Sam Witteveen explores the capabilities of this new model for multi-day agentic engineering, high-precision knowledge tasks, and 3D web development. The analysis covers benchmark performance, cost-effectiveness relative to competitors, and the potential impact of such advanced AI tools on future software development and application building workflows.
What's better than a new frontier model? Try two, both of them achieving new high scores on a variety of third-party benchmarks — all with a 20% cost reduction over API for enterprises and developers.
That's exactly what Anthropic has done today: less than an hour ago, the fast-growing generative AI startup announced Claude Fable 5.2 and Claude Mythos 5.2, its latest, most powerful frontier language models for general usage and cyber workflows, respectively, and the results across the board are, once again, nothing short of impressive — dare we say, potentially industry-changing.
Google open-sourced the Elephant-Goldfish Model (EGM) this week, and by Sunday evening the repo sat at over 14,600 stars, up more than 6,000 in seven days (I pulled that count from the GitHub API tonight). It is a plugin for AI coding agents that makes the agent write and defend a specification before it writes code. MarkTechPost covered the release today, and the pitch compresses to one line: catch wrong assumptions in a Markdown file, not in a pull request.
Amazon Web Services introduced Kiro bots on Thursday, adding persistent agents to its Kiro Crew development workspace that can oversee recurring responsibilities, coordinate other agents and remember the decisions developers make along the way.
The feature, now generally available, targets a problem that emerges when developers delegate several jobs to AI at once: Someone still has to check which agents have stalled, reconcile their results and explain the same requirements again when work resumes.
Recently JetBrains reached out to me to get my take on where I thought IDEs were going and what their role was in software development and agentic coding. They themselves have released an interesting blog post (which I refer to in the video) and are engaging with developers to get your take as well.
In this video, Sam Witteveen takes a look at the new announcement about the DeepSeek harness and the introduction of their super agent add-on for their harness.
Strata 0.2.8 shipped on October 29, 2026, adding faster paired Intel Arc B65 cards, a four-card Radeon setup that triples throughput, and a fixed memory leak that was crippling DeepSeek on Macs. The open-source engine runs Alibaba's 125-billion-parameter Qwen3.8-Flash-Next on a PC with as little as 12 GB of VRAM, using clever GPU-RAM-SSD splitting and speculative decoding.