Today in one line
Today: Nvidia CEO pushes back against report that his company’s $100B… · ChatGPT isn’t the only chatbot pulling answers from Elon Musk’s… · AI: AI coding tools hurt learning unless you ask why, Anthrop….
Top 10 — News
- Nvidia CEO pushes back against report that his company’s $100B OpenAI investment has stalled Nvidia CEO Jensen Huang said that a recent report of friction between his company and OpenAI was “nonsense.”
- ChatGPT isn’t the only chatbot pulling answers from Elon Musk’s Grokipedia More chatbots are citing Elon Musk’s AI-generated ‘Grokipedia’, raising the risk of citation laundering and feedback loops in information ecosystems.
- OpenAI still leads enterprise AI, but Anthropic is gaining fast, according to new study An oligopoly is taking shape in enterprise AI: OpenAI still leads, but Anthropic is catching up fast while Microsoft dominates applications.
- Autonomous cars, drones cheerfully obey prompt injection by road sign Physical prompt injection: road signs and other real-world text can steer autonomous cars/drones if systems treat external instructions as trusted.
- Mobile carriers can get your GPS location A technical walkthrough of how mobile carriers can infer GPS/GNSS location—highlighting a privacy surface that’s easy to underestimate.
- How YouTube and Adhesive Tape Are Disrupting Assistive Technology Assistive technology is expensive, and many people with disabilities live on fixed incomes.
- Generative AI and Wikipedia editing: What we learned in 2025 A year-end reflection on how generative AI is changing Wikipedia editing—workflows, quality control, and community norms.
- Parametric Knowledge is Not All You Need: Toward Honest Large Language Models via Retrieval of Pretraining Data arXiv:2601.21218v1 Announce Type: new Abstract: Large language models (LLMs) are highly capable of answering questions, but they are often…
- US has investigated claims WhatsApp chats aren’t private US authorities reportedly investigated claims that WhatsApp chats aren’t private—another reminder that privacy depends on more than just ‘end-to-end encryption.’
- Finland to end “uncontrolled human experiment” with ban on youth social media
Finland is considering a ban on youth social media, framing current exposure as an uncontrolled ‘human experiment.’
AI tools & model updates
- AI coding tools hurt learning unless you ask why, Anthropic study finds Developers who learn new programming skills with AI assistance score significantly worse on knowledge tests, according to a new Anthropic s…
Full list — News (14)
Nvidia CEO Jensen Huang said that a recent report of friction between his company and OpenAI was “nonsense.”
More chatbots are citing Elon Musk’s AI-generated ‘Grokipedia’, raising the risk of citation laundering and feedback loops in information ecosystems.
An oligopoly is taking shape in enterprise AI: OpenAI still leads, but Anthropic is catching up fast while Microsoft dominates applications.
Physical prompt injection: road signs and other real-world text can steer autonomous cars/drones if systems treat external instructions as trusted.
A technical walkthrough of how mobile carriers can infer GPS/GNSS location—highlighting a privacy surface that’s easy to underestimate.
Assistive technology is expensive, and many people with disabilities live on fixed incomes.
David Silver, one of the key AI researchers behind landmark Deepmind projects like AlphaGo and AlphaZero, is leaving the Google subsidiary…
A year-end reflection on how generative AI is changing Wikipedia editing—workflows, quality control, and community norms.
arXiv:2601.21218v1 Announce Type: new Abstract: Large language models (LLMs) are highly capable of answering questions, but they are often…
US authorities reportedly investigated claims that WhatsApp chats aren’t private—another reminder that privacy depends on more than just ‘end-to-end encryption.’
Finland is considering a ban on youth social media, framing current exposure as an uncontrolled ‘human experiment.’
arXiv:2601.20975v1 Announce Type: new Abstract: We introduce DeepSearchQA, a 900-prompt benchmark for evaluating agents on difficult multi-…
arXiv:2601.21132v1 Announce Type: new Abstract: I demonstrate that large language models can infer ethnicity from names with accuracy excee…
arXiv:2601.21235v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly deployed in high-stakes domains, where rare b…
Full list — AI tools & model updates (1)
Developers who learn new programming skills with AI assistance score significantly worse on knowledge tests, according to a new Anthropic s…