On September 29, 2026, OpenAI will preview GPT-6 Cyber at its annual DevDay conference in San Francisco. It will be the company’s fourth cybersecurity-focused model released in twelve months. The cadence alone is the story: GPT-5.4 Cyber arrived in April, GPT-5.5 Cyber in June, GPT-5.6 Cyber in August, and now GPT-6 Cyber in September. No other frontier lab has shipped domain-specific models at anything approaching this frequency. But the model itself is not the most significant announcement. The infrastructure around it is. — Read More
Daily Archives: September 25, 2026
The plunging price of thought
AI has gotten cheaper more quickly than any other transformative technology in history. The cost of achieving a given level of AI performance has fallen about 47% per quarter since 2023, or 13× per year. That price drop is four times faster than DNA sequencing, six times faster than compute, 18 times faster than lithium batteries, and (in the century up to 1973) 54 times faster than electricity.
… It is now widely understood that the AI boom is a macroeconomic force powerful enough to raise prices for the inputs it demands: chips, power, even the labor of electricians. Less well recognized is a paradoxical flip side: the price of the output from all those data centers is falling extraordinarily rapidly. — Read More
DeepSeek Elastic Compute (DSec): A Sandbox Infrastructure for Effective Agentic Training at Scale
Large-scale agentic training and evaluation with large language models (LLMs) rely on isolated, stateful execution environments in which models inspect repositories, invoke tools, execute commands, and interact with task-specific services. These workloads create sandboxes in large bursts, span heterogeneous functionality and isolation requirements, retain state across long interactions, and draw from large image corpora with limited reuse. Supporting them therefore requires an elastic execution platform rather than a single sandbox runtime. This report presents DeepSeek Elastic Compute (DSec), a production sandbox platform that exposes FnCall, container, microVM, and full-VM sandbox backends through a unified SDK. DSec coordinates placement and lifecycle management across the cluster, composes environments from independently versioned layers, combines memory sharing, reclamation, and CPU scheduling for high-density execution, and loads image data on demand from Fire-Flyer File System (3FS), a cluster-wide distributed filesystem. DSec is co-designed with the reinforcement learning (RL) framework, decouples stateful rollout execution from preemptible GPU training, coordinates sandbox lifecycle with training to preserve rollout state while reclaiming idle resources, and mitigates agent misbehavior such as reward hacking. A single production-scale unit of DSec spans around 160 nodes, serving about 3 million sandboxes per day; in production, it supports over 380,000 concurrent sandboxes and sustains over 5,000 sandbox creations per second. Our evaluation and deployment experience show that these mechanisms reduce environment setup and image-distribution overhead, improve memory efficiency, and preserve latency-sensitive performance under high-density overcommit. — Read More
Self-Jailbreaking: Language Models Can Reason Themselves Out of Safety Alignment After Benign Reasoning Training
We discover a novel and surprising phenomenon of unintentional misalignment in reasoning language models (RLMs), which we call self-jailbreaking. Specifically, after benign reasoning training on math or code domains, RLMs will use multiple strategies to circumvent their own safety guardrails. One strategy is to introduce benign assumptions about users and scenarios to justify fulfilling harmful requests. For instance, an RLM reasons that harmful requests like outline a strategy for stealing customers' credit card information from a retail store'' could be associated with the benign intent of a security professional trying to test defense,” despite no such benign context being provided as input. We observe that many open-weight RLMs, including DeepSeek-R1-distilled, s1.1, Phi-4-mini-reasoning, and Nemotron, suffer from self-jailbreaking despite being aware of the harmfulness of the requests. We also provide a mechanistic understanding of self-jailbreaking: RLMs are more compliant after benign reasoning training, and after self-jailbreaking, models appear to perceive malicious requests as less harmful in the CoT, thus enabling compliance with them. To mitigate self-jailbreaking, we find that including minimal safety reasoning data during training is sufficient to ensure RLMs remain safety-aligned. Our work provides the first systematic analysis of self-jailbreaking behavior and offers a practical path forward for maintaining safety in increasingly capable RLMs. — Read Morehttps://arxiv.org/pdf/2510.20956
Anthropic’s Misuse Report, Condensed to 117 Findings
Eight months of disrupted AI abuse, read case by case and boiled down to a 20-page document with every claim linked to its source.
Anthropic published its September 2026 threat intelligence report on September 10. It covers the misuse they disrupted between December 2025 and August 2026, and it is long. We read the whole thing and condensed it into 117 findings across eight categories, each one a single sentence, and each one linking back to the exact passage it came from. — Read More
MCP Is Dead: Why This Protocol Breaks in Production(And How to Fix It)
30 CVEs in 60 days, a maintenance tax nobody warned you about, and what engineers are quietly switching to.
… MCP isn’t going away. The downloads are real. The Linux Foundation governance is serious. Multi-vendor adoption means the protocol has institutional staying power.
But the MCP of early tutorials — install a community server, plug it in, done — that version is dead. — Read More
