Today, the Model Context Protocol (MCP) published its 2026-07-28 specification, the largest and most significant revision of the protocol since its launch. With this release MCP becomes a stateless protocol that scales on ordinary HTTP infrastructure. Alongside the transport changes, this new version introduces a governed extensions system, strengthens authorization by aligning more closely with…
The expulsion of five diabetes experts from the American Diabetes Association meeting last month continues to reverberate, with more than 200 members now calling for two ADA leaders to resign. In response, the ADA delivered its own message Tuesday, expressing appreciation for members’ patience while a report is prepared. The resignation demand, sent last week…
1. Medita Holdings is a pseudonym for a family-owned diversified company in southern Europe. 2. A. Sanz-García, M.P. García-Vera, and J. Sanz, “Is It Time to Replace the Big Five Personality Model? Factorial Structure of the NEO PI-R in a Community Sample of Spanish Adults,” The Journal of General Psychology 151, no. 3 (2024): 335-356,…
AI-agent development has progressed through overlapping phases: prompt engineering, context engineering, tool use, autonomous loops, memory systems, and multi-agent coordination. A newer focus is graph engineering, which treats AI applications as explicitly designed workflows rather than a single autonomous agent. Graph engineering defines how agents, tools, deterministic functions, validators, data sources, and humans coordinate to…
GPT-5.6 Sol edges Kimi K3 on single-shot quality, but Kimi wins on pass@k with k > 1 and costs 64% less per completed task. The two models succeed and fail in different ways, which makes routing between them the strongest play on the benchmark. Kimi K3 vs GPT-5.6 Sol is close on DeepSWE pass@1: Sol…
Featured projects TL;DR Helion is PyTorch’s high-level DSL for writing performance-portable ML kernels. Partnering with Google, we have built a TPU backend that compiles Helion kernels to Pallas, providing a PyTorch-friendly way to author performant TPU kernels. On a flash attention workload, the Helion-generated kernel achieves 838 TFLOPs (~79% MFU of one tensor core) on…
Listen to the article 3 min This audio is auto-generated. Please let us know if you have feedback. Dive Brief: Intermountain Health is fully acquiring two hospitals and more than 60 clinics that it jointly owns in Idaho from Surgery Partners, in a deal valued at almost $1.2 billion. Along with local physicians, the nonprofit…
The emergence of artificial intelligence has raised expectations for enterprise search, but the results do not always inspire confidence. Users may encounter fragmented or irrelevant search results or receive answers that lack the context or security controls needed to determine whether the information is accurate. Coveo Solutions Inc. is addressing these challenges through its AI-Relevance Platform,…
# Introduction The conversation in data science and AI has shifted dramatically over the past year. We’re no longer talking exclusively about large language models (LLMs) acting as reactive systems that only respond when prompted in a browser tab. The focus has moved to AI orchestration: giving these models the autonomy to execute complex workflows….
22nd July 2026 – Link Blog Are AI labs pelicanmaxxing? (via) Excellent piece of work by Dylan Castillo, who took a deep-dive into the frequently pondered question of whether the AI labs have been deliberately training models to draw pelicans riding bicycles in response to my deeply unscientific benchmark. I’ve been randomly spot-checking this in…