I don't want the details

I don't want the details
4 minutes by Michael Heap

When something goes wrong, most teams focus on explaining why it happened. A clear explanation can actually make things worse, because once everyone agrees the situation was reasonable, the urgency to fix anything fades. Instead, ask what needs to change so the same failure is less likely next time. Good corrective actions fix the system, not the people.

Your AI Automated the Steps, Not the Workflow
sponsored by Progress Telerik

AI can handle individual development tasks while engineers still connect the steps by hand. Turn those steps into an agent-run workflow and you remove those handoffs, but also some of the human checkpoints hidden inside them. Learn how to set boundaries and approvals, assign ownership, and measure whether automation actually improves delivery speed and reliability.

Delegation and trust
7 minutes by Bjorn Roche

Effective delegation is built on trust, clarity, and shared ownership. Managers should clearly define goals, timelines, milestones, and escalation rules while making it safe to report problems early. Instead of controlling every decision, Bjorn suggests giving people ownership of how to achieve the goal. Good delegation reduces frustration, builds confidence, and helps people take responsibility for successful outcomes.

The overhang
10 minutes by Ethan Mollick

AI already has capabilities far beyond how most people use it. Ethan argues that people can get more value from AI by combining its abilities with four human strengths: deep knowledge, wide knowledge, taste, and agency. These help people guide AI, recognize good or bad results, explore new possibilities, and create work that neither humans nor AI could achieve alone.

Expected goals
10 minutes by Mike Fisher

Charles Reep spent decades counting passing moves in football and found most goals came from short sequences. But he missed a crucial step: checking how often longer sequences were even attempted. Business thinking makes the same mistake, studying winners without counting losers. Tools like expected goals fix this by grading decisions on what was knowable beforehand, not what happened. Write down your reasoning before results arrive, then grade it honestly afterward.

The problem with putting people in boxes
9 minutes by John Cutler

Personality tests and labels like "strategic" or "collaborative" compress complex behavior into tidy boxes, but the same traits can produce very different actions depending on the situation. What matters more than someone's type is understanding what conditions cause them to shift modes. Organizations also quietly treat established norms as neutral while expecting individuals to adapt to them, a burden that falls unevenly. Context shapes behavior more than character does.

Your MCP can’t do everything for you. Break it down.
sponsored by Hygraph

More engineering leaders are under pressure to scale content production without adding headcount. But connecting your CMS to an MCP isn’t the finish line. Try to run everything through one workflow, and your team will quickly blow up the context window. Learn the crawl, walk, run process that scaled one team to 750+ pieces of GEO content and a 30%+ lift in AI mentions.

industry

• If you don’t have the factories, you lose the expertise
• What Sun got wrong
• AI-generated posters don’t have to be horrible
• Samsung to double HBM4 output next year
• How Cloudflare enforces engineering standards using AI

And a few interesting models were released:

• GPT-6 Sol and Luna
• Claude Opus 5.5
• System One Models & Jev

security

• Gemini hacked three companies in first known breakout
• Supply chain security in coding agents
• Be alert: targeted attacks on prominent Rustaceans
• Hacking OpenAI

And the most popular article from the last issue was:

• The broken windows theory of coding agents