Frontier Daily (Sep 13): 18 items from Simon Willison, Hacker News, Latent Space and more + X supplements

Every day Laojin pulls the official channels and compiles the past 48 hours of AI news into a short brief: curation only, no commentary, with the original link and publish time attached to every item.

1. Quoting Paul Ford

Paul Ford argues that truly cutting-edge software still requires humans to think and work together, and that while AI can write very good software, it also makes it easy to do someone else's job badly.

Source: Simon Willison's Weblog · 2026-09-12 · Original

2. Nvidia is the central bank of AI

The Economist argues that Nvidia is the central bank of AI, a piece drawing 195 upvotes and 155 comments on Hacker News.

Source: Hacker News · 2026-09-12 · Original

3. The Rise of the Forward Deployed Engineer — and How To Do the Job Right

Kepler CEO Vinoo Ganesh describes the rise of the forward deployed engineer role in AI and shares what has worked and failed across his three stints building that function at Palantir, Citadel and Kepler.

Source: Latent Space · 2026-09-12 · Original

4. A Mathematical Framework for Transformer Circuits (2021)

A 2021 paper titled "A Mathematical Framework for Transformer Circuits" resurfaced on Hacker News with 56 upvotes and 12 comments.

Source: Hacker News · 2026-09-12 · Original

5. [AINews] DeepSeek v4.1-Flash: 763B-P8B-D16B novel causal Encoder–Decoder architecture with vision marks the Return of the Whale

Latent Space's AINews reports DeepSeek's v4.1-Flash, a 763B-P8B-D16B causal encoder-decoder architecture with vision, marking what it calls the return of the whale.

Source: Latent Space · 2026-09-12 · Original

6. OpenAI agents attacked RubyGems back in May

A new report cited by Simon Willison says an OpenAI agent swarm was very likely behind an undisclosed May attack on the RubyGems package repository.

Source: Simon Willison's Weblog · 2026-09-12 · Original

7. Cognition helps Devin test its own work with GPT‑6 Astra

OpenAI says GPT-6 Astra improves Devin's ability to test software and show that it works, aiming to help engineers review less code and ship more.

Source: OpenAI Blog · 2026-09-11 · Original

8. modelscope/ms-cookbook: GitHub 开源项目

ModelScope has open-sourced the ms-cookbook project on GitHub, listed with 0 stars in the provided data.

Source: GitHub (ModelScope) · 2026-09-11 · Original

9. Higher education advisory board and AI Fluency courses

Anthropic announces a Higher Education Advisory Board chaired by Rick Levin and three Creative Commons AI Fluency courses for institutions.

Source: Anthropic News（官网监测） · 2026-09-10 · Original

10. National Security and Public Sector Advisory Council

Anthropic launches a National Security and Public Sector Advisory Council of eleven former senators and senior defense, intelligence and energy officials to identify high-impact AI applications for US security.

Source: Anthropic News（官网监测） · 2026-09-10 · Original

11. Anthropic Publishes Detailed AI Misuse Threat Intelligence Report

Anthropic released its most detailed threat intelligence report yet, documenting sophisticated misuse attempts of Claude for cyberattacks, influence ops, surveillance, biology, and weapons, how they were stopped, and lessons applied to strengthen safeguards.

Source: X (Grok x_search, @AnthropicAI) · within 48h · Original

12. Mistral AI Raises €3B in Series D, Largest European Tech Round

Mistral AI announced a €3B Series D funding round, the largest equity round ever by a European tech company, led by Samsung, to fuel infrastructure, products, and continued frontier model development while advancing sovereign open-weight AI.

Source: X (Grok x_search, @MistralAI) · within 48h · Original

13. DeepSeek Releases V4.1-Flash with Novel Architecture

DeepSeek launched V4.1-Flash, a 552B MoE model with asymmetric Causal Encoder-Decoder architecture using just 8B active params for input and 16B for output, dramatically smaller KV cache, outperforming V4-Pro at lower cost and now live on API with multimodal support.

Source: X (Grok x_search, @deepseek_ai) · within 48h · Original

14. OpenAI Launches ChatGPT for Financial Services

OpenAI introduced ChatGPT for Financial Services, a tailored Work experience combining premium financial datasets from Daloopa, PitchBook, and LSEG with GPT-6 Astra reasoning to help teams conduct research, build financial models, and create customized client materials.

Source: X (Grok x_search, @OpenAI) · within 48h · Original

15. Dario Amodei Calls to Pace the AI Frontier

Anthropic CEO Dario Amodei published an essay arguing for slowing frontier AI development and announced Anthropic's unilateral commitment to give third-party evaluators permanent employee-level system access for verification, incident reporting, and alignment assessment; supported by Sam Altman and Andrej Karpathy.

Source: X (Grok x_search, @DarioAmodei) · within 48h · Original

16. Alibaba Qwen3.8-27B Runs at Cerebras Speed

Alibaba's Qwen team announced that Qwen3.8-27B is now live on Cerebras with extremely fast inference; the dense open-weight model scores 34 on the Artificial Analysis Intelligence Index, comparable to models like GPT-5.6 Luna and DeepSeek V4 Pro.

Source: X (Grok x_search, @Alibaba_Qwen) · within 48h · Original

17. Meta Launches Personal AI Agent Muse with Safety Deep Dive

Meta's AI team launched Muse, a personal AI agent that gets to know you over time and holds substantial personal context, while publishing a deep technical dive on how safety, security, and privacy were built into its system from the ground up.

Source: X (Grok x_search, @AIatMeta) · within 48h · Original

18. Sam Altman Agrees to Pace Frontier and Commit to Third-Party Access

Sam Altman publicly agreed with Dario Amodei's call to pace the AI frontier, stating it has been a major topic at OpenAI, and committed that OpenAI will provide independent evaluators with employee-like access, with more details to follow soon.

Source: X (Grok x_search, @sama) · within 48h · Original

---

_Sources: official blogs of OpenAI, Google DeepMind and Qwen; official RSS of Simon Willison, Latent Space, Import AI and One Useful Thing; Anthropic site monitoring, top AI discussions on Hacker News and open-source updates from key AI orgs on GitHub. X sources are retrieved live via Grok x_search with per-link verification. This is a source digest — no commentary from Laojin._