AI Enthusiast Log (Page 2)

AI Enthusiast Log

Sign in Subscribe

More issues

GM tripled pull requests by redesigning engineering loops around AI agents

At VB Transform 2026, General Motors' autonomous driving division described a structured approach to AI agent adoption that led to roughly three times as many merged pull requests, faster releases, and fewer defects escaping into later stages of development. The claim comes from Rashed Haq, GM's VP

Snowflake positions Cortex AI Gateway as the agent governance layer, but preview-stage claims need production proof

Snowflake's announcement of Cortex AI Gateway frames the product as the control plane for agentic AI: a centralized layer that governs authentication, permissions, audit logging, and spending limits for both Snowflake-built agents and third-party agents including Claude Code and Cursor. The case rests on a structural

MCP's biggest update targets the enterprise AI scaling bottleneck

The Model Context Protocol has shipped its largest architectural revision since Anthropic released it twenty months ago, and the core change is a fundamental one: MCP is now stateless. The shift eliminates the session-sticky requirement that forced organizations to route requests back to the same server instance, a constraint

Kimi K3 weights are out: the real cost is in the license, not the benchmarks

Moonshot AI has released the full weights for Kimi K3, completing the rollout of what the company describes as the world's first open 3T-class model. The technical specifications are substantial: a 2.8 trillion-parameter Mixture-of-Experts architecture that activates 104 billion parameters from a pool

Microsoft's MAI-Cyber-1-Flash bets on cost over frontier model supremacy

Microsoft's security AI strategy rests on a 90/10 split, not a frontier model bet. The company announced MAI-Cyber-1-Flash, a compact model developed by its Microsoft AI division, designed to handle up to 90% of security tasks efficiently, while MDASH escalates the remaining 10% of

AI cites your deep pages but sends humans to your homepage — most sites are built for the wrong visitor

The click economy is not coming back, and the split it is leaving behind is more specific than the usual "AI is killing the web" framing suggests. Every major data source now confirms the same pattern: AI systems cite deep content when they generate answers, but they send

Ransomware built for AI weights signals a shift in attacker incentives

Sysdig's Threat Research Team documented two intrusions through the same exposed Langflow server, fourteen months apart. The first ran improvised Python against configuration files. The second brought ENCFORGE, a compiled Go binary whose extension list was written for AI infrastructure specifically. PyTorch and TensorFlow checkpoints, Hugging Face SafeTensors

Claude shared conversations indexed by Google: why the artifact era changes the stakes

A Reddit user discovered last week that shared Claude conversations were appearing in Google Search results. VentureBeat independently verified that shared Claude Artifacts, including interactive applications and business documents, were also searchable and accessible without authentication. By the following morning, some results had disappeared, suggesting Anthropic, Google, or both had

Enterprise AI agents: Most are chatbots, and the controls lag deployment

VentureBeat Research fielded five parallel surveys in June 2026, reaching 573 qualified respondents at organizations with 100 or more employees, and the central finding cuts through the agentic AI hype directly: enterprises deployed AI agents knowing the controls were not in place. That is not a discovery. It is a

Anthropic's Opus 5 bets token efficiency over raw benchmark dominance

Anthropic released Claude Opus 5 on Friday with a pricing move that says more than the benchmarks. The model holds input and output token costs at Opus 4.8 levels while roughly doubling performance on key agentic evaluations. The pitch is not that Opus 5 is the smartest model in