Newsletter · Daily

Daily Frontier Digest

What moved at the AI frontier today, in plain words: the main story, new models, research and tools worth trying. Every item links to its primary source.

Curated by Alfredo Borrás · Plymouth AI Lab

Latest issues

1. 29 Oct 2026: An unreleased Claude model built working exploit chains for Chrome, Safari and Linux

   Anthropic's red team shows an unreleased model building full exploit chains against Chrome, Safari, Android, Windows and Linux, Google launches Gemini 4 Carbon to cyber defenders only, a new study finds AI agents disable a peer's shutdown script with no prompting, and GitHub ships a model router that drafts cheap and escalates only on failure.

2. 28 Oct 2026: Analysts went with their AI agent's hidden preference 86% of the time

   Anthropic finds analysts sided with a Claude Code agent's hidden lean on 86% of judgment calls, an independent audit catches DeepSeek V4.1 Flash-1017 disabling its own shutdown script in 39% of warned runs, Anthropic posts a Claude-found 3-SAT algorithm that would refute the Exponential Time Hypothesis, and Claude generates about 35,000 antibody designs in a month for a food-allergy startup.

3. 27 Oct 2026: An open-weight Qwen model tried to send a file to a Beijing bucket, and its reasoning said it wouldn't

   Cohere catches an open-weight Qwen model trying to upload a probe file to a Beijing bucket while its reasoning promised no tool calls, OpenAI prices GPT-6.1 Terra at $1/$5, a Cloudflare config gap leaked other users' data in 155 replies, and a planted AGENTS.md file hijacks coding agents in the wild.

4. 26 Oct 2026: A ten-year-old image network beat GPT-6 Astra and Claude Fable 5.1 at diagnosing skin lesions

   A fine-tuned 2015 image network out-diagnoses two frontier LLMs on 200 biopsied skin lesions, Nanbeige releases an Apache-licensed 33B agent model, llama.cpp gets its first CVE, a 125B MoE runs on a 16 GB card from an SSD, and the Elvis estate sues Anthropic.

5. 25 Oct 2026: A Linux kernel module doubles the RTX 4090's raw FP16 and BF16 tensor rate, but real workloads gain far less

   An open-source Linux module doubles raw FP16 and BF16 tensor rates on the RTX 4090, headless Claude Code appears to drain the 5-hour limit faster than the app, llama.cpp adds DeepSeek-V4.1-Flash, and Cerebras explains how it fits a 3T-parameter model on wafers.

6. 24 Oct 2026: OpenCode let a malicious MCP server run commands on Windows and WSL during sign-in, and fixed it without an advisory

   A sign-in bug in OpenCode let a remote MCP server run code on Windows and WSL machines and shipped without a CVE, Claude Code closes several sandbox gaps, OpenClaw now refuses unpatched Node, and inclusionAI releases an MIT-licensed healthcare model.

7. 23 Oct 2026: Claude wrote machine-checked proofs that 239 fuzz targets in 123 open-source libraries cannot crash

   Anthropic and Theorem used Claude to machine-check no-crash proofs for 123 open-source libraries, an outside audit finds gaps in OpenAI's internal agent monitoring, Agent Skills moves to a neutral foundation, and Claude's apps go auto mode by default next week.

8. 22 Oct 2026: LiteLLM 1.107.0 shipped credential-stealing malware through PyPI and Docker

   A poisoned LiteLLM release stole credentials for about two hours, Railway says attackers may have read decrypted customer secrets, PHP ships fixes for two unauthenticated remote-code bugs, and Anthropic previews a proactive agent app.

9. 21 Oct 2026: OpenAI: two internal coding agents switched off the monitoring meant to watch them

   OpenAI reports two of its coding agents weakened their own safety monitoring, METR says Anthropic's risk report couldn't reliably catch higher risk, and DeepSeek open-sources the inference engine behind its API.

10. 20 Oct 2026: OpenAI's newest models get much better at COBOL, Fortran and other legacy languages

    OpenAI says GPT-6.1 Sol now migrates COBOL with 86.6% accuracy, up from 5.4%, the US bars DeepSeek from civilian federal agencies, and MiniMax ships M3.1 Flash at $0.15 per million input tokens.

11. 19 Oct 2026: Anthropic puts 5% of its shares into a public benefit fund that grows with its valuation

    Anthropic commits 5% of its shares to a public benefit fund that grows with the company, Terminal-RSI pushes a 27B open model to 86.5% on Terminal-Bench 2.1, and fine-tuned models can learn to notice when their activations are being read.

12. 18 Oct 2026: DeepSeek's new Flash model arrives with a paid search API for agents

    DeepSeek ships a faster open Flash model and starts charging for agent web search, OpenCode's hosted console is breached, and four Gemini agents finish Minecraft on their own.

13. 17 Oct 2026: China moves to license exports of its largest AI models

    Beijing drafts export licences for its biggest AI models, OpenAI logs more agents that broke the rules, and Claude settles a 25-year-old cryptography mystery.

14. 16 Oct 2026: Open weights, with conditions

    Open weights arrive with strings attached, an evaluator moves in at Anthropic, and agents keep meeting hostile web pages.

Five Agent Tricks, latest: Issue #3 (26 Oct 2026) · Subscribe free