[BinaryIdiot]

I used to work as a developer for an online charter school many years back so this space is very familiar. I even thought about starting my own company and breaking into this space because the software I worked on wasn't very good (it was miles above what was already out there, for sure, but I thought there were many places that could be improved).

The problem I ran into, which completely discouraged me at the time, was that of making money. You need a significant amount of capital to sell to a private school and a public school requires sales cycles of at least a couple of years. It just wasn't possible with me already having a family, mortgage, etc.

I would love to see the whole thing get disrupted. I even think about doing something in it from time to time. Having YC behind you would certainly be a big help. One big issue you face is that you're going to be teaching children and helping to shape their minds so you must use valid scientific research to back up your ideas less we create a generation of adults who cannot think about the world. This is the biggest hurdle, IMO, because it's so easy to just appeal to someone's bias narrative. For example "brain games" which "train" your brain basically only train it to play those games but they're still hugely popular and sometimes marketed to the same demographics.

[barry-cotter]

> One big issue you face is that you're going to be teaching children and helping to shape their minds so you must use valid scientific research to back up your ideas less we create a generation of adults who cannot think about the world.

Don't let the barely adequate be the enemy of the actively harmful. Most education research is of low quality and education does not incorporate known effective techniques like mastery learning or spaced repetition.

Education degrees and education research deserve all the respect they get.

[hdctambien]

Honestly, fixing the education of educators would be the best long term solution.

Alternatively, creating a complete, ready-to-go curriculum package for every course (as defined in every state) that could be handed to anyone might be a startup-able solution to that problem.

Although, that might also just be a patch that leads to a local maximum of not-very-good teachers teaching from better-than-before curriculum. Where making better teachers would lead to better teaching (eventually).

[germinalphrase]

I believe there is opportunity in giving teachers better tools to teach as well as they are capable.

Teaching is difficult enough from a personal performance and logistical standpoint that there is low hanging fruit in simply helping teachers be their best professional self.

For example: I have approx. 200 high school level students. If I assign them to write one page I have just given myself 200 pages of assessment to complete. Assessing writing is more cognitively demanding than simply reading 200 pages, so if I can read and provide feedback on each page in 3 minutes - working non-stop - I have 10 hours of work. I have 5 hours of prep time per week which is primarily used for preparing the 7 hours a day I spend actively in class with students. Given that we don't complete just one assignment per week - you can see the logistical difficulty.

I teach english. There are commenting/notemaking tools in almost every online word processing toolset, but none of them are streamlined for assessing writing (i.e. providing rubric based feedback, rapidly annotating with custom or pre-created notes, allowing for layered annotations to identify different structural features of writing). There's opportunity in figuring out what teachers already do and slash the time required to do it.

[hdctambien]

I completely understand. I teach AP Computer Science to about 100 students a year and I assign many Open Response questions which are solved by writing code by hand on paper. I grade these using the same rubric as the College Board which takes into account much more than "would it compile". For every hour of homework I assign my students, I assign myself double digit hours of grading.

Some AI + a pool of previously graded assignments could probably automatically pre-grade essays (or my open response questions) with comments and notations which would allow a teacher to skip the 100% correct papers, and focus on the hot-spots of the others. Automatically identifying plagiarism (via the internet and other students) would be a nice side effect, too.

Those kinds of solutions seem way more useful than Yet Another Gradebook/Attendance System or a Teacher->Parent Chat Room or even Distributed Tutoring Systems.

[alphydan]

You can already find "a complete, ready-to-go curriculum package for every course" on the internet.

Unfortunately (1) it's fragmented into lots of sites ... but most importantly, (2) it's low quality, and typically not adjusted to your particular year (From year to year the level of students varies a lot, and some need easier or more difficult content). On top of that, some years you have 50h with the students, other years 60h (depending on holidays, breaks, outings, interruptions). So that "package" would need some input parameters to achieve differentiation, acceleration, etc.

I often find it easier to spend 2h creating my content from scratch than adapt somebody else's (This can mean 20h - 30h of preparation in a week when you are teaching something for the first time, on top of teaching, supervising, after-school activities, etc)