@superteam Looking for co-founders and teammates to join to win the next hackathon!
I'm a software engineer with 6+ years experience, I won multiple hackathon before.
LLMs one-shooting apps and features actually is a really bad thing.
One-shooting app means there is no:
- No architectural decisions. Just code that exists.
- No security model. Just features that work (until they don't).
- No scalability plan. Just something that runs today.
- No maintenance strategy. Just technical debt at scale.
It's basically trading speed for technical debt at 100x leverage.
The developers winning with AI aren't using it to skip steps. They're using it to do the hard steps faster.
Yeah, Posthog is very powerful. Especially if you have a lot of users and interactions happening.
Its power comes from customization, so you have to create custom events then put them in insights and custom dashboards, based on what you want to track.
I've been using it for 2 years, but still learning.
One of the cool @tan_stack tools is "@tanstack/react-devtools".
It lets you create or install plugins for any dev feature you need.
In my case, I needed a way to generate mock data for heartbeats. Tanstack itself didn't do the heavy lifting, but it made everything much cleaner and more organized.
@charlietlamb@nextjs@EffectTS_@better_auth@autumnpricing No, Nextjs has been consistently slow since they introduced the app router across thousands projects and machines and users
It was the most common complaint on github, they improved a lot but it's still bad
the stack behind boringstatus (open source uptime monitor) ๐
- bun (to save seconds from dev)
- tanstack/starter (the 'not-slow' nextjs alternative)
- better-auth (best typescript auth library i found)
- postgresdb (nothing comes even close)
- timescaledb (handles millions of heartbeats)
- docker (one command dev/deploy)
- golang service for metrics (fast and simple)
keeping it simple pays off. code is WIP here:
https://t.co/mdvSfQ0tQd