Slack turns coding into a group project

Slack has introduced Slack Code, a new feature that transforms software development into a collaborative group project by embedding AI coding agents directly into shared messaging channels. Rather than competing to build the best individual AI agent, Slack is positioning itself as the central venue where human teams and AI workers collaborate together. The feature allows team members—engineers and non-engineers alike—to watch, direct, and steer AI agents as they write code, with live previews and human approval gates before deployment. All Slack plans include access at launch, with support for multiple AI agents including ChatGPT, Claude, Devin, Vercel, and GitHub.

Rowan's Corner addresses what makes candidates stand out to AI-first startups in 2026. Simply claiming to "use AI daily" is insufficient; top candidates demonstrate their AI workflows by sharing their screen without being asked and showing concrete examples. Candidates who identify real company problems and proactively submit solutions (via Loom videos, redesigns, or decks) advance quickly. The most important differentiator is "AI intuition" or taste—understanding precisely what unique value you bring that AI cannot replicate, and which workflows should be delegated to AI.

The newsletter includes a detailed guide on selling high-value AI workflow audits as a consultant, recommending a structured approach: use Claude to interview yourself about your expertise, build an AI assessment tool for prospects, analyze their responses to identify bottlenecks, and deliver findings in a client-ready deck.

Research from the University of Northampton demonstrates AI's value in medical analysis. Their CenSegNet platform analyzed breast tumor samples from 27 patients across 911 samples, discovering previously unseen patterns in centrosome abnormalities that correlate with patient outcomes and disease aggressiveness. This illustrates AI's less-publicized but significant capability to identify patterns in large datasets that humans and traditional systems miss, directly supporting diagnosis and personalized treatment development.

The community AI workflow highlights Eldercare Log, a tool built by Shawn using Claude Code to coordinate his father's medical care, appointments, medications, and vital sign trends. After his father's death, Shawn integrated lessons learned about estate management into the final product, demonstrating AI's practical value in managing complex family logistics.