A collection of work-related content produced between November 2024 and July 2025 pertains to technology discussions, particularly focusing on AI, platform engineering, and private cloud solutions.

In macOS Tahoe, you can finally export Apple Notes in markdown. This is great, and even exports images and handwriting from the Apple Pencil. You can also bulk export, which is great! Sadly, bulk exporting images and handwriting doesn’t work. It saves the images and handwriting in a file called Attachements, but then in the actual note markdown, all references to images are to a file called FallbackImage.png. Also, it exports an sqlite database.

OpenAI’s study indicates that users primarily utilize ChatGPT as an advisory tool to enhance decision-making and productivity in knowledge-intensive jobs.

In my searches/research, analysts like Gartner, IDC, and Forrester are invisible to ChatGTP, Claude, etc. The robots will find the reports licensed by vendors - I’m guessing only if they don’t require leadgen. The analysts should probably start publishing some juicy abstracts with just enough numbers, analysis, and advice to get into the training data and search results to get people to funnel to their pages. Of course, converting a rando-click to a $95,000 PDF would be some real funnel-management magic.

Goldman Sachs has implemented its GS AI Assistant to enhance employee efficiency in tasks like document summarization and data analysis, while acknowledging the importance of personal nuance in client services.

Somewhere between 40% and 60% of apps run on private cloud, you just never hear about it.

I’m thinking through the topic of ROI for infrastructure software again (obviously for our PaaS stuff). I get asked about this every year or two. Each time I look at it, I get more confused. This ask always come from sales, presumably because the buyers are asking it. But, I don’t think the ask is about the basic ROI calculations: the amount we paid for this tech is less than the amount of money we make using it.

An enterprise AI strategy probably means adding AI to your existing apps and workflows, not just standing up a stand-alone AI app. We experience generative AI as chatbots in the consumer space - and they’re great! - but this doesn’t seem like the best approach for business applications. Think about search. We don’t even notice it now, but at work, search is built into existing apps, it’s not usually a stand-alone app that tightly integrates with and links into existing apps.

Much like financial debt, technical debt is helpful when managed responsibly, but like real debt, tech debt can also stop growth and innovation in its tracks. My colleague Bryan Ross has a piece out on tech debt with five ways to address is. Here’s a summary of it. Think of technical debt as the accumulation of compromises on development quality to save time. Some examples of these choices are postponing unit tests, running outdated software, or focusing on end-user features at the expense of internal processes.