
EP175: Can children learn from an AI avatar? | How will they crackdown on child P*ornagraphy? | AI whirlwind updates!
The hosts dive into the Ministry of Education's launch of a new digital learning platform for secondary students, noting that fancy portals are great right up until the internet decides not to work. They question whether these shiny new online tools will actually bridge the learning gap or just give school tech support a perpetual headache. Article: https://www.ttt.live/education-ministry-launches-new-digital-learning-platform-to-support-secondary-school-students/ The government plans to crackdown on child p*ography, but how exactly do the plan to carry this out? And how effective is it in tracking down Telegram messages? Is this just all 'hot-air'? Article: https://www.guardian.co.tt/news/govt-to-deploy-new-technology-in-crackdown-on-child-pornography-6.2.2654872.3a124bc84f So, Anthropic is now sneaking invisible text watermarks into Claude outputs globally to appease EU regulators, using a secret key to subtly bias word choices without actually ruining your text quality. The real kicker for users is that while simple copy-pasting keeps the mark alive, substantive human editing will dilute it, meaning even a light proofread could leave your work tagged as AI-assisted. Article: https://www.reuters.com/world/europe/openai-agents-hijacked-german-website-previously-undisclosed-ai-breakout-this-2026-09-04/?utm_medium=Social&utm_source=twitter https://youtu.be/1QNsdr-Qx_I?si=76pGLumUguuFhwjF



