Ask HN: What weird or hard problems are you trying to solve?

I am trying to reinvent the web.

Replacing HTML/JS/CSS with a language called ALFI. It is stupidly simple in its design but still very powerful. Similarly to HTML you use it to create widgets, place them, and define their behavior. It is humanly readable like HTML but line-based instead of markup-based. Instead of nesting it uses references. This allows it to be streamed.

A big difference is that the language itself doesn't allow styling (like CSS), the downside being you get less flexibility but the upside being it will render correctly on any display with any resolution.

For this I have also written a new type of web browser called NAVI which takes ALFI code and produces (somewhat) beautiful widgets and renders them using OpenGL.

Source for both ALFI and NAVI: https://github.com/jezze/alfi

My own ALFI website: http://www.blunder.se/

You need NAVI to actually browse blunder.se properly. otherwise you will just see ALFI code. Also this is still very early so all features are not done yet.

I want to give everyone a digital identity. In some countries (including mine) basically everyone has an e-ID which we use to sign in to things like government services, banks, payment providers and much more. This is absolutely essential to everyday life and many startups are built around it.

Unfortunately, many countries don't have useful e-IDs and the ones that do are limited to that one country. I want to create a single digital identity which works for everyone, for all applications, across borders. The basic features are:

- App based with no special hardware necessary.

- Privacy friendly with the user always fully aware of what data they are revealing.

- Simple to integrate for developers. It's a standard SSO flow over OAuth/OIDC.

I'm currently calling it Pass: https://getpass.app. If anyone wants to have a chat about digital identities you can reach me at fabian (at) flapplabs.se

Safe is interesting but it’s still pretty much still a ledger. I am not sure this will work at current-internet-scale, let alone in a couple of decades worth of data. In addition to that they still pretty much keep building on top of the web, which is really not accessible to machines.

Think more in terms or ipfs/ipld.

Sure! Just ignore my dwm desktop row on the top.

http://www.blunder.se/screenshots/s1.png http://www.blunder.se/screenshots/s2.png http://www.blunder.se/screenshots/s3.png

Education.

Schools are functionally no different from part-time prisons. You must attend daily under penalty of law.

Many teachers are plain awful.

They rely on students being additionally taught by their parents. That forces parents to go to school with their children, which perpetuates the vicious cycle. The teacher recalibrates the class to the students who either understood everything the first time or had supplemental education and the rest languish.

Schools assign homework that is not easy to do when the student hasn’t fully grasped the concept. That burns time they could use to get better.

So... I am implementing Math common core in software. The first part is an automatic homework solver for math. Once we solved the student’s homework, we can teach them how to do it with generate problems. Crucially, there will be multiple perspectives and an ontology of topics do the student can backtrack to where they got lost in class weeks ago.

After we are good with math, we’ll do the same with English. It will probably not go too deep, but it will let students obtain the missing foundation of knowledge.