Chandler Brown
Technical Notes on Data, AI, and Product Systems
Implementation-first writing on architecture decisions, reliability tradeoffs, and lessons from shipping real systems.
Latest Writing
View full archiveAnthropic's Emotion Concepts Are a Safety Signal
Anthropic's emotion-concept paper is useful less as a sentience claim than as evidence that internal model states can predict and shape risky behavior.
Gemma 4 Pushes Google's Open Model Strategy Into Real Deployment Territory
Gemma 4 matters because Google is no longer treating open models as a single product tier. It now spans phones, laptops, workstations, and agentic workflows.
Cursor's Bugbot Is a Preview of Agentic Code Review
Cursor's Bugbot learned rules and MCP support show code review moving from static checks toward persistent, tool-using systems that improve from feedback.
Project Glasswing Shows Where Frontier Security Is Going
Anthropic's Project Glasswing shows frontier models moving into defensive security work, while also making exploit generation cheaper and faster for attackers.
OpenAI's Safety Fellowship Is a Map of Open Problems
OpenAI's Safety Fellowship is less a branding move than a list of safety and alignment problems the company still wants outside help to solve.
TimesFM 2.5: A Practical Read for Forecasting Teams
TimesFM 2.5 is one of the most credible open forecasting foundation models right now, especially for fast multi-series experiments with long context and quantile output.
Axios Supply-Chain Attack: What Actually Happened
In late March 2026, two malicious Axios versions were published to npm and weaponized install-time scripts through a fake dependency, creating a short but high-impact supply-chain window.
Cisco IMC Advisories: What Operators Need to Patch First
Cisco published two IMC advisories on April 1, 2026, including an auth bypass and several root-level command execution flaws across server and appliance families.
Are We Moving Beyond React? The Rise of Visual State Systems
React still leads frontend development, but signals and graph-oriented visual systems are changing how teams manage complexity and rendering cost.
Making RAG Respect Permissions
A practical ACL-aware RAG blueprint that captures NTFS permissions at index time and enforces authorization inside retrieval.
Why Quantization Is About to Make Local AI Explode
Local AI adoption is accelerating because quantization shifts model memory and latency into ranges that regular developer hardware can actually sustain.
The IDE Is Dying. Agent Environments Are Replacing It
The center of software development is shifting from writing every line manually to directing agents across editor, terminal, and cloud environments.
When Declarative Systems Break, and Where Imperative Still Wins
Declarative systems scale intent and reliability, but they introduce failure modes where explicit imperative control is still the safer engineering choice.
Databricks Medallion Architecture in Production
A practical guide to running bronze, silver, and gold in production when schemas drift, sources are messy, and reliability actually matters.