(2025-08-03) Strong How To Give Everyone A Great Education For Free
Michael Strong: How to Give Everyone a Great Education for Free. After graduating from one of my schools, Cade Summers got a few friends together on a Discord server and they read and discussed Napoleon Hill’s classic Think and Grow Rich for a few months. Then they started to work through Harvard CS50 together and finished it. With an official Harvard certificate of successfully completing Harvard CS50 in hand, at eighteen Cade was able to get an entry level job at a software company in Austin and gradually work his way up to a software developer position.
In general Massive Open Online Courses (MOOCs) have low completion rates (single digits) but when one considers those who say they intend to complete a course, as opposed to casual drop-ins, then completion rates rise to 25-66%. But Harvard CS50 has a completion rate of only 1%, even among those who plan on completing the course.
The biggest factor that is neglected in edtech is that most learning is social. Human beings are above all social creatures. I suspect that many of the tech geeks who are inclined to learn from MOOCs (or now AI) are the single digit percentage of the population who happen to be motivated and capable of learning in isolation. (autodidact)
Over the years I’ve spoken to a number of edtech developers who don’t disagree that learning is social, but none of them have made a sustained commitment to nurture a high quality social environment
AI continues to become ever more sophisticated, I see human relationships, including mentoring, coaching, and engaging with each other, as our lasting value. Henrik Karlsson captures something that is likely to remain importantly human for some time:
"When I meet a person who is truly, profoundly themselves, I sometimes think of a letter Charles Darwin sent to Joseph Hooker in January 1862 after receiving a package of orchids.
The moth and the orchid must have evolved in dialogue. This is what I infer when I see someone who is comfortable in their unique strangeness, too. There probably exists someone who enabled that evolution of personality. A parent, a friend group, a spouse. It is rare for people to come into themselves if no one is excited and curious about their core, their potential. We need someone who gives us space to unfold.
Even before AI tutors, educational content had become a commodity. The entire world of edtech is focused on “How do we teach X.” AI has simply made it more obvious that the task of “Teaching X” is banal. It is already free for everyone. Now all that matters is having the motivation to learn.
Alpha School has recently received a lot of attention for its system of external incentives, giving kids prizes for learning. They also hype “AI learning,” but as the article makes clear, the incentive system is perhaps more distinctive. (extrinsic motivation)
At present most of their education content simply uses off the shelf, widely available apps such as IXL (mediocre as Brandon Hendrickson of the excellent “Science is Weird” argues here) and Math Academy (excellent).
Zvi Mowshowitz provides a helpful analysis of exactly how they have designed their external incentive system appropriately to encourage learning.
I’ve spent many years at Montessori schools where trained Montessori educators are adamently against using incentives of any kind
I’m not against the use of external incentives, and acknowledge that the Alpha external incentive approach will be easier to scale more quickly than are systems such as Montessori that rely more on developing internal motivation in a systemic way.
Because I see standard schooling as so damaging to so many human beings, I celebrate Alpha’s aggressive, well-funded, and actionable campaign to promote their alternative model.
My hope is that in the next 20-30 years, high agency learning models such as Montessori, Acton, Alpha, Prenda, Primer, and my Socratic approach, The Socratic Experience, will have 20-30% market share as creative destruction gradually puts the old system out of business.
But ultimately while using incentives from time to time is fine, I’m glad that I raised my children without such incentives and my heart is completely on the side of Montessori. Richard Ryan and Edward Deci’s Self-Determination Theory is a well developed theory of motivation aligned with Maria Montessori’s original insights (and program design).
“Relatedness” is the remaining ingredient
Since edtech/AI can trivially provide personalized learning within a child’s zone of proximal development, thereby allowing children to experience competence by doing work at their level. In such an environment, autonomy should be the norm.
Thus the remaining question regarding, “How do we scale education for free?,” at least in the absence of external incentive systems such as those used by Alpha, is about human relationships, not content transmission.
Creating Scaleable Cultures of Learning
Many people learn for free online all the time. Often that includes a substantial social component. For decades there have been various online forums
But for the most part, these are made up of adults with particular interests finding each other.
One can think of these communities as de facto apprenticeship communities
online conversation and dispute itself is a powerful learning opportunity.
Because of the neglect of the social/human relationships side of edtech, we haven’t yet been able to replace schooling by low cost scalable human networks in K12. At the lower grades it may not be possible, so let’s focus on high school for now
The closest anyone has gotten is Andrew Sachs’ Nobel Navigators, which has on the order of 10,000 mostly high school age students around the world teaching each other a combination of tech skills and social skills. In some ways he is recreating the Bell Lancaster Method.
Recess.gg, founded by Ben Somers, is another potential partner. Ben is developing his own AI learning system designed to optimize existing edtech platforms (which is analogous to how Alpha is using AI). But as a firm advocate of self-directed learning, what is most interesting to me is how he is thinking about community formation
key functionality for a scalable virtual learning community is well-considered parental controls both on content and relationships, and Ben has thought this out more than anyone I’ve met.
Qbix.com, founded by Greg Magarshak, is an open source community platform that could be customized. It includes eye-tracking and could easily add engagement tools based on facial expression or vocal expression. Engagement tools will be useful to evaluate at scale when and where children are engaged, and eye tracking allows the system to know what is engaging the child
Beth Porter’s Riff Analytics, which measured interruptions, dominance, and influence in virtual conversations, was also a potential partner. But they were acquired by Esme Learning, which now seems to have vanished
Going in a different direction, Alan Tang’s Collaboration Laboratory specializes in game-like activities designed to train students in more collaborative activities.
Pulling this Altogether to Create Scalable Communities of Autodidacts
Tomorrow I will be leading a group of middle school students in a Socratic discussion of an article on “Dogs are the Reason Neanderthals Disappeared.” It summarizes Pat Shipman’s thesis that wolves and humans together became a “super predator” that led to the extinction of the Neanderthals.
I predict that these 11-14 year olds will have a good time spending 90 minutes in the middle of the summer talking about dogs and humans being a super predator. They will have fun learning together.
Over the decades of leading thousands of hours of intellectual conversations with children, I’ve come to regard the critical factor to be a set of social norms around respectful dialogue
At St. John’s College, popularly described as a “Great Books” college, all classes are taught Socratically. The “tutors,” deliberately not “professors,” for the most part simply ask questions in class
Remarkably, all tutors are required to teach all subjects, regardless of their educational background or training.
As a result, as a mathematically capable student, I found myself helping a French Ph.D. who had never taught physics figure out the math in most classes.
By conventional academic standards and expectations, this would be the height of incompetence and unprofessionalism. But I loved it
I found the classes at St. John’s to be de facto learning clubs for autodidacts in the making
I left St. John’s feeling terrifically empowered as a learner.
my focus has shifted to, “How do we facilitate the forming of learning communities with the necessary norms of learning together effectively?”
Beyond Global Free Online Learning to In Person Learning Communities
once people realize that the next stage in operationalizing free learning consists of social norms on behalf of communities of autodidacts, then more people will be able to deliberately develop, share, and design communities of such people in person.
My original intention in launching my Socratic Michael Strong YouTube channel was to share the norms of intellectual dialogue that I’ve found effective at developing children’s minds for free.
She turns thirteen next month and is not only near completing Harvard CS50, but she also scored 680 on the SAT verbal, putting her in the top 1%.
Once people realize that the next phase in making the world a better place is simply designing better subcultures, then we can all get working on this in earnest
Once more people at all levels of the socioeconomic spectrum understand this, and realize that all the free online content they ever need is already out there, they can begin focusing more on developing the social norms that matter to provide their children with a better life.
Edited: | Tweet this! | Search Twitter for discussion