AI Digest: A Dispute Over Book Authorship, DraftKings, and the Neurochip by STC “Module”

Today: the consequences of using AI in advertising and questions of authorship, as well as Russian neural chips and practical experiments with models.

🔥 Hot:

🔹 In Canada, people are debating whether an award-winning book was written with AI 🔹 EFF warns that DraftKings is using AI to amplify the harms of behavioral advertising

➡️ News:

🔹 virtio-nvgpu allows KVM virtual machines to access NVIDIA GPUs — The project uses a virtio device to provide access to the graphics card with minimal performance loss. 🔹 Botbin launched a pastebin for AI agent artifacts

➡️ Useful materials:

🔹 Why you shouldn’t outsource all your thinking to AI — An analysis of the trap of blindly trusting AI assistants and the consequences of delegating more than just routine tasks to them. 🔹 A review of the Russian neural chip from the Module Research and Development Center — The author compares its capabilities, based on the datasheet and implementations, with Jetson and Ascend 310. 🔹 F-Secure tested whether AI shopping agents can be trusted

➡️ Discussions and case studies:

🔹 An LLM beat NetHack 🔹 The community is discussing a local model for 3D-printing monitoring — It is proposed for detecting failed prints and issuing timely warnings about problems. 🔹 A community member reported a speed of 85.6 tokens/s for Qwen3.8:27b on a single RTX 5090

📝 If you would like to add other news and materials to the list, write in the comments.