AI Briefing#software-engineering
The latest AI and developer news about #software-engineering, with the original source and a short summary.
The latest AI and developer news about #software-engineering, with the original source and a short summary.
A new benchmark of 202 real-world tasks across 13 languages finds Opus 5.5 maintains performance on functional languages while open-weight models like DeepSeek, GLM, and Qwen fall behind.
The model features a 1M-token context window for complex software engineering and cybersecurity tasks, with a 50% price reduction ending November 30, 2026.
The new experience allows users to ask Dropbox to organize files and answer questions about content, with all results remaining within the platform.
Developers built a closed loop that controls agent progress without writing code directly.
Microsoft's Copilot Cowork reached GA in 4.5 months, with a record 1,211 PRs merged in a single day, while a live Python-to-Rust migration reduced p50 time-to-first-token from 7.9s to 2.7s.
In a single-file test of 100 WordPress 4.0 files, the best LLM left only 59 passing behavioral checks compared to 83 in the unchanged originals.
The improved speed of AI implementation has increased the importance of determining task direction, structure, verification, and the scope of delegation.
The model achieves 6 confirmed strengths across 11 categories, including a #1 rank in Humanity's Sixth Sense and #2 in SWE-Bench Pro V2.
Claude Fable 5.1 ranks #3 on SWE-Bench Pro V2 and #2 on MCP Atlas, but its low Remote Labor Index score (17.92%) indicates it requires human verification for autonomous professional tasks.
In the Yogiyo PE pilot, code generation has sped up, but verification and processes remain bottlenecks.
The new benchmark features tasks 67% larger than previous versions and reveals that models struggle with unstated constraints and thoroughness, with the best model achieving a 45.5% score.
The study demonstrates that AI can produce formal proofs for software properties, a distinct capability from code generation that helps identify defects missed by conventional testing.
SlopCodeBench, a UW–Madison benchmark, tests 15 agents on evolving coding tasks and reports that structural erosion increased in 77% of runs.
The framework ranks codebase improvements as the most efficient method for ensuring agent reliability, followed by static analysis, prose, and human review.
While AI agents drastically lower code production costs, they amplify existing organizational weaknesses in decision-making, verification, and coordination rather than solving them.
Seemann argues that legal responsibility for AI-generated code defects will likely fall on human sign-offs, mirroring Tesla FSD liability models.
The rewrites exhibited significant architectural inconsistencies and performance issues, such as a 1% notification delivery rate in the Rust version under heavy load, demonstrating that AI agents cannot replace human oversight in system design.
Block leverages Claude Fable 5 as an orchestrator to direct smaller models like Opus and Sonnet, enabling the automated coordination of up to a thousand pull requests per migration.
The 4B-parameter model is derived from Qwen3.5-4B and trained on 1,500 synthetic SWE tasks using the Leaf harness.
GitHub advises developers to focus on directing AI agents, critically evaluating outputs, and solving higher-level architectural problems.
The new model launches at an introductory price of $2 per million input tokens and $10 per million output tokens.
The pipeline generated over 3,000 machine-checked safety and liveness proofs from real-world TLA+ specifications using an agentic prover-reviewer loop.
A six-month development of a Bach music generation system using Claude Code revealed that subscription fees accounted for only 2% of total costs, while the remaining 98% was spent on human judgment and verification.
Evaluation results from the Real-SWE benchmark showed that AI coding agents had low resolution rates due to unverified assumptions and missed requirements when handling ambiguous specifications.