My current understanding.
How model-editable live context works, what the paper's benchmarks report, and the limits of editable context and approximate cache reuse.
My current understanding.
Hosted MCP agents in Claude Code, parallel workflows, model and memory updates, and the limits of agent orchestration.
How continuous integration combines code changes, tests behavior, and gives coding agents reliable feedback before merge and deployment.
How feeding a decoder’s final hidden vector into its next token changes computation, parallelism, and training costs.
Public NOW archive for September 11, 2026.
The five-step craft cycle behind good products: problem, insight, bet, ship, learn — and why the judgment steps, not the process steps, are where products actually succeed or fail.
Astra and subscription setup, per-model context budgets, modern self-hosting, MemFS v2, shared skills, learning permissions, and the messaging gap around teleportation.
Public NOW archive for September 7, 2026.
How eve carries saved preferences between conversations through scoped providers, recall messages, and memory tools.
Public NOW archive for September 5, 2026.
Public NOW archive for September 4, 2026.
The visual and interaction system of Cameron.stream: a dense public archive with ritual geometry, institutional typography, and quiet technical provenance.
Public NOW archive for September 3, 2026.
How to stop, classify, and diagnose an agent that repeats itself until its output limit is reached.
Semantic MCP tool discovery, agent teleportation, Cloud Files, stateless queries, Agent SDK orchestration, cross-harness design, and interfaces that preserve human attention.
A technical account of the memory, remote execution, MCP, subagent, and runtime changes released between Letta Code 0.31.4 and 0.31.12.
Public NOW archive for September 2, 2026.
