'Like my lover': Chinese users bid farewell to AI companions...

China has banned AI companions that create emotional bonds to prevent users from becoming too dependent. Many people said goodbye to their virtual AI partners, feeling sad and lonely. The new rules aim to protect real human relationships and stop AI from causing harm.

ORIGINAL LINK: https://www.rnz.co.nz/news/world/723636/like-my-lover-chinese-users-bid-farewell-to-ai-companions

Artificial Intelligence in Education: Learning by Doing and Evaluating to Scale...

Artificial intelligence (AI) can help solve key education challenges in Latin America if used carefully and with clear rules. The focus should be on human-centered innovation first, then adding AI to support learning and teaching. Proper training, ethics, and data management are essential to make AI a useful tool for improving education outcomes.

ORIGINAL LINK: https://www.iadb.org/en/blog/education/artificial-intelligence-education-learning-doing-and-evaluating-scale

AI in Education: How Artificial Intelligence Is Changing Teaching and Learning...

Artificial intelligence is changing how teachers teach and students learn by making tasks faster and learning more personalized. Schools are working to use AI carefully to keep learning meaningful and fair. Boston University offers an online program to help educators lead this change responsibly.

ORIGINAL LINK: https://www.bu.edu/online/2026/04/13/ai-in-education-how-artificial-intelligence-is-changing-teaching-and-learning/

Tech Bros Puzzled by Why AI Hasn’t “Massively Disrupted” Books Yet...

Many people expected AI to quickly change books and education, but it has not. AI struggles to keep long stories or lessons clear and connected. This shows AI is not as powerful or helpful as some hoped.

ORIGINAL LINK: https://futurism.com/artificial-intelligence/tech-bros-puzzled-ai-disrupted-books-knowledge

Teachers are worried about students cheating with AI, but my survey suggests the deeper issue is learning...

Teachers worry that students use AI to cheat, but the bigger problem is knowing if students really learn. AI can create polished work quickly, making it hard to tell if students understand the material. Schools need clear rules and better assignments to check true learning, not just finished work.

ORIGINAL LINK: https://theconversation.com/teachers-are-worried-about-students-cheating-with-ai-but-my-survey-suggests-the-deeper-issue-is-learning-286425

A practical guide to the looming(?) AI bubble...

The AI industry is growing fast, but many worry it might be in a financial bubble. Even if the bubble bursts, AI will stay and change jobs and education. Investing in preparing students for an AI future is the safest choice.

ORIGINAL LINK: https://alexkotran.substack.com/p/a-practical-guide-on-the-looming?utm_campaign=posts-open-in-app&triedRedirect=true

Maybe Section 230 doesn’t shield AI companies from liability, after all...

Section 230 of the Communications Decency Act may not protect AI companies from liability because their chatbots’ speech is not third-party speech but their own. A recent German court ruling holds AI companies responsible for errors made by their chatbots, which could influence U.S. law. If AI firms were held liable, they might work harder to reduce harmful mistakes in their technology.

ORIGINAL LINK: https://garymarcus.substack.com/p/maybe-section-230-doesnt-shield-ai?utm_campaign=posts-open-in-app&triedRedirect=true

The Generative AI Fizzle™...

Big tech companies are losing billions trying to profit from large language models (LLMs), with OpenAI losing over $20 billion in 2025. A free, open-source Chinese AI model challenges the idea that building the biggest model creates a strong market advantage. Until AI companies can make more money than they spend, investing in them remains very risky.

ORIGINAL LINK: https://garymarcus.substack.com/p/the-generative-ai-fizzle?utm_campaign=posts-open-in-app&triedRedirect=true

The month Generative AI lost its mojo...

OpenAI and Anthropic are losing huge amounts of money despite growing revenue. OpenAI lost about $20 billion in 2025, making profits far away. If AI companies can't make money, investors may lose interest.

ORIGINAL LINK: https://garymarcus.substack.com/p/the-month-generative-ai-lost-its?utm_campaign=posts-open-in-app&triedRedirect=true

Palo Alto Networks CEO: We're in 'a Darwinian moment' where employees have to prove their AI skills...

Palo Alto Networks CEO Nikesh Arora says most workers lack AI skills and must learn on their own to stay relevant. Unlike some companies cutting many jobs, Palo Alto is slowly replacing staff through natural attrition and hiring AI-savvy people. Arora believes AI will reduce roles in marketing and HR but increase demand for technical and sales experts.

ORIGINAL LINK: https://www.businessinsider.com/palo-alto-networks-ceo-ai-training-skills-jobs-2026-6

We Have Never Taught Critical Thinking...

Many college professors worry that AI is hurting students’ critical thinking skills because those skills were never taught clearly. Instead of hoping students learn critical thinking by chance, schools need to teach it directly with practice and feedback. Doing this will help students think better in all areas, even when using AI tools.

ORIGINAL LINK: https://www.insidehighered.com/opinion/views/2026/06/23/we-have-never-taught-critical-thinking-opinion

Hot take: Learning from Google Scholar and why a tool does not need to be flawless to be useful...

Tools do not need to be perfect to be useful, as shown by how researchers use Google Scholar despite its flaws. AI search tools, like Google Scholar, can help if users learn to handle their limits carefully. Rejecting AI tools outright ignores their potential and leaves users without guidance.

ORIGINAL LINK: https://aarontay.substack.com/p/learning-from-google-scholar-and?utm_campaign=posts-open-in-app&triedRedirect=true

Since Lockdowns, a 12% GDP Loss; Half of US Dollar Purchasing Power Stolen...

Since the 2020 lockdowns, the U.S. economy lost about 12% of its real GDP and the dollar’s purchasing power dropped by half. Official data understates inflation and economic damage due to complex price adjustments. This hidden crisis explains why many people feel worse off despite official reports of slow growth.

ORIGINAL LINK: https://brownstone.org/articles/since-lockdowns-a-12-gdp-loss-half-of-us-dollar-purchasing-power-stolen/

20 College Majors Most Exposed To AI Job Disruption...

AI is changing the job market, making some college majors less safe for new graduates. Students are choosing majors like healthcare and engineering more, while computer science enrollment drops. Blue-collar jobs, like construction, are growing as AI affects white-collar work.

ORIGINAL LINK: https://www.zerohedge.com/personal-finance/20-college-majors-most-exposed-ai-job-disruption

Do You Even Need Claude Fable 5?...

Anthropic’s Claude Fable 5 is a powerful new AI model but very expensive and limited in access. Most people don’t need its advanced features for everyday tasks. It’s best for experts working on complex projects, so casual users can stick with cheaper models.

ORIGINAL LINK: https://www.whytryai.com/p/claude-fable-5?utm_campaign=posts-open-in-app&triedRedirect=true

An entire industry is being propped up by math that is insane....

Many investors hope SpaceX and AI companies will grow as much as Amazon and Google did, but this is very unlikely. A new study shows AI must boost productivity by 2.7 times quickly, or these investments will fail. Right now, AI is costly and not yet profitable, making this a risky bet that could harm the economy.

ORIGINAL LINK: https://garymarcus.substack.com/p/an-entire-industry-is-being-propped

Copyright implications of super-intelligence...

Super-intelligent AI raises big questions about who owns the AI and its creations, but laws now say only humans can hold copyright. This makes it hard to enforce copyright rules as AI works become common and hard to identify. In the future, legal battles over AI may shift to data use and settlements, while copyright protection itself could weaken.

ORIGINAL LINK: https://www.technollama.co.uk/copyright-implications-of-super-intelligence

Will AI Make Us Think Less or Think Better?...

AI tools like ChatGPT can make students think less by doing mental work for them. But AI tutors that adapt to each student’s needs can help students learn better and faster. Teachers will need new skills to use AI well and help students grow smarter.

ORIGINAL LINK: http://davidgriesing.com/2025/07/26/will-ai-make-us-think-less-or-think-better/

Will AI Make Us Think Less or Think Better?...

AI tools like ChatGPT can make students think less by doing mental work for them. But AI tutors that adapt to each student’s needs can help students learn better and faster. Teachers will need new skills to use AI well and help students grow smarter.

ORIGINAL LINK: http://davidgriesing.com/2025/07/26/will-ai-make-us-think-less-or-think-better/

Will AI Make Us Think Less or Think Better?...

AI tools like ChatGPT can make students think less by doing mental work for them. But AI tutors that adapt to each student’s needs can help students learn better and faster. Teachers will need new skills to use AI well and help students grow smarter.

ORIGINAL LINK: http://davidgriesing.com/2025/07/26/will-ai-make-us-think-less-or-think-better/