Justin Baeder, PhD

Education philosopher & instructional leadership author. Creator of Repertoire, the professional writing app for instructional leaders.

[Justin Baeder, PhD @eduleadership]

I’m thrilled to announce that my new book with Keith Fickel is available for pre-order!

Cultivate & Activate: Building Teacher Capacity for Instructional Leadership

Ships May 2026

https://t.co/To9aFRLCIT

[Justin Baeder, PhD @eduleadership]

@Shadow_Rebbe No, because I have a lot of background knowledge on this subject and no reason to take him seriously. Every ed tech founder claims to be on the cusp of reinventing learning, and it’s all the same nonsense.

[Justin Baeder, PhD @eduleadership]

I mean, you can keep trying stuff, but there’s nothing to be found in this “unexplored territory.”

Personalization has been tried for decades and has never worked, because kids don’t learn in different ways.

[Alex Milenkovic @CuriaCEO]

@eduleadership @MichaelPauleen Our data shows significant variation in how individuals perform tasks, and respond to interventions. I consider it encouraging that so much unexplored territory remains in individualizing instruction.

[Justin Baeder, PhD @eduleadership]

@tracewoodgrains Amazing that he never thought of “teach the kids more stuff.”

Remarkably little attention to the most obvious lever.

[Justin Baeder, PhD @eduleadership]

@MichaelPauleen I’m not saying there’s no personalization in medicine. I’m using it as an example of a field in which it’s clearly ridiculous for non-experts to propose miracle solutions.

Personalization has been attempted for generations in K-12 with nothing to show for it. Not a new idea!

[Justin Baeder, PhD @eduleadership]

@kutluokan @MichaelPauleen OK, but student agency has zero overlap with “an algorithm decided to give you this task next” personalization

[Justin Baeder, PhD @eduleadership]

This works on TikTok and Netflix because people have different preferences.

But—popular slogans notwithstanding—all kids learn the same way.

You can split-test to find the most effective materials, and should, but can’t personalize.

[Ben Somers @ben_m_somers]

Alpha School made a groundbreaking change to how we teach kids academics that every school should copy

They started A/B testing every piece of learning material to give kids the most effective materials for every topic

To accomplish this, they needed to divorce mastery tracking from curriculum. They needed to build a "routing" layer that has independent diagnostics/assessment in order to be able to run these A/B tests

We're taking this one step farther at Recess; our product has a recursive/self-improving loop similar to how tiktok works.

We're building an independent assessment and tracking layer, and then using classic ML-based-recommendation systems to show kids content for each node. This will automate the A/B tests and also allow different kids to get different types of content based on what will perform the best for them, vs. having one population-level curriculum that all kids follow.

It's a bit like the difference between a cable television network and Youtube. Cable television creates a "program schedule". There's a set playlist of shows coming up and you follow that. Alpha is a GREAT network cable show; think HBO.

Our approach looks more like HBO Max or Netflix or Youtube; we still syndicate all that amazing content, but every user is recommended slightly different content that makes sense for them. Not all users will see the same content; some will get content that no one else on the platform gets

[Justin Baeder, PhD @eduleadership]

This is all widely known among education experts, but apparently we’re too stupid to bother asking.

Imagine treating the field of medicine that way.

[Justin Baeder, PhD @eduleadership]

The “butbutbut we’re different” cope in ed tech is really something to behold.

No, you all think the same thing, and are all wrong for the same reason.

[Justin Baeder, PhD @eduleadership]

@benwaldman @ben_m_somers Nope, why would kids learn in different ways?

There are persistent myths like learning styles, even among educators, but it’s not news.

[Justin Baeder, PhD @eduleadership]

@EdTechMuser @MichaelPauleen They treat it just like any book. Potentially useful, but not going to create breakthrough results compared to anything else.

[Justin Baeder, PhD @eduleadership]

Just to clarify, no one in K-12 takes this paper seriously, and few are even aware of it, because it’s so ridiculous.

[Michael Pauleen @MichaelPauleen]

Imagine the most important research in your field was a hand drawn graph based on data from an unpublished grad student's three week long experiment.

Glad we're finally testing it for real.

[Justin Baeder, PhD @eduleadership]

@effortfuleduktr The argument for Alpha is basically the same as the argument for the Internet a generation ago, for libraries centuries ago, and for the printing press half a millennium ago.

Learning is not an information problem.

[Justin Baeder, PhD @eduleadership]

@EdTechMuser @MichaelPauleen K-12 educators and education researchers absolutely do not take Sal Khan seriously, because none of his stuff makes a difference.

[Justin Baeder, PhD @eduleadership]

@MummaLabs @evernote Yuuuup

https://t.co/T8vCKew5f3

[Justin Baeder, PhD @eduleadership]

Cancelled my @evernote subscription after 18 years.

Sorry, $249 a year to hold some text files, PDFs, and images is just not worth it. Nor is $99 for 1GB. Come on.

Claude got everything into Obsidian for me, and now it's free.

[Justin Baeder, PhD @eduleadership]

@MichaelPauleen So it’s important because people who don’t know what they’re talking about like it 😅

[Justin Baeder, PhD @eduleadership]

@kutluokan Sure! There’s tons of research on tutoring, and it’s about a 0.4 effect size, not 2.0.

At a quick glance this is probably a good overview:

https://t.co/hfkeh8MN1x

[Justin Baeder, PhD @eduleadership]

@gtmom @evernote Yeah, kind of the Bending Spoons business model—buy a stable business and pull the levers to make it more profitable.

But this kind of thing is trivial to build now, and I was already using Obsidian which is crazy fast and talks beautifully to Claude.

[Justin Baeder, PhD @eduleadership]

@ben_m_somers Good on you! Not a bad use of philanthropic dollars.

[Justin Baeder, PhD @eduleadership]

@Shadow_Rebbe And of course 2.0 has never been replicated, whereas 0.4 is based on tons of studies.

So it’s kind of like saying “Yeah actually there’s a QB you’ve never heard of who’s 5x as good as Tom Brady.” Ok bro.

[Justin Baeder, PhD @eduleadership]

@Shadow_Rebbe So, a good basis for comparison is John Hattie’s effect size meta-analysis work, which certainly has faced much criticism, but gives a rough sense of how outlandish the two-sigma claim is. Hattie found tutoring to be about 0.4, in line with other interventions.

[Daisy Christodoulou @daisychristo]

Bloom's 2 sigma study is one of the most poorly designed and misleading papers in all education research.

It's based on a couple of hundred kids who got 7 hours of intro lessons on cartography. Their progress was measured with unstandardised assessments.

It cannot bear any of the enormous weight that people place on it, and it baffles me that 40 years on people are talking about recreating it like it is some kind of holy grail.

We will know education is a mature science when people stop talking about Bloom's 2 sigma study!

https://t.co/7ayeADPXh0