Posted by david927 11/9/2025
Ask HN: What Are You Working On? (Nov 2025)
I’m responsible for product research, analysis, design, development, and promotion. Most of the workflow was powered by AI tools — more than 60% of the research, design, and coding involved AI assistance. Initially we used Bootstrap for responsive design, but later switched to a lighter TailwindCSS framework.
Right now traffic is very low (just a few dozen users per day), and I’m trying to figure out whether that’s a product issue or simply a lack of promotion. Honestly, it’s a bit discouraging, but I’m hoping to learn from feedback and iterate.
Curious if anyone here has experience growing early-stage products with niche financial data — how did you approach the balance between product refinement and marketing?
But thanks to LLMs, I finally decided to give it a go and got something basic working in a short time, hurrey for AI assisted coding!
Feels empowering to be honest. No idea if I will really implement the main ideas, that I have since a long time, but I know that I can now if I want to.
==> Here is a photo of the Editor : https://ibb.co/FC9Hzj2
==> website with an example of a FPS visit https://free-visit.net
Please be honest, tell me why I don't have traction.
It's great that you're working on this. If you want to continue on this, I'd consider: - Cleaning up the design of the website-- it looks kind of crappy. Get an AI agent to clean it up for you, it's better to look like "generic professional website" rather than "crappy amateur". - Use the more common words for creating 3d models. A "Visit" sounds like an experience, but what you're really making is a "scene", or a "spatial capture", or a "floor plan". - Maybe try to figure out a niche. Is your niche that people can edit this the 3d object afterwards? Or is the niche integration with video games? You gotta find something that doesn't directly compete with polycam.
And yes.It's not 100% automatic. I agree, my low point.
Too many things I wanted to analyse went to nothing because I was too lazy to fetch information and the put it inside a spreadsheet cell by cell. So I developed this to help me extract docs into spreadsheets while also having access to the web.
Now working on a vibe-coded version of it where instead of showing a spreadsheet, it will be able to generate data-focused tiles and apps
Decided to pivot and start learning about databases and their internals more. Currently pulling down Clickhouse and reading some code along with the reading the book Database Internals by Alex Petrov.
So I'm technically not "working on" an app...I am working on myself to branch out and attempt to specialize a bit more as I progress in my career.
Any advice/papers/books to read is very welcomed!
Building native applications for iOS, Android and Huawei devices in Haskell.
Projects like lynx and react-native automate this process using something akin to node-gyp, exposing kotlin / swift libraries via C ABI w/ a JS API. Miso accesses the kotlin / swift native modules by FFI'ing into the JS that exposes them.
The JS doesn't get compiled, but on Android it does get JIT'd. So it's "native" in the sense that the views drawn are native (not WebViews), and the device APIs are native, but not "native" in the sense that it's compiled.
Working on an app that helps me (and other people) do household management on autopilot. It helps me manage things, food, expiration dates, shopping, chores, and I get notified periodically to review my lists. I waste way less food and I actually do my chores instead of procrastinating. https://okthings.app
Generally, i tend to buy the same things, so the "big" data entry job happens only once.
Yeah I'm not gonna touch it and I'm going to actively encourage people to disable it. Use signal instead.