exists and is based on where they are getting the information they are using to train the AI. Another thing is the original ideas behind the many visualizations had to stem from somewhere because they did not just spring up from thin air! I do agree that inspiration derives from building upon what has come before more often than not!

@@mind_of_a_darkhorse This is bullshit. For the following reasons: 1. Copyright is absurd, no one should own intellectual standards, it's an incredibly dystopian notion in itself 2. Artificial intelligences just consume the content and try to learn from it, not much different from what humans do. Artists lay eyes on something and obviously draw inspiration from that thing, it doesn't matter if it was "copyright material" or whatever, and then they create "their own art", if you say we should pay for just seeing one thing and use it as training material then you would effectively be advocating a dystopia, because firstly it is impossible to charge to see something or to train with it, and secondly because creativity should not be limited by any law. Artificial intelligences learn about the contents and then "re-create" them, they don't actually copy or "blend" anything, because that would mean that we literally store the images inside them and then use some merge function, which is absolutely far from truth. They work through a process of diffusion and regeneration, their neural networks acquire information about the patterns contained in the images through trial and error processes and they manage to generate results through various mathematical functions, they are effectively "new" results in the sense that all that exists in a neural network is a series of millions of neurons and millions or billions of connections between its various layers, there is no "image storage" there, only storage of the patterns that were understood by the neural network, say that this is somehow "blending human contents" is nonsense.

@@diadetediotedio6918 I hope the law and lawyers agree with you! But anywhere there is money to be made will always draw the capitalistic vultures!

Thanks for the comment! Yes I agree this doesn't replace the skills of a designer, but I also think a lot of lower-mid tier range apps aren't using professional designers to create a carefully thought out design/UX. A lot of these applications do just use things like general purpose themes from marketplaces, or just use reasonably standard patterns and don't worry too much about the design. This seems to be a good way to make those sorts of apps a bit "prettier" where there isn't budget for a designer. And yes, I just generated a more or less random icon set and used that here, but I could certainly put some more effort it in here and get Midjourney to design icons for the specific situation (e.g. I could create one icon at a time giving specific instructions as to what it should be).

It's more like prototyping an downgraded low budget version of a nice concept.

It is useable though, that was the point of the video - I used the generated concepts to create an actual app. You could debate whether what I created was crap or not, but it is a useable app.

@@JoshuaMorony I mean this concepts generated by AI are far from real business tasks and app functionality and real userflows or scenarios Sorry for misunderstanding, you did great job 👍

@@JoshuaMorony it is not usable because you can't use or steal designs from dribble for your app.

plenty of trash designs made by "actual designers"

you are missing all the points here

@@minhuang8848 I thought the point was that AI can design an entire app in minutes. Which yes, it can, just badly, and then you still need to put it together yourself, so minutes still turn to hours of time spent on a design that's not great. what did I miss?

​@@Hazarth You missed that this enables people without any design experience whatsoever put together themes that would otherwise take them months It doesn't matter that this might not be a perfect UX straight out of the box, that's something we'll quickly be able to abstract away. Just consider how far temporal coherence has come with chatgpt and the next iteration will give you thorough pointers on basic design specs you might want to consider - which even this example still way, way overkills. Maybe it's missing some aspects, but even so it'd still outpace every single billion-dollar company's UX design endeavors, regardless of backend considerations. Yeah duh you still have to put it together, but the sheer amount of work you save by not hiring a concept artist (or, lmao, doing it yourself) alone is off the charts and more than indicative of a trend. Noisy my ass, people already care little about good design, basic UX is good enough if it accomplishes your task. This stuff looks like a million bucks in comparison, and that's just a preliminary attempt at putting together something with the earliest tools. Picking it apart at the seams when it really is more of a proof of concept is just silly.

@@minhuang8848 Are you a manager at some company? You talk exactly like a manager: Nothing matters, but we make money, so it's good

no, am waiting for skynet, been out of time for too long

A designer wrote this

@@JC-jz6rx 100%