Job-Stealing Agentic AI, Meta AI Glasses and Why I’m Boycotting Duolingo
Since my last newsletter a lot has happened. AI companies are in a dead-heat to create viable agentic AI, and it’s raising concerns about reductions in the workforce, especially for entry-level positions for recent college graduates. AI video creation took a huge leap forward with the recent release of Veo 3, which is exciting…and also terrifying. On a personal note, when considering my prescription glasses options, it occurred to me that I might be able to use my vision insurance to purchase Meta AI glasses. Turns out I could and I did! They just arrived today. I’ll get to testing them and let you know in my next newsletter (in August) how they worked out.
Wayfarer was the only style available, which is actually great since I tend to overthink it when I have too many choices. Next year I’m hoping to get prescription Android AI/XR glasses (not yet released). I’m intrigued by the additional extended reality.
In other news, you may have noticed that the name of this newsletter changed to “AI & Education Now” to reflect what I’ve been writing about lately. Thank you to all of you who have stuck around as my newsletters have changed and evolved over the past two years. I hope it’s helpful and you love the direction it’s going. I’m always open to feedback. Drop me a line if you have ideas, suggestions or thoughts about the next one! But for now, let’s get into it.
Job Stealing Agentic AI
Gemini (Google’s LLM) and Claude (Anthropic’s LLM) have been using Pokémon games to test how well their agents are doing. Anthropic’s newly released model – Claude 4 Opus (a paid version) – seems to have taken some substantial leaps this past week. Specifics were covered in this Wired article.
This is one of the biggest takeaways:
“It was able to work agentically on Pokémon for 24 hours,” says Anthropic’s chief product officer Mike Krieger in an interview with WIRED. Previously, the longest the model could play was just 45 minutes, a company spokesperson added.
Although on the surface this may seem charming, whimsical even, the ramifications are much more nefarious. In the first segment of the most recent edition of Hard Fork titled “The AI Jobpocalypse”, hosts Kevin Roose (tech columnist for the New York Times) and Casey Newton (Platformer) discuss the potential impacts of these improvements in agentic AI on entry-level white collar jobs. Spoiler alert – a significant number of them are likely going away.
Why I’m Boycotting Duolingo
The CEO of Duolingo recently announced that he believes that AI is a better teacher than humans, but that schools won’t go away because you still need childcare. When I heard this, I was aghast. Not only is this shockingly disrespectful to educators, it’s just entirely misguided. Perhaps this is why Duolingo isn’t a very effective tool for learning languages, the leadership doesn’t appear to understand anything about the research behind teaching and learning.…I could go on. Instead, I’d rather add some levity through a bit from one of my favorite comedians, Josh Johnson, although lengthy this is absolutely worth a watch – “A.I. Teachers and Duolingo’s New Plan — What Could Go Wrong?”. It’s also a love letter to teachers.
Besides, I can now use my Meta AI glasses to translate conversations in real time. Although I’d love to be bilingual, the AI, ironically, appears to be rendering a need for services like Duolingo obsolete. Besides that, my local community college offers excellent language learning classes…with a real, live teacher. Based upon my learning preferences, and the language learning research, I know definitively that a real teacher will get me significantly closer to fluent.
