Running a system in production against real work is different from running it against test cases you designed to make it succeed. The production gap — the difference between what you built and what you needed — shows up in ways that are difficult to anticipate and impossible to fully test for.
With Llama 4 Scout's 10-million-token context and Claude Sonnet 4.5's 1-million-token beta window, a reasonable question is whether RAG is still necessary. If you can fit your entire knowledge base in a context window, why build the chunking, embedding, and retrieval infrastructure? It's
Adopting a framework means committing to its mental model. The longer you use it, the more that mental model shapes how you think about the problem space. LangGraph's mental model — workflows as directed graphs with typed state flowing through nodes — turned out to be a good fit for
There's a question anyone faces before building custom tooling: why not use what already exists? It's a reasonable question, and the answer shouldn't be "because I prefer to build things." It should be a specific list of requirements that existing solutions don&
GPT-5 shipped this week. The announcement positioned it as a unified system: a fast model and a deeper reasoning model behind a real-time router that decides how much thinking the query deserves. You don't choose o-series vs. GPT-series anymore — you just send the request and the system figures
Lakeflow hit general availability at DAIS, and the scope of what it covers is broader than the old DLT story. End-to-end data engineering: ingestion from external sources, transformation via Delta Live Tables, orchestration, and monitoring — all under one product surface. The declarative pipelines authoring experience in particular is a significant
Serverless GPU availability in Databricks was one of the quieter DAIS announcements, but it's worth talking through because the "serverless = cheaper and easier" assumption that applies to SQL warehouses and notebook compute doesn't hold the same way for GPU workloads. The economics are meaningfully
The model evaluation was useful, but it produced a conclusion that might seem discouraging: none of the alternatives I tested solved the problem I was trying to solve. Let me be more precise about what that means, because "they don't work" and "none of them
Databricks announced full Iceberg support in Unity Catalog at DAIS — managed Iceberg tables, the Iceberg REST Catalog API, read/write from external engines like Spark, Flink, and Kafka. The format wars between Delta Lake and Iceberg have been running for years. This announcement doesn't end them, but it
Databricks announced Lakebase at DAIS — a fully managed, Postgres-compatible operational database built on their acquisition of Neon. The announcement landed harder than I expected, because it isn't just a new product. It's a structural argument that the lakehouse can absorb the OLTP workload that's
When you've decided the current tool isn't doing the job, the responsible move is to evaluate the alternatives before building from scratch. I spent the better part of three months doing that evaluation. The alternatives were Claude, Pi Coder, and Hermes — each with a different set
The headline from DAIS this week was Agent Bricks — Databricks' take on what it should feel like to build a capable agent without hand-tuning every component. You describe the task, connect your data, and the platform auto-optimizes the agent: selects the model, tunes prompts, wires the retrieval layer. The