Why mid-level engineers are burning out under the cost of AI productivity and four changes to make before your best people leave.

Many leaders believe that diverse teams naturally produce better ideas and decisions. But difference creates value only when leaders can navigate the friction that accompanies different perspectives. Four common leadership traps—certainty, inconsistency between words and actions, emotional reactivity, and self-justification—can quietly erode trust, collaboration, and performance.

The metrics used to measure engineering productivity are obsolete in the AI era. Shifting from velocity to validation fixes roadmaps that keep slipping.

Research on more than 6,000 knowledge workers suggests that the most effective teams are defined less by office perks or work arrangements than by their ability to protect focused work. High-performing teams deliberately reduce interruptions through focus blocks, meeting-free periods, and systems that minimize constant messaging and check-ins. Rather than emphasizing constant interaction, the best workplaces recognize that collaboration depends on giving people the time and conditions to think deeply, solve problems independently, and contribute better ideas to the group.

Staying on top of new research can help your company make more informed decisions. This research roundup, drawn from the July-August magazine issue, offers insights into topics including middle managers’ impact on upskilling efforts, why intrinsically motivated employees are most at risk of burnout, and how experience influences customer loyalty.

Draw the prompts by hand — Prompt-Brush, the first non-AI generative art model.

AI for nostalgia's sake