Blog
Read about what we're up to and what we're learning along the way
Read about what we're up to and what we're learning along the way
24 articles
Block launched Buzz — an open workspace where humans and AI agents work side by side as equals, same identity, same permissions, same audit trail. This is the week multi-player AI stopped being a conviction held by a few teams and became a statement made by a major public company. Why we're celebrating a bigger megaphone, what it's like to actually use it, and where a shared mind fits in a world that now has a shared place.
Every coordination tool a software team has ever used was an answer to a question about scarcity. A genealogy of waterfall, agile, DevOps, and platform engineering — and how the agentic moment breaks the pattern. Chapter 2 of a book we're writing in public.
Everyone is asking what AI coding has changed. The more useful question (the one Bezos taught me to ask) is what it hasn't. Two stories from one week: a Heisenbug that turned out to be heap fragmentation no packet capture could see, and a whiskey-fueled argument with the architects behind Xet about a design choice no model would have made. The fat middle melted. The two ends got more valuable, not less.
The most important technology in a product is usually the part the customer never sees, built by someone solving a different problem. Alan Frindell spent years getting Media-over-QUIC right for live video at Facebook. A weekend conversation, six weeks, and one Seattle dinner later, it had become a standup that transcribes and illustrates itself in real time — told, fittingly, through a recording the system drew while we talked about it.
YAGNI isn't wrong — the floor it operates on has risen. When implementation cost collapses, scoping discipline belongs at the systems level, not the code level.
Scale is no longer about how many people you have. It's about how much throughput moves across your team relative to the attention you can bring to bear on it. The first chapter of a book we're writing in public.
A new idea is the fastest-acting antigen known to science — and the deeper problem is that we can't even perceive an idea that doesn't fit what we already believe. Investors, physicists, and jazz musicians each built a ritual to beat the reflex. Art schools built the most literal one of all, and recording everything is how we run it all day.
Every engineer we've invited to a work-trial would have been a confident 'hire' under the methods I trusted my whole career. Only two of six clicked. This is what the trials taught us about finding who actually thrives in the A.D. era — and why we let people find that out for themselves.
Steve Yegge and Ajit Banerjee on AI literacy cohorts, the death of job titles, why human connection is getting more important — not less — and what it actually means to 'cook' with AI. Hosted by AI House Seattle, May 20, 2026.
Ajit Banerjee and Zoom AI's Zhenbin Xu on how Zoom's brand-new intelligence APIs — topping Hugging Face's ASR leaderboards — are now powering SageOx's real-time speech recognition and speaker identification, and what it took to integrate them in days.
Dr. Rupak Majumdar — Scientific Director at the Max Planck Institute for Software Systems and SageOx Advisor — on the collapsing cost of code generation, the value of proof, and the trust signals that matter when software is abundant.
Dr. Rupak Majumdar continues his reflection — this time through the lens of economics. If code generation is the good whose price has dropped, where does the value flow? Substitutes, complements, and the rise of cognitive debt.