Design frameworks that let regulated teams ship AI-built software that passes review. eBay, Microsoft, Riot Games, Department of War. Free audit: link below.
I'm Jeff. I've run Cartisien, a design and engineering studio, since 2007.
What I post about: AI tools now ship screens faster than anyone can review them. The design rules live in someone's head or a 40-page PDF, so the product drifts.
I write about closing that gap with design rules your agents and engineers can actually check against.
If you lead product or engineering at a software company in a regulated space, follow along. And tell me: which screen does your team argue about most?
"What should the design system team build next?"
In most companies the honest answer is whatever the loudest person asked for in the last planning meeting. Nobody knows which components get used, which get detached in Figma, or which get quietly rebuilt in code.
Heads of product: what does your system team measure, and has it ever changed what they built next?
@datavendor_ai@seeklis Seems like a good idea till itโs not. Wait this is never a good idea.
You do realize how much people despise data brokers.
What a clear vc cash grab this is
๐จ #BREAKING: Bank of America, Chase, Well Fargo, & other major U.S. banks reveal rogue AI agents have been found โlivingโ in customerโs bank accounts.
@atmoio Seriously... right. Like when is a country going to just get crippled from an cyber attack. Security is no longer in legacy systems, that shits gonna be $$$$ to retro too
Although not sad for Getty, but for photography in general if this is a sign for the stock sites to come. Design has changed, advertising has changed and budgets smaller so its not a suprise, but wow.
Why do AI agent demos look so safe right up until the agent touches a real user's account?
Is the demo showing what the agent can do, or what the team can actually contain and review?
@themgmtconsult Thats the million dollar questions right.. I think there is going to be real channel fatigue coming soon. Most will find one main X, linkedin etc possibly smaller niche communities not infiltrated by AI bots and or 1:1 / face to face interactions.
What are you seeing?
"Why is the token called blue-dark-2 when it's green?"
Every new hire asks it in their first standup. It made sense when the brand was blue. Then marketing picked green, and the names started lying. Tokens named for their job, like action or muted text, survive a rebrand. Tokens named after a color don't.
Design leads covering two jobs: how many of your token names would still be true after a rebrand?
@akanksha_j20 Engagement farming has become the norm unfortunately for a lot of founders & instead of building a solid product they want to ascend to guru status
@sqs Mini-apps beat text whenever the answer is picking from a set. On a compliance tracker we built, the whole design question was how fast one person spots the wrong record, and no chat thread gets close to that
@theo dude agents are the most toxic interactions sometimes. Its like having a psycho coworker, you know their work is good but shit... talking them is like eating glass.
Looking to connect with people building in:
๐ง AI tools
๐จ Design
๐ Software
๐ค AI agents
โ๏ธ Automation
๐ฆ Product development
๐ Web apps
Drop what you're working on ๐