OpenAI Daybreak·GPT‑5.6‑Cyber와 Muse Glimmer 오픈웨이트 중심의 최근 동향

We’re expanding our cybersecurity initiative Daybreak and introducing GPT-5.6-Cyber, a new model for advanced, authorized cybersecurity work. As the threat landscape evolves, we’re putting frontier intelligence in the hands of trusted defenders before attackers can deploy offensive AI at scale.

Daybreak Blue provides access to frontier models, including GPT-5.6 Sol, with safeguards calibrated for broad defensive work. It’s the recommended starting point for most defenders, supporting vulnerability discovery, secure code review, malware analysis, incident response, and patch validation.

Andrew Ng

@AndrewYNg

Thank you Mark, Alex and the whole Meta team for your contributions to open weight AI.

Today @AIatMeta introduced Muse Glimmer, an open-weight, 30-billion-parameter model distilled from Meta’s Muse Spark for on-device agentic workflows. Alongside, ExecuTorch is adding end-to-end support for running Muse Glimmer on @NVIDIA GPUs and Macs with @Apple silicon. Why ExecuTorch? ExecuTorch offers a smarter, native approach: Build in PyTorch: Implement your model and decoding strategies right in PyTorch as you normally would Seamless Export: Export directly to ExecuTorch when you're ready for deployment Automated Optimization: The framework automatically handles backend-specific lowering (Triton on CUDA, MLX-native, or Metal on Apple silicon) and uses ahead-of-time compilation to optimize the full execution path end-to-end Learn more: https:// pytorch.org/blog/fast-onde vice-agentic-ai-with-executorch …

Stagehand

@Stagehanddev

Introducing Stagehand v4: the SDK for browser agents. Playwright was built for testing, we built Stagehand for your agent: with improved context management, self-healing actions, and iframe support.

Harrison Chase

@hwchase17

How to build a web browsing agent Stagehand v4 is an SDK to let agents browse the web from @browserbase We created a tutorial to show how you could use Stagehand with Managed Deep Agents to create a production ready web browsing agent https:// youtu.be/O0hkpChFBkM

Jeff Dean's new company already has competition. Jeff Dean's Discovery Loop targets one key bottleneck with research today, i.e, research is human-intensive and runs one step at a time. His argument is that automating the full cycle raises both the count and the quality of experiments, starting with ML research itself. Transformer Lab's Primus is built around a similar premise, with the research question as the only thing a human provides. It runs for roughly a day and returns a complete paper. The loop runs in six stages: - Literature review: reads and synthesizes existing work before designing anything - Experiment design: scopes the study, picks models, datasets, and metrics - Provisioning: brings up GPUs and schedules the jobs - Execution: runs training and evaluation, catches failures, and retries - Analysis: checks results against the design it committed to - Writeup: produces the paper with methodology, findings, and caveats Each stage on its own is usually a routine task, since agents already do all of them. But holding them together across a day of wall-clock time wasn't solved yet, since a crash after a few hours invalidates everything after it. That is a scheduling and state problem more than a model problem. I tested Primus on a quantization research question, asking it to compare INT8 and FP16 on a small open-source vision model and pick the better tradeoff for consumer GPU deployment. The video below shows the run and the paper it came back with. Try it yourself here: https:// lab.cloud In my own run, as it worked, it caught two problems on its own. The image set it first downloaded was a compressed copy, and the compression was quietly inflating accuracy by about 2.9 points on all three models, so it went back and rebuilt every result on the original photos. It then noticed its own timing was unfair, since running the two precisions one after the other favoured INT8 at three of the four batch sizes, and switched to alternating them in short blocks. The result I would not have thought to look for is that INT8 costs ResNet-50 only 0.05 points of accuracy, which looks free, while quietly changing the answer on 4.3 percent of images. To know whether 4.3 percent was even a lot, it first built the same model twice in FP16 and measured how much two identical builds disagree with each other, which is 0.14 percent. Nobody has published that baseline, so there was never anything to compare against.

Impressive new paper from Meta. (bookmark it) Scaling laws assume model size and training data act on loss independently. This work introduces Skaling law, which couples capacity and data through a single interaction exponent. The extra term cuts mean absolute percentage error by 1.5x to 3x across both interpolation and extrapolation. The largest corrections land in the data-scarce and heavy-overtraining regimes where the standard Chinchilla and Kaplan forms drift. Paired with a sparse grid restricted to low-compute runs, it extrapolates the full grid using roughly 10x less compute than a uniform sweep. Why does it matter? Deployment now happens well past compute optimal. A law that stays accurate there, and that can be fit from small runs, changes how a pretraining budget gets planned. Paper: https:// arxiv.org/abs/2608.07222 Track more trending AI papers in our academy: https:// academy.dair.ai

Across broader evaluations, Ling-3.0-tiny reaches 71.03 on IMO-AnswerBench, 73.40 on GPQA Diamond, 83.15 on Multi-IF and a 69.54 non-hallucination rate. It delivers balanced coverage across reasoning, coding agents, instruction following and long-context tasks.

Data Protection Controls with LangSmith LLM Gateway Detect, redact, and replace PII and secrets in requests before they reach model providers Redact sensitive data from the traces Gateway records in LangSmith Now available in the LLM Gateway Public Beta for Enterprise users. Request access: https:// langchain.com/langsmith-llm- gateway-data-protection-request …