Wastebook

• “We live in the age of ‘fuck around and find out’ - of iteration and experimentation.” 8Ball wisdom.

• “15 years ago, the internet was an escape from the real world. Now, the real world is an escape from the internet.” Noah Smith, via Ibid..

• “everything is optimized for engagement instead of meaning,” the Curmudgeon Era of life.

• If the likely outcome is the same, you might as well do a good job, so long as that’s fun too. If it’s not fun, do a bad job, if any at all.

• Never let a slide template tell you how to live your life.

• “The key to sales relationships, in my experience, is accountability. Take responsibility for mistakes. If you commit to something, do it. If you aren’t sure you can do something, don’t commit to it. This is harder than it looks.” James Dillard.

• “the death of the author and the return of pleasure” promises of Barthes-lore.

• “I blew off everything on Thursday and sat in a local place reading Ionesco with a glass of wine and an excellent coffee made with beans roasted in Naples.” Warren Ellis.

• I always want to know the “therefore”: What it means, what effect it has, how to think different, what to do next. Knowing the diagnosis is helpful, knowing what to do next is more important.

• “déformation professionnelle.” Found in Russel Davies' neck of the woods.

• And: “The Nobel laureate Alexis Carrel has observed that ‘[e]very specialist, owing to a well-known professional bias, believes that he understands the entire human being, while in reality he only grasps a tiny part of him.’”

• “that’s what might occlude our catsup.”

• “tweakments”

• “To read - and announce oneself as having read - literature in translation is to be tasteful and intelligent, a latter-day cosmopolitan in an age of blighted provincialism.” Bros of literary criticism.

• “obscurantism” Ibid.

• “I don’t think that knowing anything helps. I don’t think there is anything to know.” -Rick Ruben.

• Anything she says is correct. She just don’t talk much.

• “To under­stand the San Joaquin Valley, or any pro­duc­tive ag region, as “rural” misses the point. This is a vast, open-air fac­tory floor, totally wired up, care­fully monitored. I say that with appre­ci­a­tion bor­dering on awe.” Robin Sloan.

Relative to your interests

• Google’s Sergey Brin Asks Workers to Spend More Time In the Office - The 60 hour work week. // '“A number of folks work less than 60 hours and a small number put in the bare minimum to get by,” he wrote. “This last group is not only unproductive but also can be highly demoralizing to everyone else.” // These rich guys just can’t read the room. They seem to have failed to surround themselves with humane people and entered some bizarre land of productivity. Theory: they turned this hobby into their job and can’t imagine people who don’t do the same. They don’t even realize they did it. Cynical theory: they’re just crude capitalists. // Meanwhile:

• The Work from Home Divide: Insights from Six US Surveys - A study mapping WFH potential vs. actual adoption but sidestepping the bigger question of productivity.

• Build vs. Buy: Compare Your Kubernetes Platform Options - Don’t build your own platform. And especially don’t build your own Kubernetes-based platform. It’s not going to turn out well.

• Developers spend most of their time not coding - Developers spending something like 50% to 60% of their time on stuff that should be automated and built into the process. // And, now, we’re about to pile a bunch new stuff on them:

• AI Adoption: Why Businesses Struggle to Move from Development to Production - Day Two AI Operations: “This interchangeability means the real differentiators lie elsewhere: how you integrate your company data, design safety and guardrails, and adapt your development processes.”

• Model Context Protocol Bridges LLMs to the Apps They Need - The idea of having the AI sort out which tools to use is cool. What’s also cool is that you natural language to tell the AI what the tools do and when to use them. It then sorts them out. What’s also cool is that Spring AI is the official Java implementation.

• Model Context Protocol explained as simply as possible

• Affording your AI chatbot friends - ‘An “AI Agent” is just a model with access to tools like “escalate ticket”, “run SQL query”, or “draw an image”. The rest of the hype comes from fitting it into existing workloads like ETL nonsense with MuleSoft or something banal like that. This is really what all the hype is about: hooking AI models up to existing infrastructure so that they can do “useful things”.’

• Evolving From Pre-AI to Agentic AI Apps: A 4-Step Model

• The Big Shrink in LLMs - AI is shifting towards smaller, more efficient language models to address sustainability and data quality issues.

• Unfolding Time: The Medieval Pocket Calendar

• Build a Campaign-Unique Faction List - D&D stuff: “faction list turn our world’s lore to specific things the characters interact with during the game. Faction lists turns fuzzy concepts into a practical list we can use in the next game we run.”

• The Hobby’s Cult of Personality - Recap of the problematic stuff in the history of RPGs.

• Tyler Cowen, the man who wants to know everything - good if you like Cowen work, rather, the style of his work, his “production function.”

• Canadian Brewery Offers ‘Presidential Pack’

• Even Our Sex Scandals Are Sexless - Modern “sex scandals” often lack explicit sexual content, reflecting changing cultural and technological norms.

The agentic AI hype-cycle is nearly done

Good news, everyone. We can start doing “useful things” with agentic AI.

The hype cycle for agentic AI has been one the fastest ever. And I think we’re almost through hit without very many people actually have written agent AI code.

Let’s say Soaked McConaughey (Super Bowl, Feb 9th 2025) was the peak. And the beginning was sometime around October 2024 is when an NVIDIA blog asked “What is Agentic AI?”

Now, each thing I read about agentic AI, from a technical person, is all like “this is just hooking up genAI to your database.”

One sample:

An “AI Agent” is just a model with access to tools like “escalate ticket”, “run SQL query”, or “draw an image”. The rest of the hype comes from fitting it into existing workloads like ETL nonsense with MuleSoft or something banal like that. This is really what all the hype is about: hooking AI models up to existing infrastructure so that they can do “useful things.”

And MCP has made Richard Seroter’s list every week, sometimes daily, in the past three or so weeks, so we’re hopefully clinging up the trough to the Plateau of Productivity. Hopefully this means we’ll finally start doing “useful things” at scale in enterprises.

That said, the odd thing is I don’t think very many people have even written and deployed agentic AI apps yet.

(I put together a small timeline here, minus the Super Bowl part.)

Conferences

Events I’ll either be speaking at or just attending.

VMUG NL, Den Bosch, March 12th, speaking. SREday London, March 27th to 28th, speaking. Monki Gras, London, March 27th to 28th, speaking. CF Day US, Palo Alto, CA, May 14th. NDC Oslo, May 21st to 23rd, speaking. SREDay Cologne, June 12th, speaking.

Discounts: 10% off SREDay London with the code LDN10.

Logoff

I should re-learn programming. It’d make my marketing work a lot easier and truer. A programmer who can market is a good place to be.