2026-11-03 · TuesdayPublished daily at 08:00
AI Daily
OpenAI launches Grid Protectors initiative to support grid peak shaving, starting with its Georgia campus

OpenAI launched the Grid Protectors initiative, working with utilities, regulators, and policymakers to reduce electricity demand at future compute campuses when the grid is under strain, beginning with its Effingham County campus.
- 1
- 2
- 3Digital Life Kazike shares the architecture of an Agent automation system built over a year of tinkeringWeChat Official Account: Digital Life Kazike
01Product Releases/Updates
Claude Platform launches message threads in public beta to reduce latency for long-conversation agents
Anthropic launched message threads in public beta on Claude Platform, allowing agents to send only new messages each turn instead of resending the entire conversation history, reducing time to first token.
Original textTencent Hunyuan releases Hyra 2.0: Enabling general-purpose agents to develop reusable scientific tools and models
Tencent Hunyuan released Hyra 2.0, shifting the focus of automated scientific research from solving problems directly to having agents develop and upgrade scientific models, workflows, and software tools. The system uses a three-tier structure of Overall, Mission, and Worker, connecting to a real development environment that includes WeCom, computing platforms, and code repositories. Its results have been published in a GitHub repository.
Original text02Industry News
OpenAI launches Grid Protectors initiative to support grid peak shaving, starting with its Georgia campus
OpenAI launched the Grid Protectors initiative, working with utilities, regulators, and policymakers to reduce electricity demand at future compute campuses when the grid is under strain, beginning with its Effingham County campus.
Original textAnthropic’s quarterly revenue grows twelvefold to nearly $17 billion, but net losses reach $23.3 billion
According to Bloomberg and The Information, Anthropic’s revenue for the quarter was nearly $17 billion, twelve times the figure a year earlier. However, operating costs reached $17.5 billion, and quarterly net losses totaled $23.3 billion, up from $12.1 billion in the same period last year.
Original text03Paper research
OpenAI trains and deploys activation classifiers as a safety monitoring layer for frontier models
Starting with GPT-5.6, OpenAI has deployed activation classifiers (probes) on internal and external traffic for its most capable models. These probes use internal model activations to score each token in real time, allowing noncompliant responses and tool calls to be blocked mid-generation and referred to a reasoning model for review. Experiments show that probes are more robust than larger text classifiers against the jailbreak attacks tested, improve in accuracy as model capabilities increase, and transfer well across post-training checkpoints. However, they require retraining for new base models and can detect only known violation types for which they can be explicitly trained.
Original text04Tips and perspectives
Digital Life Kazike shares the architecture of an Agent automation system built over a year of tinkering
Independent content creator Kazike shares the Agent automation system he has spent nearly a year building. The core idea is to shift task initiation away from people and let events trigger Agents to work. The system has four layers: an automation hub, Agent execution, a CI hub, and Feishu-based decision-making. The automation hub uses rules.json to centrally manage event triggers and Agent scheduling. A local MacMini runs Agent tasks, while a self-built CI hub handles five steps: review, checks, merging, release, and verification. Only critical decisions, such as deleting data, spending money, or changing direction, are sent to the author through Feishu cards for approval.
Original text