That title says it all. There are two people I think of here: Tyler Cowen is eternally optimistic. He may say he doesn’t like something or he thinks something is “not the best it could be,” but he’s rarely “negative,” and never bitter. James Watters, despite a rough couple of decades for PaaS (his life’s work) is eternally optimistic. He only talks about positive things and potential, not bitterness about rival technologies.
Posts in "links"
What's on TV?
Meanwhile, humans are still making good Internet videos.
🔗 ✺ CHANNELVUE
$30bn of $6tn
‘“The availability of AI devices has also boosted overall spending by more than $30 billion,” Lovelock said. “With the replacement cycle unchanged, the stronger performance in 2025 will result in a lower relative growth rate for 2026, as demand has been pulled forward."’
🔗 Global IT spend to exceed $6 trillion in 2026
What's next in Sora
Lots of new features coming in Sora. Clearly not just a side project.
🔗 Coming in the next versions of Sora
A representative case of the oddly simple things that LLMs are bad at
Does it seem right that computer systems that use billions of dollars of hardware, electricity, and clean water need my help to add line numbers to a few kilobytes of plain text? No, it does not. But at least this gives me deterministic results for one part of the processing.
Recently, though, I’ve observed that Claude Skills are interestingly clever on the command line. Maybe agentic just means “can use Unix pipes.
Vibe coding toolchain and method
A lot of effort went into making this effortless.
🔗 Just Talk To It - the no-bs Way of Agentic Engineering, Peter Steinberger
Hockey stick on wheels
Great, if cynical phrase: hockey stick on wheels:
“Consider a team which presents their forecasts in the form of a hockey stick graph,” said Chen on Tuesday. “They come back the next year with their revised forecasts, and they are the same as last year’s forecast, just delayed one year. If you overlay this revised hockey stick forecast on top of the previous year’s forecast, it looks like what happened is that the hockey stick slid forward one year.
Anthropic is building a PaaS.
Anthropic is building a PaaS. They’ve got isolated VM’s/containers, they’ve thought out support languages for the major languages. You can even write code in the browser…with the AI. And it checks your stuff into GitHub? Plus, of course, it runs your code for you, at least with Claude Skills.
🔗 Claude Code for web - a new asynchronous coding agent from Anthropic
You need a proxy for enterprise AI
Google Cloud suggests using a centralized proxy to mediate all communication between clients and remote MCP servers. This proxy enforces access control, audit logging, secret policies, and secure transport, helping reduce the attack surface by having one enforced point rather than many decentralized servers. In addition, Google emphasizes identifying particular risk vectors like unauthorized tool exposure, session hijacking, and weak authentication, and treating identity, transport, and policy enforcement as foundational rather than optional.
Step one in using AI in enterprises (MCP or just plain access to inference): get a proxy. It feels like this will get all your KPI’s for the next 12 months done. Stop farting around, and put one in place.
🔗 Google Cloud Outlines Key Strategies for Securing Remote MCP Servers
Can't stop hitting yourself
The problems feel like something we should have moved passed long ago. The general problems and patterns are not new. Why do they persist?
🔗 Why Up to 70% of Platform Engineering Teams Fail to Deliver Impact