The goal is quite simple, allow developers to host their application with easy straight forward pricing. We are about to launch very soon. Everything is built on Laravel/PHP.

We are open to beta testers, so if you feel you want to test this please drop me and email in my profile.

An AI-native DocuSign

It's been around a month I've been working on it. Struggling with getting people to actually use it - this week I've set the ambitious goal of 10 new contracts sent *and completed* by people I don't know (last week's was 10...by people I do know).

It's hard because I feel I'm in a weird hole - in order to have a good product I need people to use it and give me feedback, but in order for people to use it and give me feedback I need a good product. It's like wth!

Another thing I'm struggling with - enjoying the process. I get daydreams like mad. I feel I'm always living in the future in some way, especially with this software, and it's taking away from being present in this work. Which sucks, because I want to be excited to *work* on this and NOT fake my own excitement towards this as a manifestation of my greed to get rich off it.

But MAN am I greedy. It's ugly sometimes, to myself.

But god how I love to work on software also. How I love making stupid bash commands on my terminal. How I love to feel like the old gods, who conquered the infant digital world.

I'm trying to see if I can "get away with it": no schema migration, no fixed views, one tenant per DB, local-first-friendliness.

The general approach is "Datomic meets XTDB meets redplanetlabs/Rama meets Local First". Conceptually, the lynchpin "WORLD FACTs" table looks like this:

  | tx_id  | valid_id | tx_t    | valid_t | origin_t | entity | attribute | value | assert | namespace     | user | role |
  |--------+----------+---------+---------+----------+--------+-----------+-------+--------+---------------+------+------|
  | uuidv7 | uuidv7   | unix ms | unix ms | uuid7    | adi    | problems  | sql   |      1 | org.evalapply | adi  | boss |

A workshop I've been experimenting with, using willing friends as guinea pigs. To help people remove friction from being able to "spool brain to disk". The sales-y part is here, with more context / explanation about what it is about and what it is not about: https://www.evalapply.org/index.html#writing-for-nerds

You upload interviews with family members (text, audio or video all work) and the system automatically transcribes the text, finds key people or events, and puts it together with other information you may have gathered about those events or people before. Like building a genealogical tree but with the actual details about people's lives.

In the works to also attach pictures of said people and events to give it some life.

I live in Switzerland and am (like many people here) an avid hiker. There are a lot of great hiking websites but they all suffer from the same problem: They are ultimately just a list of hiking routes that you need to plan around. Because I do a hike almost every week, the extra planning has become an overhead that takes time out of my life: how far away is it, what train should I take, whats the weather situation like, do I need to bring snowshoes, etc. The 65,000km of trails in this country also gives me decision paralysis!

So I'm building an app (React native/django) which takes a users current situation and preferences and then algorithmically suggests a few best options for them that they can quickly give a yes/no to. It's integrated with a lot of data like the train timetables, snow data, weather forecast etc.

I was able to reduce an hour of planning down to 5 minutes last week, so it's definitely working for me. What I am currently trying to do is figure out if other people have this problem and there's interest in the app concept.

That's why I am building Overcentric - a simple and affordable toolkit that combines web & product analytics, session replays, error reporting, chat support and help center - all in one place.

Been building it and testing with several startups and improving based on their feedback. I am also using Overcentric for Overcentric itself, so I always get ideas for improvement.

What's next: more tools that are useful for startups are on the roadmap and I am exploring how LLMs can be further utilised (apart from support, session replay summaries, aiding in writing help center articles) and refining pricing.

Check it out at https://overcentric.com/

Would love to connect with other SaaS founders and have Overcentric help them grow their startups.

My next item is to add AbuseIPDB IP addresses to my "Uninvited Activity"[1] IP address blocking system, implementing xRuffKez's script here: https://github.com/xRuffKez/AbuseIPDB-to-Blackhole

Unfortunately, but also understandably, AbuseIPDB limit their free-access (account required) API to 10,000 IP address records. So I might be putting it into a database to hopefully aggregate multiples of the 10k results if they're not always the same 10k.

[0]: https://github.com/UninvitedActivity/PiHoleLists

[1]: https://github.com/UninvitedActivity/UninvitedActivity

FOSS toolkit for SRS and adaptive tutoring systems. Inching closer to proper demos and inviting usage.

In essence, I'm looking to decouple ed-tech content authoring (eg, a flash card, an exercise, a text) from content navigation (eg, personalizing paths and priorities given individual goals and demonstrated competencies), allowing for something like a multi-sided marketplace or general A/B engine over content that can greatly diminish the need to "build your own deck" for SRS to be effective.

Project became my main focus recently after ~8 years of tiny dabbling, and I've largely succeeded at pulling spaghetti monolith into a sensible assembly of packages and abstractions. EG, the web UI can now pull from either a 'live' couchdb datalayer or from statically served JSON (with converters between), and I'm 75% through an MVP tui interface to the same system as well.

https://github.com/dahlend/kete

Research grade orbit calculations for asteroids and comets (rust/python).

I began working on this when I worked at caltech on the Near Earth Object Surveyor telescope project. It was originally designed to predict the location of asteroids in images. I have moved to germany for a PhD. I am actively extending this code for my phd research (comet dust dynamics).

Its made to compute the entire asteroid catalog at once on a laptop. There is always a tradeoff between accuracy and speed, this is tuned to be <10km over a decade for basically the entire catalog, but giving up that small amount of accuracy gained a lot of speed.

Example, here is the close approach of Apophis in 2029:

https://dahlend.github.io/kete/auto_examples/plot_close_appr...