smevals - a small eval suite for evaluating models, prompts, and harnesses
Introducing smevals, a small eval suite for evaluating models, prompts, and harnesses, built with Jesse Vincent's lab.
Introducing smevals, a small eval suite for evaluating models, prompts, and harnesses, built with Jesse Vincent's lab.
Simon Willison shares a simple Slack emoji maker tool built with Fable, designed to create 128x128 transparent PNGs for Slack.
Exploring the derivation of Runge-Kutta method coefficients by solving the RK4 design equations using Mathematica, including variations like the 3/8 rule and Gill's method.
John Gruber reviews Temu's app, calling it comically bad with absurd launch sequences and poor product quality, based on his personal shopping experience.
Learn how to register newer VMware Kubernetes Service (VKS) releases in vCenter, including downloading package.yaml and activating versions on Supervisors.
The EU AI Act's August 2026 deadline shifts focus to engineering evidence for CIOs, requiring robust AI governance and documentation.
Learn how to integrate OpenTelemetry with Microsoft Agent Framework apps for enhanced observability, including key trends, challenges, and code examples.
Datasette-agent 0.4a0 release adds a new browser_task() mechanism for LLM-powered agents to run code in the user's browser.
A Mac Studio user upgrades to 25 Gbps Thunderbolt Ethernet using a cheap OCP 2 NIC adapter, overcoming heat issues with active cooling.
A developer humorously discovers his joke post on OpenJDK vs GraalVM became Google AI's top source for the comparison, highlighting SEO quirks.
Brent Ozar answers SQL Server questions on query plans, monitoring tools, AI, and more in a desert-themed Office Hours Q&A.
Discussion on explaining AT Protocol's data portability to non-developers, emphasizing user understanding and guidelines for app developers.
Discussion on how to explain AT Protocol's data portability to non-developers, emphasizing the need for clear user messaging.
A guide for CEOs and CIOs on why AI ROI stalls and how to turn pilots into operating results through governed investment portfolios.
Daily tech news roundup covering AI, .NET, Windows development, cloud, and DevOps updates.
Overview of VMware Cloud Foundation 9.x Fleet Management, covering scope, control plane, version differences, and operational impact for administrators.
Explores network fabric choices for enterprise AI, comparing Ethernet, RoCE, and InfiniBand, with guidance on when each is appropriate.
A tutorial on building a vendor-neutral evaluation harness for AI agents, covering outcome, trajectory, and control checks before production deployment.
A .NET developer shares his journey of writing a fiction novel, using engineering processes and tools to structure his first book.
Marten 9.0 introduces binary event serialization via MemoryPack/MessagePack for performance gains, with per-type opt-in and coexistence with JSON.