Exploring the echoes reverberating through time left by the technology of yesterday as we embrace the technology of tomorrow.
Most organizations can't compete on AI spend. The strategic question isn't how to spend less than you'd like; it's how to extract disproportionate value from the smaller spend you actually have. Worth being plain about what frugal AI strategy looks like.
Three years past the original encoding-a-person framing, the question of what data actually encodes a person has gotten clearer. The answer is less about volume and more about the specific kinds of context that compound over time.
Direct commits to main are forbidden. Every change goes through a pull request, even when you're the only reviewer, even when you're the only person on the project. The discipline of writing the change down twice is what makes the work reviewable later, including by your future self.
The cheapest way to train an LLM in 2026 isn't to spin up a giant cluster or to negotiate with a hyperscaler. It's to start from a pretrained base, optimize the data, and use the cheap-tier infrastructure that's matured around small-scale fine-tuning.
CES 2026 week. NVIDIA unveils Rubin and the DGX Spark/Station desktop AI line, AMD ships Ryzen AI 400 with a 60-TOPS NPU, Qualcomm pushes Snapdragon X2 to 80 TOPS, Samsung wires Bixby and Alexa+ into every appliance, and Mercedes ships Nvidia-powered L2 in the CLA. My take.
Three layers, top to bottom: applications (the AI features users see), platform (model registry, serving, observability, governance), substrate (K8s, GPUs, storage). Decisions as Code runs through every layer. Centralize the decisions. Project them everywhere.
What it actually looks like to run a serious training pipeline at home in early 2026, the data prep, the orchestration, the evaluation, the operational discipline. Less hand-wavy than the typical write-up, more boring than the marketing pitch.
Twelve specific predictions for 2026, calibrated against the patterns I missed in 2025. Less hand-wavy than the typical year-ahead piece. I'll grade them plainly at the next checkpoint.
The bridge week. SB 53 went live on January 1, the year-end retrospectives all said roughly the same thing about 2025, the 2026 prediction pieces from Sequoia and a16z dropped with surprisingly compatible takes, and CES 2026 is sitting on Tuesday's runway. What I make of the turn of the year.
End-of-year recap. Most of the year's narrative was hype layered over a smaller set of substantive shifts. Worth being explicit about what actually changed and what didn't, before the 2026-predictions piece tries to extrapolate from the noise.
Multi-tenant from day one is roughly 2x the cost of the single-tenant version. The hidden costs aren't in the code, they're in auth, audit, query routing, observability, support tiering, and billing. Often the cheaper move is to start single-tenant and earn your way in.
MCP went from emerging standard to genuine substrate in 2025. The use cases that have actually shipped on it are more interesting than the broad-platform-narrative version. Worth being specific about which ones to track and why.
