Gary Marcus: OpenAI probably can’t make ends meet. That’s where you come in. (Nov. 5, 2025) Sam Altman got defensive when questioned about the company’s significant debt and questionable revenue, hinting at future growth without concrete details. Now, OpenAI’s CFO is suggesting that the U.S. government should subsidize AI development to compete with China, effectively asking taxpayers to bail out the company’s financial risks. (Altman’s reply is below)
Sam Altman: Government Guarantees (Nov. 7, 2025) Altman doesn’t want governmental guarantees for data centers, and he remains bullish on OpenAI’s prospects to generate revenue over the next 8 years. I still expect to see ChatGPT prices rise significantly in the next 3-4 years, even for consumers.
WSJ: I Loved Being Social. Then I Started Talking to a Chatbot. (Nov. 2, 2025) An extrovert details how consistent interaction with an AI chatbot, intended for productivity and brainstorming, led to social isolation and a decline in her ability to connect with people. “Indeed, on the days when I talked to AI for a few hours, I was all talked out by the evening, with neither the craving nor the energy (nor the practical need) to have an extended human conversation.”
Inside Higher Ed: Higher Ed Tech Leaders Pursue Consolidation and Savings (Oct 31, 2025) Higher education technology leaders expressed caution about investing in new AI technologies due to budget constraints and staffing limitations, emphasizing the need for clear ROI and strategic planning.
arXiv.org: A Definition of AGI (Oct 25, 2025) Exploring a quantifiable framework for defining and measuring Artificial General Intelligence (AGI) based on the cognitive abilities of a well-educated adult, using the Cattell-Horn-Carroll theory of human cognition. The framework assesses AI systems across ten core cognitive domains using adapted psychometric batteries, revealing that current models have strengths in knowledge but weaknesses in areas like long-term memory. The resulting AGI scores, like GPT-5 at 57%, highlight the progress and remaining gap towards achieving true AGI.
Simon Willison: A quote from Nathan Lambert (Nov 6, 2025) Chinese AI labs like DeepSeek, Qwen, and Kimi are quickly gaining recognition and catching up to the performance of leading models. This shift indicates a growing concentration of cutting-edge AI development in China.
Google Research: Exploring a space-based, scalable AI infrastructure system design (Nov. 4, 2025) Google’s Project Suncatcher aims to build a space-based AI infrastructure using solar-powered satellites equipped with Google TPUs and high-bandwidth optical links. This project addresses key challenges like inter-satellite communication, orbital dynamics, and radiation tolerance, with plans for a learning mission involving prototype satellites by 2027.
Lukew: AI Has Flipped Software Development (Jul 27, 2025) AI coding agents are drastically accelerating software development, allowing engineers to build features much faster than designers can refine them, effectively flipping the traditional design-to-build process.
GenAI Image Showdown: GenAI Image Showdown (Nov 30, -0001) A comparison of state-of-the-art generative image models, evaluating their performance on specific prompts designed to test adherence to complex instructions and concepts. T