The latest V3 webinar was a bit unfocused. I still have lots of questions, about the software, about the company, but once I had a moment to piece together all the content, I still find myself intrigued. Here’s hoping I get to kick the tires on V3 soon.
The webinar software let us all down with multiple instances of frozen screen sharing, and the presenter could not seem to keep the narrative going in one direction for more than a minute, we did learn quite a bit.
Design improvements
Minimal User interface hides out of your way until you need it.
Improved media management, let the AI change things around, or edit directly.
Secure
Encrypted content creation (not even the Grid company can read what you are creating unless you invite them to a project)
Encrypted swarm – real time collaboration
Speed
WebGL based code is faster to display on all devices and much faster for the AI to juggle variations.
Convenient new integration services like Unsplash help you gather creative commons, copyright free images directly.
Social
Screen recording directly in app lets you share what you see, and trim the video before saving it to your page or the media library.
Real Time collaboration lets you work with friends and coworkers from all around the world at the same time.
Nerd/Developer friendly
An in app editor lets you create new hgml pages or alter parts of an AI created page you’d like to fine tune.
Hopefully, an updated API will be released to allow developers to create and evolve new experiences for the Grid and the Multiverse.
I’ve been working on this post for almost a year, I don’t want to bother with more screenshots or examples. v3 should be on its way shortly. I’m posting this now so I can compare my v2 thoughts vs v3. I did purchase a lifetime membership, so I haven’t given up on thegrid, fwiw.
To make my art-portfolio grid site present mostly the way I want, I’ve pretty much resorted to just importing Instagram posts.
The Backend User Interface is… clunky.It does look more polished than the beta version, and has some pretty animated menus, but it’s also less clear on how you and the AI work together to make a page.
!) The new site dashboard wastes a whole screen of real estate.
2) Post ordering does not place the latest post at the top, and while you can drag and drop items to move them around, I can see this easily becoming onerous with 3 dozen or so images or posts. You can actually zoom out the thumbnail views for easier drag and dropping, but there is no only text version to help you.
A little investigation has shown that the AI does not read the content or title of your posts, and no way to force thegrid to sort items on either the backend, or exposed as a feature for the front end. No way except to drag and drop the articles back in order.
Posts with images in them do tend to stay in the same order as listed.
Search
There is no search option for either the front or backend.
Navigation
There are more navigation varieties this time around, so that’s nice, but rarely surprising.
Some of the options are ones I would never use, so I have to keep hitting redesign until I get one that’s ok.
Typography
Typography is strange, sometimes fonts are larger when placed next to similar designed smaller fonts. This example is using 2 links to two separate pages (same thing, shown differently, beside itself)
Additionally, every page load/reload seems to suffer from a CSS Flash of Unstyled content, which means the text/page jumps around a bit until it figures out what font style/sizes should be.
Redesign Process
This whole process irks me. We’re asked a series of questions everytime we ask for a redesign.
The Layout: Do you like it?
(where 1-3 means redesign, and 4 and 5 mean…)
The color: Do you like it?
Choosing 1-2 here will prompt you to choose another color palette
Typography:
Molly.AI is supposed to be able to identify what parts of an image are important. Faces, skylines, or text text should not have text overtop of them, similar rules apply to cropping. This seems to be hit or miss. We used to have a quick key to check to see what part of the image the AI thought was the focus.
I’d be ok if you could at least click on the text overlay and see the original, or uncropped version.
Yes, you do have options to turn off the cropping, overlays and color manipulation. I want to see theGrid do things with the design that I wouldn’t have thought of, colors, layouts, navigation.
I’ve made a wishlist of things I want Molly.AI to be able to do, and put it on my Ghost In The Well grid site, but I’m going to copy it here as well (the flexible updating wishlist will stay on that site.
But, as of now here it is:
Regarding Function
Have a media manager, so I don’t need to upload the same image more than once. (confirmed for V3)
Iterate designs much faster, allow me to give input.(coming in v3?)
Auto pull in feeds of my other sites or social networks. (Instagram/twitter – perhaps into my drafts mode so I can curate later?)
Give Molly some initial direction in the purpose of my site. (maybe in v3)
Or, enter content and have Molly intuit, summarize and categorize what I may want the direction to be. (this feature may become the clustering idea and may be in v3)
Really surprise me
Design some crazy bad things
Design some crazy good things
Allow for tagging of posts
Have option to attach the user profile/sig to appropriate article.
Date and Time stamps available to be included on the posts.
Add Accessibility warnings for color/font choices.
Ability to pin certain articles or objects to the top of the homepage
Choice of different sizes for headers images/videos (both in app and on site) Currently wasting space ‘above the fold’
Ability to merge posts together
Usability:
Add settings/preferences for the admin experience. IE color theme/font sizes, animations on or off.
Add Revision history. Date/time/last edited by which user information.
Have you heard of The Grid, an AI that builds your website?
They have an impressive demo video, and a strong marketing strategy.
I made it into the beta early, and have been able to start poking around with both the documentation and the site builder apps.
I’m intrigued…
There are a lot of things going on under the hood. The system is analyzing my photos, adding color filters to match my branding (which was in turn created from the system analyzing a picture I chose, in the first place). Then it attempts to place title text in places that don’t overlap important parts of the images. You make a few other choices, layout, type formality and then the AI builds you a site.
It’s a bit of trial and error now, and the tech bits seem more mature than the user interface. That jives, the system does work, but it’s a bit difficult to figure out how to ‘make it go.’
It’s beta, things like the on-boarding of new users, tutorials and walk-throughs are pretty light on info.
but that’s part of the adventure, right?
There have provided a lot of API documentation, sites live on thegrid.ai and the files live github folders.
It’s not a magic bullet, someone will still need to make some choices about the direction of things, even when the AI has a better idea about what successful sites look like and operate.
But, I think it’s a step in the right direction, and every time the AI learns a new trick it goes through all sites and republishes them, making better choices.
That kind of thing will keep things fresh, evolving, and hopefully – eventually engaging.
It’s going to be a little while before truly professional/beautiful sites just ‘happen’ through the grid. I’m more than willing to wait to see where this takes web design and development.
Questions:
How do I sync my website/twitter feed/other RSS feeds to my grid site?
I see something in the documentation, for the API, about how to write ruby/python/ect to link feeds in, but this seems like a standard base requirement that many grid sites will be looking for.
Why doesn’t theGrid chrome extension or the IOS app use the application token authorization for apps I’ve already logged in to/authorized.
Good that it uses the twitter account from IOS, bad that all other accounts (g+. Facebook, ect) it sends me to a login screen even if I have an app open and logged in already.
Why does my grid extension have two share options when I right-click on something to share to the grid?
They both seem to work.
ok, a day after grabbing that, it’s working properly, glitch or fix, it’s better:
It seems like the company, like their AI, is iterating features/improvements daily.
Why is the ‘Review Design Changes’ page bright yellow?
New Crayola color: ‘Augh My Eyes! Yellow’
Every other page of the grid on the app or in the chrome extension is dark gray, soothing to the eyes and not a bright hazardous yellow background?
I shared a post that seems to mysteriously be shrinking with each iteration.
The fonts for those 2 posts are pretty much unreadable unless I look at the responsive phone sized version. This may be an @media size issue. If I make the browser window larger or smaller, the font sizes become readable again.
incredible shrinking content.
phone size is easier to read
What does the AI see as a ‘Successful’ design/iteration?
I’d love to see 6 possible designs that the AI discarded in favor of the one I get.
I’d love a mode where I can thumbs up or rate the results when I get them. I want to help train the system!
A feature I’m looking forward to is focused ‘purpose‘ options. The company has spoken about how choosing purpose categories like ‘sales‘, ‘newsletter sign ups‘, or ‘increase social exposure‘ will help influence how the system makes some design choices.
Could the Grid AI start to learn my personal preferences as well as successful designs over all?
We’ll find out as the year progresses. I’ll post a link to the sites I’m experimenting on when I have more time to play.
