![[The AI Show Episode 143]: ChatGPT Revenue Surge, New AGI Timelines, Amazon’s AI Agent, Claude for Education, Model Context Protocol & LLMs Pass the Turing Test](https://www.marketingaiinstitute.com/hubfs/ep%20143%20cover.png)
![Why modern UI is rarely written in C++? [closed]](https://cdn.sstatic.net/Sites/softwareengineering/Img/apple-touch-icon@2.png?v=1ef7363febba)
![From Accountant to Data Engineer with Alyson La [Podcast #168]](https://cdn.hashnode.com/res/hashnode/image/upload/v1744420903260/fae4b593-d653-41eb-b70b-031591aa2f35.png?#)
.png?#)
![Pixel 9a vs. Pixel 8a: The A-series is all grown up [Video]](https://i0.wp.com/9to5google.com/wp-content/uploads/sites/4/2025/04/Pixel-8a-and-Pixel-9a-wide.jpg?resize=1200%2C628&quality=82&strip=all&ssl=1){kind=tracking}
![Apple Posts Full First Episode of 'Your Friends & Neighbors' on YouTube [Video]](https://www.iclarified.com/images/news/96990/96990/96990-640.jpg)
![Apple May Implement Global iPhone Price Increases to Mitigate Tariff Impacts [Report]](https://www.iclarified.com/images/news/96987/96987/96987-640.jpg)
Weekly Review 11 April 2025
Some interesting links that I Tweeted about in the last week (I also post these on Mastodon, Threads, Newsmast, and Bluesky):Human educators are still better than AI chatbots, but the gap is narrowing: https://www.datasciencecentral.com/how-do-ai-chatbots-compare-to-human-educators/I don't think that relying on AI to fact check news is necessarily a good idea: https://dataconomy.com/2025/04/02/broadcast-journalists-ai-real-time-fact-checking/More evidence that AI are being trained on copyrighted works without permission: https://dataconomy.com/2025/04/02/openai-might-have-trained-its-ai-on-stolen-books/The adaptive abilities of AI chatbots are making them essential for business: https://dataconomy.com/2025/04/03/are-ai-agents-the-new-must-have/AI training data bots are now scraping Wikipedia so much that the entire site is under threat: https://arstechnica.com/information-technology/2025/04/ai-bots-strain-wikimedia-as-bandwidth-surges-50/Applications of AI in cinema: https://www.datasciencecentral.com/creative-applications-of-ai-in-cinema/All the ways artificial general intelligence could ruin the world, and how to stop it doing so: https://arstechnica.com/ai/2025/04/google-deepmind-releases-its-plan-to-keep-agi-from-running-wild/So if you want to hack Gemini AI, use Gemini AI? https://arstechnica.com/security/2025/03/geminiding-hackers-can-deliver-more-potent-attacks-with-a-helping-hand-from-gemini/A list of the current lea AI models and what they're good for: https://techcrunch.com/2025/03/30/the-hottest-ai-models-what-they-do-and-how-to-use-them/The race to get AI to college students: https://www.theverge.com/ai-artificial-intelligence/641193/openai-anthropic-education-tool-collegeWhy AI tend to hallucinate: https://arstechnica.com/ai/2025/03/why-do-llms-make-stuff-up-new-research-peers-under-the-hood/It's OK for AI to be stupid. But when stupid people use what AI says to set policy, we have real problems: https://arstechnica.com/tech-policy/2025/04/critics-suspect-trumps-weird-tariff-math-came-from-chatbots/Do AI generated twins of models mean less work for models in the future? Maybe not for live fashion shows, but for everything else? https://www.theguardian.com/fashion/2025/mar/30/fashion-models-ai-job-lossesAI researchers in the USA are more optimistic about AI than the general public: https://arstechnica.com/tech-policy/2025/04/survey-americans-fear-ai-will-hurt-them-experts-expect-the-opposite/While Google's AI search is working reasonably well, they are struggling to make money from it: https://www.computerworld.com/article/3951907/the-future-of-ai-search-is-googles-to-lose.htmlAn open standard to connect AI: https://arstechnica.com/information-technology/2025/04/mcp-the-new-usb-c-for-ai-thats-bringing-fierce-rivals-together/OpenAI is not following copyright rules when it comes to scraping AI training data: https://www.theregister.com/2025/04/03/openai_copyright_bypass/If you want to future-proof your AI, you need to keep your data in good order: https://www.informationweek.com/machine-learning-ai/when-it-comes-to-futureproofing-ai-it-s-all-about-the-dataBuilding large AI is now so expensive, a lot of providers are going to get wiped out. In other words, the AI bubble is close to bursting: https://www.theregister.com/2025/03/31/llm_providers_extinction/Biased data results in biased models. This is a fundamental principle of AI yet it still creeps into models everywhere: https://dataconomy.com/2025/04/03/the-llm-wears-prada-why-ai-still-shops-in-stereotypes/Anthropic's Claude AI is being introduced to campuses in the USA, offering a 'Learning mode' that guides students through reasoning rather than providing direct answers: https://dataconomy.com/2025/04/03/claude-ai-is-coming-to-a-campus-near-you/I feel that OpenAI's valuation being greater than McDonald's is a pretty strong signal that the AI bubble is peaking: https://dataconomy.com/2025/04/02/openai-is-now-worth-more-than-mcdonalds/AI movie generators are improving, now they can maintain continuity between scenes: https://www.theverge.com/news/640821/runway-gen-4-artificial-intelligence-video-generator-filmmakingThe difference between training and inference in AI: https://www.kdnuggets.com/training-vs-inference-the-ultimate-alliance
Some interesting links that I Tweeted about in the last week (I also post these on Mastodon, Threads, Newsmast, and Bluesky):
- Human educators are still better than AI chatbots, but the gap is narrowing: https://www.datasciencecentral.com/how-do-ai-chatbots-compare-to-human-educators/
- I don't think that relying on AI to fact check news is necessarily a good idea: https://dataconomy.com/2025/04/02/broadcast-journalists-ai-real-time-fact-checking/
- More evidence that AI are being trained on copyrighted works without permission: https://dataconomy.com/2025/04/02/openai-might-have-trained-its-ai-on-stolen-books/
- The adaptive abilities of AI chatbots are making them essential for business: https://dataconomy.com/2025/04/03/are-ai-agents-the-new-must-have/
- AI training data bots are now scraping Wikipedia so much that the entire site is under threat: https://arstechnica.com/information-technology/2025/04/ai-bots-strain-wikimedia-as-bandwidth-surges-50/
- Applications of AI in cinema: https://www.datasciencecentral.com/creative-applications-of-ai-in-cinema/
- All the ways artificial general intelligence could ruin the world, and how to stop it doing so: https://arstechnica.com/ai/2025/04/google-deepmind-releases-its-plan-to-keep-agi-from-running-wild/
- So if you want to hack Gemini AI, use Gemini AI? https://arstechnica.com/security/2025/03/geminiding-hackers-can-deliver-more-potent-attacks-with-a-helping-hand-from-gemini/
- A list of the current lea AI models and what they're good for: https://techcrunch.com/2025/03/30/the-hottest-ai-models-what-they-do-and-how-to-use-them/
- The race to get AI to college students: https://www.theverge.com/ai-artificial-intelligence/641193/openai-anthropic-education-tool-college
- Why AI tend to hallucinate: https://arstechnica.com/ai/2025/03/why-do-llms-make-stuff-up-new-research-peers-under-the-hood/
- It's OK for AI to be stupid. But when stupid people use what AI says to set policy, we have real problems: https://arstechnica.com/tech-policy/2025/04/critics-suspect-trumps-weird-tariff-math-came-from-chatbots/
- Do AI generated twins of models mean less work for models in the future? Maybe not for live fashion shows, but for everything else? https://www.theguardian.com/fashion/2025/mar/30/fashion-models-ai-job-losses
- AI researchers in the USA are more optimistic about AI than the general public: https://arstechnica.com/tech-policy/2025/04/survey-americans-fear-ai-will-hurt-them-experts-expect-the-opposite/
- While Google's AI search is working reasonably well, they are struggling to make money from it: https://www.computerworld.com/article/3951907/the-future-of-ai-search-is-googles-to-lose.html
- An open standard to connect AI: https://arstechnica.com/information-technology/2025/04/mcp-the-new-usb-c-for-ai-thats-bringing-fierce-rivals-together/
- OpenAI is not following copyright rules when it comes to scraping AI training data: https://www.theregister.com/2025/04/03/openai_copyright_bypass/
- If you want to future-proof your AI, you need to keep your data in good order: https://www.informationweek.com/machine-learning-ai/when-it-comes-to-futureproofing-ai-it-s-all-about-the-data
- Building large AI is now so expensive, a lot of providers are going to get wiped out. In other words, the AI bubble is close to bursting: https://www.theregister.com/2025/03/31/llm_providers_extinction/
- Biased data results in biased models. This is a fundamental principle of AI yet it still creeps into models everywhere: https://dataconomy.com/2025/04/03/the-llm-wears-prada-why-ai-still-shops-in-stereotypes/
- Anthropic's Claude AI is being introduced to campuses in the USA, offering a 'Learning mode' that guides students through reasoning rather than providing direct answers: https://dataconomy.com/2025/04/03/claude-ai-is-coming-to-a-campus-near-you/
- I feel that OpenAI's valuation being greater than McDonald's is a pretty strong signal that the AI bubble is peaking: https://dataconomy.com/2025/04/02/openai-is-now-worth-more-than-mcdonalds/
- AI movie generators are improving, now they can maintain continuity between scenes: https://www.theverge.com/news/640821/runway-gen-4-artificial-intelligence-video-generator-filmmaking
- The difference between training and inference in AI: https://www.kdnuggets.com/training-vs-inference-the-ultimate-alliance