305 karma · joined July 9, 2014
At the moment, it's not really a monetize-able product because it's so niche, and competitors are well established. My goal is to have stable, consistent pricing for years (hopefully decades), while also making it easy to switch to self-hosted alternatives.
Stats: $0 MRR
I've been using another memory plugin [1] for the past couple months. Since this change, I almost never run session compaction. Instead, I opt to just make a new session, give the agent a task, and have it figure out what happened in previous session(s).
Decision summaries look really cool -- though the agent/model info probably won't be useful for me because I don't actually ever let agents commit code autonomously (even in branches).
- [1] https://ctx.rs/
It’s a good editor, but the AI features are forgettable, and probably a distraction away from the core value proposition. They should sell commercial licenses like IntelliJ and be done with this AI monetization dilemma.
Most senior devs can't make a complete app from scratch - and that's not a bad thing. Like most disciplines we specialize, and so a "full stack senior dev" ends up being very good at a small slice of the whole software production pipeline. Even with LLMs, it takes a lot of effort and iteration before even a small product can reasonably serve a broad group of people.
Don't worry too much about "architecture" - there's never a perfect architecture. I think what you're describing is judgement. Senior devs aren't born with good judgement, nor do they acquire it by solely consuming content (books, videos, posts). The most successful pattern I've seen is: try your best, consider how difficult it would be to change things later, and favour simple solutions in general.
Since you're using this project to learn, I would advise you to use Claude Code with intention. You should read code 10x more than you write code (it's okay if Claude does the writing). Try to avoid asking Claude to explain the code to you. It's nice if you need a quick verification, but generally just try to find things out for yourself.
You're on the right track. Good luck!
This has bothered me for years. I think the solution lies in personal, private archives, and lending access to archived content within small communities.
Linkwarden does this well. You can share a collection for a small group of people.
It looks like this:
├── files
│ └── 632daffb-2f4f-4795-bb4d-3149d24f4264
│ ├── original.html
│ ├── readerview.html
│ └── screenshot.png
├── manifest.json
└── metadata.csv
This trend of outright banning the Internet Archive has me extremely worried. I fear a future where news articles are memoryholed, and no one can remember exactly what was reported and how sensational it all seemed.
I've been working on this project [0] for a while. Originally, I started with a tool that would allow people to snapshot webpages in their own browser, and they could selectively share their snapshots. Then by consensus, everyone could understand what exactly had changed, and they could draw their own conclusion about why.
While working on it, I realized that an authoritative answer to "what did it look like on $DATE" can't be produced by a no-name company. It's gotta be a non-commercial entity that's got a track record of integrity. The dream would be to allow MemoryHole customers to submit their snapshots to the Internet Archive (or other non-commercial entity). It's definitely a copyright nightmare - so no clue how this could work.
[0] - https://memoryhole.app
You may run into some issues with Docker and native deps once you get to production. Don’t forget to cache the bumblebee files.
https://hexdocs.pm/elixir/List.html#myers_difference/2
https://github.com/elixir-lang/elixir/blob/master/lib/elixir...