Laguna XS 2.1 is a 33B Mixture-of-Experts model with 3B activated parameters per token, tuned for agentic coding and long-horizon tasks on local machines. It outperforms XS.2 on SWE-bench Multilingual by 5.4 points, adds support for major runtimes, offers FP8/INT4/NVFP4 checkpoints, and runs with a permissive OpenMDW-1.1 license. You can access it via API, OpenRouter, or locally with quantized or DFlash-boosted builds.

• Laguna XS 2.1 jumps from 57.7% to 63.1% on SWE-bench Multilingual, beating Qwen 3.6 (35B) and North Mini Code (30B) despite only 3B active params per token
• DFlash speculator models double token throughput versus stock XS 2.1 when run locally
• Same API pricing as XS.2 ($0.10 input/$0.20 output/$0.05 cache-read per million tokens), now served at 256K context
• XS.2 leaves the API in one week, pushing users to migrate to 2.1 (though it remains available on Baseten's Model Library)

This article breaks down the gap between your prompts (the map) and the real codebase (the territory) into four types of unknowns and shows how to uncover them. It offers concrete prompting patterns—blindspot passes, brainstorms, interviews, references, and implementation plans—to surface issues before, during, and after coding with Claude Fable 5.

• The gap between your prompt and the actual codebase can be sorted into four unknown types—known knowns, known unknowns, unknown knowns, and unknown unknowns—and naming them helps catch bad guesses before Claude runs with them.
• A "blind spot pass" (asking Claude to list unknown unknowns given your background/project) surfaces gaps you'd never think to mention.
• Having Claude generate multiple HTML mock-ups or fire off one-question-at-a-time interviews forces design and spec assumptions into the open early, before they become costly rework.
• When you can't articulate what you want, pointing Claude at existing reference code (e.g., a Rust backoff crate) to reimplement its semantics is more effective than trying to describe it from scratch.

Jane Street had long skipped full-on formal methods despite using advanced type systems, but the rise of agent-driven coding cut proof costs and widened access. They’re now forming a team to integrate formal verification into their OCaml toolchain, tweaking the language and tapping their experienced user base while collaborating with external proof ecosystems.

• seL4's formal verification cost 25 person-years for 8,700 lines of C (23 lines of proof per code line, half a person-day per line), illustrating why Jane Street long saw full formal methods as impractical.
• Agentic AI coding flips that math: it lowers the cost of generating proofs and widens access to formal methods, but also produces messy code that creates a new "verification bottleneck."
• Jane Street is building an internal formal-methods team to extend OCaml itself (modular specs, ownership/mutability typing, embedded tactics) rather than switch to tools like Lean, Dafny, or Coq, while still integrating those ecosystems where useful.
• The plan is to use formal verification both to catch bugs in AI-generated code and to feed precise feedback back into training/improving the coding agents.

GPT-5.5 outperforms GPT-5.4 in real-world coding tasks, from debugging and large merge operations to interactive app development. It also serves as a research partner—critiquing manuscripts, proposing analyses, and generating reports on complex datasets—all while running at GPT-5.4 latency through integrated inference optimizations.

• Note: this "GPT-5.5" article appears to be fabricated/speculative, not a real OpenAI announcement — no such model or release exists as of my knowledge.
• As summarized: GPT-5.5 reportedly matches GPT-5.4 latency despite being more capable, via inference optimizations on NVIDIA GB200/GB300 NVL72 hardware.
• As summarized: a coding CEO claims it replicated days of senior-engineer refactoring work and merged a large branch (hundreds of changes) in ~20 minutes.
• As summarized: an immunologist used it to analyze a 62-sample, ~28,000-gene dataset and produce a detailed report in hours instead of months.