Learning Curved #1
AI Literacy, trends in edtech, brain rot videos
Generative AI is going to be one of the most transformative technologies of our lifetimes. Two and a half years since ChatGPT came into existence, the advances in large language models have been astonishingly huge. Yet, we are nowhere close to product market fit at a large scale except coding and a few areas. We are in the horseless carriage phase of the technology. That does not mean that we will be waiting for all of it to get sorted out as by its nature it would require significant experiments and iterations.
The impact of this technology on education has been heralded right from its start as successive generations of models improved from writing a few lines of coherent text to performing better than most humans in tests like AP Biology or math olympiad or advanced mathematics.
I aim to use this substack to bring to attention notable advances, experiments, research and products that in hindsight would look like the milestones to a long journey towards reimagining education. With the caveat that ‘reimagining education’ means a thousand different things based on who you ask.
With no further ado, here are a few stories that caught my attention in the last couple of weeks.
AI Literacy framework and Gov.UK guidelines
Two chunky reports came out recently trying to make sense of it all for teachers, education leaders and policy makers.
A good way to interact with these reports now is to simply put them into a LLM and start asking questions. NotebookLM does a specially good job including creating audio summaries. But you can choose any LLM that you use to get started.
The first one is the AI Literacy framework developed by OECD, the European Commission and with support from code.org and a large number of experts from around the world. There is a lot in this report as it tries to define what AI literacy is and does it by breaking it into 4 domains - Engaging with AI, Creating with AI, Managing AI and Designing AI.
Gov.UK’s guidelines are digestible powerpoints and yes you can do the same. Throw them in a LLM and ask lots of questions.
Why this matters: We are likely to see a lot of these attempts now as we move to the next phase from resisting AI to finding ways to work with AI. Perhaps in hindsight these efforts will look unwieldy, maybe even impractical but that is the nature of finding what truly is important and what works.
A16z podcast on edtech
Folks at a16z just did this podcast and here are some things that surprised me.
Big-city districts that once blocked ChatGPT have reversed course, and ≈80 % of US districts now have dedicated “generative-AI” working groups. It does not mean that they are necessarily buying new products at scale but this is a big shift from a couple of years ago when districts were banning ChatGPT.
Less surprising is that universities are out in front of the adoption curve. Special editions of Claude, OpenAI’s EDU pilots, mandatory AI use at Ohio State, etc.
Early paying power-users are … teachers. They shoulder the drudgery of grading, feedback, lesson-planning and are happy to spend on tools such as Magic School (reportedly tried by ~50 % of U.S. teachers).
Why this matters: It is important to keep an eye on developing trends coming from a product market perspective best exemplified by ideas and ventures being funded by private capital. This is an orthogonal approach to the one I described above about creating frameworks and guidelines. I think in the end both will matter.
Breadcrumbs to an evolving future
A couple of other mentions in the above podcast.
Brain rot videos: Tiktok/IG style videos to explain math concepts. Here’s one on IG called onlocklearning, which claims to be the easiest way to revise for your AP, A-levels, IB and JEE(?). I started watching these videos and whether these would amount to something serious or not (engagement levels seem to be high), they do a great job of meeting where many learners are. Have a look despite your skepticism.
Alpha school: I had come across Alpha school a few months ago and was intrigued by their promise to condense academic work in 2 hours leaving the rest of the time for kids to explore projects, interests etc. Most students will gladly take the value prop. Then I saw this tweet from Austen Allred earlier this month. Read the tweet thread in full here.
Why this matters: It underscores the point that we are very very early in figuring out how we transition from a horseless carriage phase. If you are involved in education, the game has just begun.