The real cost, security, and culture problems behind enterprise AI agents
THE SO WHAT
Agentic AI is stalling less on model quality and more on org friction — security, data access, and ownership of failure modes. If your pilots are stuck, treat agents as a cross-functional product with budget, SRE, and risk owners, not a sidecar to the data science team.
READ THE SOURCE
MORE FROM THE WIRE
Applied AIMeta rolls out Muse, a new AI image generator
Meta is turning its consumer surface area into a default creative engine—Muse is less about art and more about compressing ad and content production costs. Brand and performance teams should assume creative iteration speed is now a commodity and shift effort to testing, targeting, and rights management.
Applied AIAnthropic plans to lease a 16-story building in Lower Manhattan and double its NYC workforce to 1,000 people this year, as AI companies expand in the city
Anthropic doubling to 1,000 people in NYC this year is a clear bet that frontier AI R&D and go-to-market can be run from dense urban hubs, not just Bay Area campuses. For operators, it means New York is now a primary talent and partnership market for AI—not a satellite.
Applied AIMeta Now Lets Anyone Use Your Instagram Photos in AI Images—Unless You Opt Out
Meta is treating public Instagram content as default training and generation fuel—shifting the burden to users to opt out. Any product training on user data needs a clear consent story now, or you’re inviting the same backlash and regulatory scrutiny.
Applied AIDiscord Goes Ban-Happy, Suspends People for Benign Images
Over-aggressive auto-moderation is the dark side of scaling trust and safety with ML. If your product leans on automated enforcement, you need a fast human appeal path and clear comms — or you’ll burn your power users while trying to protect them.