Weeknotes vol. 17: business, schmizness
By Garrett,
Hello and happy casual weeknotes Friday.
I stopped writing these about a year ago when I began the transition into consulting (solving fun and challenging problems), and to say a lot has changed since then would be the understatement of the century.
In summary: Iain joined full time, we’re helping people solve operational problems and optimize their work across pretty much all aspects of business, and we’re having a lot of fun doing it. Iain has his masters in AI for Business, which has pushed me to go down the biggest rabbit hole I’ve been down since HTML/CSS in college (and we know where that led).
What have I been doing, specifically, you ask? Great question.
The answer is mostly ‘Claude’. Figuring out it’s (his?) inner workings and how all of the tools in the Claude toolbox work together and can be used to solve actual business problems.
What I love about Claude is the ability to route prompts through a filesystem of standardized folders, files and instructions using a set of rules and naming conventions, and BOOM, you’ve got an automated process. It’s not coding, it’s just writing text like a robot wants it written (which we’re in luck folks because I’ve been thinking like a robot since the day I was born).
I use Claude Code for this, but it’s not actually code at all. It’s just writing robotic instructions that say go here for this and there for that and branch off for this and use your little (big) Claude brain to figure out the details. Then, out pops the pop-tart!
Figure out how to do that at scale for lots of different cross-departmental things without making the system too complex and unmaintainable and you’re in business (that’s the part I’ve been focusing on).
While the fact that you can ‘just do stuff’ is great, it’s also making the fragmented work problem 100x worse for people who aren’t careful. We all have data all over the place (dozens+ of specialized Saas apps, for instance) and now we’ve got Frankenstiens of all of the above on every employees desktop (Thanks, Obama AI).
The future is figuring out how to normalize all of that data into a structured brain that can help you and your AI agents and skills and all of that jazz do your work together more efficiently and transparently.
What a world to live in!
Like this? Get email updates or grab the RSS feed like it’s 2008.
More from the blog
-
We have ways of making you pay
> The true cost of AI work is hard to measure; the value of AI work is also hard to measure, and metering changes which of those two blindnesses you notice first. It drags the cost into the light, itemised and arriving monthly, while the value stays diffuse, lagging and easy to argue about. That asymmetry is exactly why the panic is showing up now, ahead of any definitive verdict on whether the spending was worth it.Simon Willison did the arithmetic on himself. He pays $200 a month across his Anthropic and OpenAI consumer plans, and when he ran the [ccusage](https://github.com/ryoppippi/ccusa…
-
Bloated: how chat made you fat
> It helps to remember the time you save generating a document is not free. It is borrowed from every person who has to read it, at interest, and the longer the distribution list the worse the rate of return.The pitch for writing with a language model is that it saves you time: you describe the memo, the model produces it and 90 seconds later you have four pages (okay, maybe forty) instead of a blank document. Someone still has to read those pages though. The model did not remove that work. It just moved it downstream to your colleagues or suppliers, and on the way it produced more than any h…
-
Apple’s bicycle without a chain
Steve Jobs described the computer as a bicycle for the mind. Apple Intelligence so far is more like a bicycle with no chain. The frame is gorgeous, and the engineering is extraordinary, but you cannot get far with it.In early 2025, Xe Iaso published a [piece that landed like a brick through a window](https://xeiaso.net/blog/2025/squandered-holy-grail/) in the Apple developer community. The argument was simple and damning: Apple had built the holy grail of trusted compute with Private Cloud Compute, a genuinely unprecedented piece of security infrastructure, only to fill it with half-baked not…
-
The ten trillion dollar gamble
In November 2025, on stage at the Wall Street Journal’s Tech Live event, the chief financial officer of OpenAI was asked how her company planned to honor roughly $1.4 trillion in compute contracts on $13 billion of revenue. Sarah Friar said she was looking to assemble a network of banks, private equity, and a federal “backstop” or “guarantee.” By the following evening, she had posted to LinkedIn explaining that “backstop” had muddied the point, that what she meant was something more like a public-private partnership, and that the United States government has been “incredibly forward-leaning” …
-
Never talk about goblins
Buried in a JSON file that OpenAI posted to GitHub recently, inside the configuration for its newest coding agent, sits an instruction that reads like a footnote written by someone losing their composure. “Never talk about goblins, gremlins, raccoons, trolls, ogres, pigeons, or other animals or creatures unless it is absolutely and unambiguously relevant to the user’s query.” The line appears more than once. Whoever wrote it wanted to be sure the model understood.Most readers, including ones who follow AI closely, may be unaware of what a “base instruction” is, where it lives, or why anyone a…
