AI
Cache-augmented generation: the cheapest token is the one you never re-read
How CAG caches a frozen corpus to cut RAG cost — the O(n²) attention bottleneck, Redis TTL traps, and where a Swiss BIM studio should actually reach for it.
Vector Databases: The Nearest-Neighbour Geometry Under Every RAG Pipeline
Vector-database adoption grew 377% in 2025. Strip the marketing and it is nearest-neighbour search — a Voronoi partition of meaning-space. The maths, explained.
Opus 5 Rebuilt a Part From Pixels Alone — Read the Provenance
Claude Opus 5 reconstructed a machine part from a drawing it couldn't see. What that means for AEC reconstruction — and why every agent output needs a source stamp.
Open weights are a provenance instrument — pin the checkpoint before you ship the drawing
NVIDIA put 650 open models and 250 datasets on Hugging Face. For architecture offices the payload is not speed — it is a checkpoint you can cite.
One Endpoint, Every Model: Routing Your AI Coding Agents Through a Local Proxy
Weave's open-source router picks a model per request using an on-box embedder, not a prompt. Setup, a routing histogram hack, and the trade-offs.
Inkling and the Fine-Tune Question Every AEC Office Will Face in 2026
Thinking Machines' open-weights Inkling (975B/41B active) shifts the AEC question from which model to which data. What an architecture office should do Monday.
Own Your AI: Run Hugging Face Models Locally Before You Rent Them
A hands-on PAZ tutorial: install Hugging Face transformers, run inference and embeddings on your own hardware, and keep your studio's tender data on-site.
Grok Build Went Open Source — Now Read the Agent Loop Before It Reads Your Repo
SpaceXAI open-sourced Grok Build's agent loop and TUI. For an architecture office, the harness — not the model — decides what leaves the building.
AGI, Annotated: What the Term Actually Means Before You Put It in a Fee Proposal
AGI has no agreed definition in 2026. DeepMind's five capability levels and five autonomy rungs give an architecture office the vocabulary it actually needs.
GPT-5.6 Sol: Read the Access Graph, Not the Benchmark Table
OpenAI's GPT-5.6 preview passed through a US government review step. Why that gate — not Terminal-Bench — is the number your office should be reading.
