Relative to your interests

• MCP Authorization in practice with Spring AI and OAuth2 - Filling in the missing piece of MCP: security.

• The Battle For Grounding Your AI Agents Has Begun - Data gravity considerations for AI.

• We did the math on AI’s energy footprint. Here’s the story you haven’t heard. - It’s still hard to know how much energy AI uses…until the big AI companies start telling us.

• Minimum Viable Humans. - “it’s only possible with algorithmic management taking over functions previously handled by human managers: performance monitoring, task allocation, basic feedback and guidance, coordination and information flow. On the positive side, this should translate into a sunset for the endless run of pointless meetings. Good riddance.”

• After months of coding with LLMs, I’m going back to using my brain - “I’m leveraging them to learn Go, to upskill myself. And then I apply this new knowledge when I code.” And: “But I’m not asking it to write new things from scratch, to come up with ideas or to write a whole new plan. I’m writing the plan. I’m the senior dev. The LLM is the assistant.”

• morning computer sociomediapath - As ever, the way to improve productivity is stop interrupting people, be they writers, programmers, or any “make something” type.

• Generation Z leading shift toward GOP, survey shows - ”voters aged 18 to 21 now favor Republicans by 11.7 points, challenging the common perception of Gen Z as ‘uniformly progressive.’" And: more updates on The Kids.

• Reports of Deno’s Demise Have Been Greatly Exaggerated - ”Most developers weren’t deploying simple stateless functions. They were building full-stack apps: apps that talk to a database, that almost always is located in a single region.” // People love the CRUD app. // Also, a tales from PaaS-land.

• Proof (again!) of why enterprises need to focus on AI strategic value over cost-cutting - “Forrester believes that a maximum of 1% of core business processes will be orchestrated by generative AI this year. A core issue is trust, since these systems tend to hallucinate, and it’s hard to troubleshoot bias.”

• The new economics of enterprise technology in an AI world - This is a strange collage of IT project success and failure. I think it’s saying that if your IT projects don’t show legible business improvement. They’re considered a failure. // Also, you have to look at an IT project as a big system, not just one point in time like the developers shipping an app. // “This pattern is clear in AI initiatives, where only 1 percent of company executives describe their gen AI rollouts as “mature”10 and only 10 to 20 percent of isolated AI experiments in the past two years scaled to create value. Our analysis of the impact of FinOps programs reveals that this misalignment of incentives leads to poor spend decisions on enterprise technology and results in a 20 to 30 percent loss of value.” // And commentary on IT projects in general, namely, it’s hard to get perceived ROI on them and stick to original (incorrect) ongoing budget estimates. On the other hand, would the company survive without them? Analogously, what is the ROI on electricity? // Also: “Some 10 to 20 percent of productivity resulting from work-from-home investments benefits employees rather than the enterprise (such as improved working conditions and freed-up time for personal activities).” Employees are better off, but who cares if it doesn’t make the shareholder more money?

Wastebook

• “Catastrophic forgetting.”

• “Parking lot with a taco hole.” RotL #580.

Logoff

I’m done with another talk, this time at NDC Oslo. I’ll be at SREDay Cologne week after next, then not too much for a while.