1. [The Future of Work Is About Doing Better - Not Less - Rhea Purohit](https://every.to/learning-curve/the-future-of-work-isn-t-about-doing-less-it-s-about-doing-better) 1. Using AI in writing is not cheating, it's only a way of better using one's time on different aspects of what it takes to create good writing (eg: strategy, humour). 2. I personally try to use AI for converting my bulleted thoughts into cohesive language. This tends to be massively helpful as it frees me to do the thinking and experimentation with ideas 2. [AGI's Paradox: When Perfect Answers Aren't Enough - Cassie Kozyrkov](https://kozyrkov.medium.com/agis-paradox-when-perfect-answers-aren-t-enough-9770e4d1717a) 1. Having an all powerful AI means nothing if the user does not understand the question they are looking to tackle with AI. Human expertise is still very much a necessary requirement with AI 2. The author calls AI a "thoughtlessness enabler"; the better the AI gets, the lesser we need to "think". This could be dangerous as we do not want to outsource our thinking to black box models 3. Quote - "This is why the most critical skill in today's AI-driven world isn't computational prowess; it's the ability to think deeply and clearly about the problems you're trying to solve." 3. [AGI could drive wages below subsistence level](https://epochai.substack.com/p/agi-could-drive-wages-below-subsistence) 1. The author uses the [Cobb-Douglas production function](https://en.wikipedia.org/wiki/Cobb%E2%80%93Douglas_production_function) to explain his argument 2. In the short term, deployment of AGI could help increase productivity, thus revenue and therefore wages 3. However, in the long term, the amount of productivity increase would be minimal, coz everyone has access to an AGI. This could reduce wages below the baseline threshold needed for human beings to survive 4. As wages are set based on requirements of the "labor", if AGI replaces human beings, then this labor would become cheaper for corporations. Wages are then determined not by the energy needs of humans, but by the energy needs of AI (which could be cheaper in the future) 5. Author predicts effects of this by around 2045 (which I personally am not sure will happen) 6. Some thoughts I have 1. Would be interesting to see how this varies from sector to sector, country to country 2. Also, I realise there may be different production functions too 4. [The AI Black Swan: How capitalism's obsession with profit will trigger the next collapse](https://fafi25.substack.com/p/the-ai-black-swan-how-capitalisms?r=6awgq&utm_medium=ios&triedRedirect=true) 1. Economy - If more jobs get automated, and lesser jobs get created; then the middle class will struggle and purchasing power falls; this will destabilise the economy 2. Ecology 3. Energy 4. Hidden dimensions 1. Cultural dissolution 2. Knowledge degradation 3. Social cohesion 5. [AI in 2025: Building Blocks Firmly in Place - David Cahn (Sequoia)](https://www.sequoiacap.com/article/ai-in-2025/) 1. Clearer differentiation between LLM providers as each now has their own distinct abilities 1. Google - Vertical Integration; owns everything in the supply chain 2. OpenAI - Brand; ChatGPT is more famous than any other LLM (similar to how "Googling" became synonymous with internet search, "ChatGPT" is becoming synonymous with AI use) 3. Anthropic - Talent 4. xAI - Data Centre Construction (see [Steel, servers and power](https://www.sequoiacap.com/article/steel-servers-and-power/)) 5. Meta - Open source 2. AI search might proliferate the most in 2025 as an AI use case 3. ROI (Return on Investment) looks uncertain still and CapEx (Capital Expenditure) is expected to stabilise (see [AI's $600B question](https://www.sequoiacap.com/article/ais-600b-question/))