Hello, I have been observing the progress, hesitations, changes and requests for Elements for a while now. I notice that “AI” productions are more frequent (this is normal since it is a future that is already here) and that external “appropriations” are also increasing. “Not having to reinvent the wheel every time” is the explanation. I wonder at this point if an AI tool directly integrated into Elements would not be useful to do all this?
The rest of this email will obviously give rise to controversy, it is not my intention so let’s try not to repeat what happened two weeks ago.
Let me explain: when the advances consist of copying and pasting what comes from AI or other sites, I can only wonder about the usefulness of Elements since I can do without it by pushing my requests to AI one step further. By having these resources available in Elements, Elements clearly becomes excellent and I no longer feel the need to resort to an external workflow with three AIs. Am I the only one who thinks this?
The basics need to be completed first before considering integrating AI, i.e. get Elements 1.0 out of the door, which I think is what @dan has already stated.
Using Cursor to generate code is by no way foolproof. It gets things wrong, multiple times. Every response I get from Cursor (OK, maybe a slight exaggeration) starts with “I apologise, you are correct …” when I pull it up over something that is not right. It also requires the user to be knowledgeable enough to be able to check and understand what is being produced by the AI tool.
I think embedding AI in Elements might be something to be considered once Apple Intelligence has bedded down and more functionality is released. I suspect that is some way off before it becomes really useful?
There is no way I would try and design, let alone maintain a website generated in its entirety from Cursor or any other AI tool at the moment. I haven’t got that much time to waste.
The only reason I am using Cursor is to generate php code is to provide extra functionality via Custom Components for Elements.
The average user of Elements probably won’t want to go anywhere near this because they (actually, I include myself as well) want an App that is easy to use to generate a website. I think Elements is well on the way to achieving that.
Hi @logrunner Thank you for your detailed response. Yes extra functionality, but I can’t see where Elements goes : no code, code with AI, no code + code + AI… I’m using Divi AI in Wordpress world for fun, well 117 minutes to set up a ready to use site for a psychologist online ; Homepage, services, contact, blog, ressources pages includes (with newsletter subscription of course). It’s not future, it’s here and I just made it for fun. I don’t like Wordpress for reasons I won’t explain here but despite my reluctance it’s here and it works ; it’s not the only one. I can understand a long term goal if I can Clearly see one. That’s one of the reason of this post I made.
I’m not too bothered if Elements doesn’t have built in AI. There are so many options out there I’m happy enough for now.
Apple Intelligence is also an unknown - we have no idea how Apple may integrate it into apps and the greater OS in the future. Maybe it will end up as a ‘Service’ and be available to anything, anywhere.
I’ve been dabbling a fair bit with ChatGPT, primarily because I don’t know HTML or CSS. I suspect what Bruno created in Wordpress would be quite doable when Elements reaches a release version.
It’s obviously going to take a while to polish everything but it’s already showing great promise and the custom components mean it should be quickly expanded by the community of users.
You are correct that Cursor’s AI (effectiveness depends on the AI model you use and your personal skill in using AI–they are not magic…well…if they are you must up level up your mage skills to use effectively). If only the AIs were perfect coder like human developers…oh…wait…
Right. Human developers make mistakes, too. I know that I do. But with AI I can literally feed an error back to it (or another model) and usually get a fix and then I move on. That takes seconds…a minute at most.
I think that Elements blending in AI and not depending on Apple’s annual release cycle is going to be critical for Elements success as a product and in the market. At least as long as the traditional website remains a thing–that is a topic for a different discussion.
While there are many possible options for using AI with Elements, some depend on what others might do and when they might do what is needed (looking at you Apple). I would suggest that an early Elements (pre or post first release is a business/marketing decision for Realmac to make) could do some things that are obvious today. Some examples:
There is a Theme studio. A Custom Component Studio could be added that uses either Realmac provided AI models (thus charging a monthly subscription like so many other services) or allow for users to use a model API.
An Elements Copilot similar to Microsoft Copilot for its productivity applications. Using RAG to access current Elements application usage information, it could be used to guide users through chatting with them specifically about using Elements, general best practices for designing websites, and perhaps even customizing copy or images for the website via user provided context (see next point).
Again, through RAG methods allow users to add context to the site through an additional pane in Settings that is similar to the Notes pane. However, this pane would be used by the Copilot for context regarding the website. For example, what kind of business is going to use this website, the mission of the company, its return policies, its physical address and other contact methods, etc. The GenAI would access this for writing copy or even suggesting the website pages (think Reloom sitemap automation except without the automatic page generation…at least until App Intents supports that kind of thing).
This is just a few quick ideas. There are many ways Elements can use GenAI today and I think it needs to be a priority but only after the foundational release model of Elements is released. Of course, that foundation needs anticipate these new features being added and should have a structure that allows for this as easily as possible.
I think that Elements could be that native, Mac-only website development tool that is successful in this short window where we still use actual websites as they are viewed today (before the mass adoption of AI agents like Knowledge Navigator). In fact, I think it could open up website development for even more people because of the assistance that GenAI provides. There is real value for Elements customers with effective GenAI application…this is not AI-washing.
Yes, AI, is here and for those that are willing to put in the time and effort, it can help to mitigate the time to do certain things. As always, integrating that ‘new thing’ into a work flow and understanding the output and what is and isn’t correct still takes experience and knowledge to get a good end result.
We are all still trying to figure out how to best use these new tools while keeping a firm balance measured against copyright and other ethical concerns. The overall trick will be how can we make it easier for the vast majority of users to build and maintain websites while still providing opportunities for the 3rd-party ecosystem that will grow around Elements.
I appreciate all the viewpoints as it certainly is helping me to think about the future. I think we are all doing the right things, just keep asking questions and giving viewpoints.
