Discover more from michael's writing
exploring a new model in higher ed (pt.1)
e/acc - accelerated learning principles for the 21st century
“what are your personal views of consciousness, specifically as it relates to information packing at a quanta level? i started an internship at zapata computing and have been thinking about the hardware-software divide in quantum computing wondering if this may hold the key.” asked a curious 17 year old.
what impressed me was not only the thoughtfulness in the question trying to form new hypotheses on personal research, but also the earnestness in the curiosity. she was genuinely curious to seek understanding.
having recently finished my masters in boston, this is the kind of question i wish more people around me were asking rather than ‘what’s a quick route to a 30% pay increase?’
no, this is not a prodigal story of individual privilege, but instead my day 1 greeting in toronto starting the role as director for the knowledge society (tks). quick intro for those who don’t know: tks a 10 month global accelerator helping teens learn and build in emerging tech and science. this was the collective norm.
for context, ~75% of students don’t have a technical background in coding before the program, and all students are accepted on a need blind basis with scholarships provided for those who need it.
these are knowledge and skills learned through an accelerated learning environment with no pre-requisites. how did this kind of exceptional curiosity become the norm for hundreds (now thousands) of teens?
3hr hackathon style sessions weekly covering ambitious challenges like socratic discussions around ‘how should governments regulate ai (if at all)’ or research sprints for ‘what are novel applications of gene editing?’
students build their own portfolio of projects throughout the 10 months in an different areas of emerging tech + science
1-1 personal coaching with former founders, researchers, and startup employees
they get more done in ten months than most students don’t get done in 5x that amount.
the future is already here – it's just not evenly distributed. william gibson
while tks is set up similar to a club sports team commitment with ~10-15hrs a week for high school students with high powered coaches. i’m curious what principles from current startup accelerator models can be carried forward to change higher education for 18-25 year olds.
to keep up with the accelerating pace of technology, the future of education needs learning environments that can adapt to the speed of change.
colleges are losing public trust with high costs and slow to adapt
post covid, 2/3 of gen z think college overall is not that important. 59% of americans agree with them, where they do not think a college degree is important—more in this recent piece in the new york times. the average student loan debt for american college students is now $36,520 and still growing.
times are changing with major companies are now hiring people without degrees including apple, tesla, space x, e+y, google and more. hiring for skillsets is so important, google has started it’s own certificates across information technology, user experience design, product management and more to fill their own hiring needs.
some how the acceleration of quality of knowledge on the internet has gone up, but the promises of quality experiences have not kept up. for example, coursera has a 97% drop out rate despite having some of the best instructors in the world designing new classes. cohort based courses like on deck catalyst program had to shut down because it’s hard to scale quality of experience with clear outcomes.
the internet age of education has under delivered as a collective despite having some of the highest potential for those who are self motivated to learn different topics.
we need new principles for the foundation of environments, both physical and digital, that can culitvate curiosity and build understanding.
the limitations of batched learning
traditional college curriculum operates in a batched system. students are treated uniformly regardless of their indiviual aptitudes or interests.
this approach feels like a three legged race, where two people tie a leg together to try to move as one.
imagine one person (orange) can run at 1.5 m/s
another person (blue) can run at 3 m/s
when tied together (purple) they can only move at 1 m/s
the difference of individual speed and being tied together:
person one has a lost potential of -0.5 m/s
person two has a lost potential of -1.5 m/s
over time, that lost potential adds up as people move at different speeds when building knowledge, practicing skillsets, and exploring what they are interested in.
what’s worse is that the batching is not a one to one pairing, it’s a 25 to 1, 100 to 1, or 500 to 1 depending on class sizes. this reduces everyone’s baseline to an averaged pace for whatever makes collective sense.
self paced with directional focus
what if instead of predetermined and pre-timed curriculums, there was the ability to customize based on personal potential and curiosity in different topics?
the baseline output in chat gpt with 1 plug in (scholar ai) looks to be able to create a core outline to get started. this is without feedback or fine tuning. this is just a starting point. eventually this could be tweaked to individual curiosity or biased towards outcomes. there could be personalized feedback loops in the journey.
students can rapidly iterate through different directions rather be confined to a linear path.
enabling every student to learn at their own pace enables freedom + exploration. i’ve seen this first hand in the first few months of previous accelerators. for example, some students playing the great online game writing their first technical blog post:
one might finish in 2 days because they are excited, have the time, and figure out how to hit publish
another one might take 2-3 weeks as they want to take time exploring what to write about and like the ability to write in sections with feedback loops along the way.
or in a conventional academic environment, one might never even complete because it’s not on the syllabus so they don’t even try.
outside of self paced learning, hackathons are probably the closest proxy environment for accelerated learning. with the limit of a day or weekend, students have 24 hours to explore global problems or pitch new ideas.
once you get a taste of that freedom, it’s hard to go back to the banality of the batch.
time-to-[x] can be one indicator for measuring acceleration in personalized learning
as time is an important ingredient in the pursit of athletic peak physical performance, why not use the same for cognitive or creative performance?
time to learn… it’s as simple as a 2 hour youtube rabbit hole
time to publish online… less than an hour after exploring and researching
time to build a project… this can be 2-3 weeks with self guided tutorials and online courses
time to seek feedback… integrated into the self learning process
time to new skill… get 5 reps in a week
… all the way through time to 100 reps…
one reason i say ‘indicator’ and not ‘main metric,’ is i don’t think time is the most important thing to optimize for, but it’s one main ingredient to challenge the quality of current learning environments.
this is part one in a series where i hope to distill my perspective from running accelerator programs in the last ten years to figure out what may make sense for a new model in hgiher ed. without challenging the norm and reasoning up what makes sense, we’ll be stuck with education systems designed over a century ago.
if you have ideas for what you want to see or comments, please share below.
to keep up with the accelerating pace of technology, the future of education needs accelerated learning environments.
tks is one example for teenagers to learn + build in emerging tech (ai, quantum computing, blockchain, and more)
batching individual learning journeys is like having people run a 3 legged race, or sometimes a 300 legged race. this reduces individual potential down to the lowest shared average ability.
self paced learning enables freedom and exploration, two core ingredients to real learning
time to act could be one to indication for an accelerated learning environment