Blog
Short updates, essays, and experiments — all in one feed.
AI Water Usage and Orders of MagnitudeSorting through conflicting claims about AI data-center water use and why orders of magnitude matter when comparing numbers.
October 18, 2025Vibe coding an Atari 2600 port of FortniteBuilding a rough Atari 2600-style Fortnite with Amazon Q: constraints, glitches, and a playable demo.
June 28, 2025Doom unto allNYT embedded Doom; a quick note on the web port and the irresistibility of playing it.
May 29, 2025Fireflies are on the declineChildhood memories of abundant fireflies and the unsettling reality of their decline.
May 26, 2025Sometimes AI remembers the wrong meChatGPT's memory can overfit past queries, producing the wrong 'me' and raising personalization concerns.
May 23, 2025Turns Out I’m Not Famous (or in the Force)A funny 'Separated at Birth' anecdote about look-alikes and a reality check on my supposed fame.
May 22, 2025Microsoft’s AI Security Chief eats some dogfoodA protest at Microsoft Build led to an accidental screen-share leak of Walmart AI plans, highlighting security optics.
May 21, 2025AI has the most dangerous demosTrying low-stakes AI experiments shows how demos hide the hard parts; real value comes from skill, iteration, and context, not flashy outcomes.
May 21, 2025Shortcuts Don’t Skip the ClimbA conversation with my son about AI and school, and why shortcuts do not replace foundational skills.
April 14, 2025The one about Shopify's AI positioningTobi's memo on making reflexive AI use a baseline expectation at Shopify, preserved in full.
April 10, 2025John Carmack on AI in game devCarmack's quote on tooling and compute sparks a reflection on AI as a force multiplier in game dev.
April 10, 2025"Can we use AI?"On the reflexive 'can we use AI?' impulse in civic and organizational settings and the real gap: skills to use it well.
April 10, 2025Pinterest Gets SloppedPinterest search is getting flooded with AI-generated slop; quick sampling and what it means for discovery.
March 3, 2025Asserts > AlertsWhy assertions beat alerts for risky code: fail fast on bad assumptions and reduce silent failures.
March 3, 2025Custom Sorting in SAQL, Domo, and BI Dashboards Without Ugly Number PrefixesHow to implement custom sort orders in Salesforce CRMA/SAQL without ugly numeric prefixes.
February 25, 2025The one about tabsToo many browser tabs as a form of memory and curiosity; a small reflection on attention and knowledge work.
February 23, 2025AI Tinkerers, Jan 30 2025Recap of the Jan 30 AI Tinkerers meetup in Toronto: talks, demos, and why the community event mattered.
January 31, 2025The Blind Spot in Internal AI: Security and Reputation RisksInternal chatbots are easy to jailbreak and can create reputational risk; practical steps to harden them.
October 19, 2024Three Mindset Shifts Required for Successful AI Adoption in TeamsThree mindset shifts - AI is not cheating, speed does not reduce value, and creativity can grow with tools - plus ways to reinforce them.
October 13, 2024