i-dont-remember
Joined 8 years ago
Karma
28
Submissions
22
Member since
Jul 17
Recent submissions
Comment · 7 months ago
Comment · 7 months ago
This is really cool, i love seeing a fun interactive game tied in to deep knowledge about an industry.
Comment · 10 months ago
I've been helping with a small portfolio of Slack apps under the brand https://happybara.io, including Nightowl (multi-DM broadcast for Slack, like BCC for email), and Channitor (channel management & auto-archiving).
It's been a really interesting journey making mistakes & learning from them. One mistake that made me feel foolish: I assumed no one would ever pay for one of the apps. A paywall was added "just in case" -> boom, customers almost immediately started signing up. I felt incredibly silly after that, especially for the number of times I've read on HN and similar sites "your work is valuable" and "charge more".
Comment · 10 months ago
This is neat, I appreciate that even if I can't use the service, i still learned something new about what's possible. Gonna keep it in the back-pocket for sharing with friends.
Also reminded me of when Justin Jackson talked about how he hired his kids for help with real projects and the positives from it https://justinjackson.ca/flipping-tables
Comment · 1 year ago
if you want to get started writing super quick, you can also take a look at some of the super minimalist platforms that are basically as you describe - like https://bearblog.dev/ and https://mataroa.blog/.
I've stumbled on both from submissions to HN and they have become my recommendations for friends that want to write, especially the couple who are prone to jumping down technical rabbit holes & not ever getting to the writing.
Comment · 2 years ago
I see resolutions for protecting against the now-known error modes discussed, and better alerting to get the on-call engineer (aka always Zeke :D) looking into things quicker, but curious how they might approach preparing for "unknown-unknowns" that will come in the future.
Are there good ways for a small-team to proactively stress test a system without mucking up customers? Open question.
Comment · 3 years ago
Video[0] isn't a direct answer, but I found it helpful for understanding the trade offs that come when considering using electric power for a plane vs regular fuel. They show the math in an easy to follow diagram.
tl;dr for their small kit aircraft the weight of batteries they would need to match the stored energy of equivalent fuel (even with a battery at 500wh/kg) would be 5-10x heavier, and also not get lighter during the flight. They said for long range it doesn't make sense, but that there are lots of companies iterating in the short range electric space.
Comment · 3 years ago
I've seen a couple tools pop up around this idea. Haven't explored them much, but:
- MonicaHQ (open-source)
- Dex (free version seems good enough for most people, i'm trying this one out rn) getdex.com/
- Custom system in Airtable - https://news.ycombinator.com/item?id=30329475
- Infrequent app - blog post talks more about it, linked in comments - https://news.ycombinator.com/item?id=33308084
Comment · 4 years ago
These aren't my projects, but inspiration I've found around the web and started keeping track of recently. https://indielurker.com/#8dff6575218049a78b6b86e977c8d66b
Some of my favorites:
- Nomad List started as a Google Sheet.
- Product Hunt began as a simple list emailed out to people.
- Levels.fyi was powered by a Google Sheet until 2021.
- Crave Cookies was making $200k/mon according to their interview using a SQlite DB which started as a JSON file.
- Headlime started as a JSON file sitting on his frontend with no DB - then sold for $60k a week later
Tools I personally am exploring:
- DB: Google Sheets - either through API (numerous services like sheet.best for this or self-hosted)
- DB: Airtable - it provides API out of the box and has forms and whatnot too, keeps things real quick to start
- Whole Shebang: have only built toy examples, but Glideapps and Softr seem like handy options to spin an idea up really quickly and not getting distracted by tech rabbit holes
With the call to action of "raise your own model", what might that actually look like? Will it be feasible for individuals, or would there need to be some trusted open source foundation that creates the "Linux" of models?