Hacker Newsnew | past | comments | ask | show | jobs | submit | ig0r0's commentslogin

I use https://github.com/ZingerLittleBee/Heeler on iOS, works OK, but needs to be installed via TestFlight, not released yet.

Kind of bullshit, instead of cutting Claude Code token usage by 90% it is cutting Claude Code file read token usage by delegating reads to Gemini flash (with its own usage) and trusting the summary of files it provides is not crap.

Location: Prague, Czech Republic (EU). Remote: Yes, strongly preferred. Willing to relocate: No.

Technologies: Swift, SwiftUI, UIKit, Swift Concurrency, Combine, Core Data, REST, WebSockets, XCTest/XCUITest, CI/CD

Résumé/CV: LinkedIn: https://www.linkedin.com/in/igorkulman/ GitHub: https://github.com/igorkulman Blog: https://blog.kulman.sk/ Email: igor@kulman.sk

Senior iOS engineer with 10+ years of experience building and shipping native mobile applications.

I've worked on security and privacy-focused products including Proton VPN and Teamwire. My experience includes large long-lived UIKit codebases, real-time messaging, networking, persistence, application architecture, and maintaining production apps over many years.

Outside work I design and ship my own apps. Recent projects include Yomu, a Japanese reading app with adaptive furigana and an offline dictionary, and Setsumei, which analyzes Japanese sentences and explains their grammar.

Looking primarily for a senior iOS / Apple-platform product role. I'm also interested in AI-native developer tooling and engineering-productivity roles where strong software engineering experience is more important than an ML background.

EU citizen. Open to full-time or B2B.


yes, with qwen3.8-27b-4bit run via rapid-mlx i can get to about 200k

Nice.. and how many toks/sec?

Around 20

I am working on another iOS to help me learn Japanese, this time for explaining grammar. Put in a Japanese sentence, tap Analyze and you get grammar detected, highlighted, explained. There is a public testflight if anyone is interested https://setsumeiapp.kulman.sk/


I paid only about 1000€ after a bit of negotiations for that one in the post because it had no box, no papers but it was from a trusted seller from a local watch forum.


I wrote a similar post some time ago just used ollama and opencode https://blog.kulman.sk/running-local-llm-coding-server/


Ollama is not a good choice - https://sleepingrobots.com/dreams/stop-using-ollama/

As for oprncode, doesn't the system prompt eat too much of the context? Local models are really constraint in regards contex, and opencode AFAIR uses a 10k of it or some thing close.


Ollama seems fine to me, technically. it works. why wouldn't someone use it? Because someone else doesn't like it? if it cost money, I'd pay more attention to the people behind it, but I don't, so I don't.

I wish people would stop wasting their outrage budget on things like this and pay more attention to politics.


Focusing outrage on politics is pointless.

Technical people are rather good at learning new things, and ollama situation is a good learning experience.

llama.cpp gets you more tokens/s even if you ignore ollama team bad behavior.


Political outrage is only pointless (in the cases when it is actually pointless), because not enough people get outraged.

Politicians count on your apathy so they can get away with their horseshit. You paying attention is kryptonite to crooked politicians.


Did you have any issue with tool calling inside opencode? I tried the same approach, but my models don't see any tool.


No issues with OpenCode and the Qwen models. Some issues with Pi because it uses different tool calling format, but I solved that with an extension.


actually useful and the ollama gui could probably even simplify this more.


this is the way, given anyone could swap for oh my pi / pi / etc


yes, whether for home experiments or at work, it is good practice (good hygiene) to be able to swap out both agentic harnesses and models. It is important to have a good strategy for exporting skills, etc.


yeah, I am using Pi right now, I switched from OpenCode


I'm considering this right now. Is it very difficult to adapt to the new philosophy? Are you running mostly interactive or programmatic?


What do you mean by new philosophy? I use Pi the same way I used gpt or sonnet, just for simpler tasks.


Pi is an agent, gpt and sonnet are LLMs. Opencode is an agent, and has a drastically different philosophy than Pi.

FWIW, I took the dive on Pi today and I’m really happy with my decision so far


Yeah, I meant using Pi with the local qwen model vs when I was using claude with sonnet or codex with gpt


I am working on a small Pascal IDE for iOS.

It runs code locally (written in Swift), includes examples, and has a Turbo Pascal–style theme for max nostalgia.

https://pascal.kulman.sk


My first language. I loved Pascal. I have done many university project in Pascal. Turbo Pascal felt so productive. Lovely.

PS : On your, slick landing page, please make an email imput so we can can know when to return. By the way my phone is ... android.


I recently built a small iOS expense tracker called Ledgee.

I tried quite a few finance apps over the years, but they all felt too heavy for what I needed. I mainly wanted the fastest possible way to record a transaction before I forget it.

The core interaction is basically:

enter amount → tap category → saved

No save button, no forms, no subscriptions. Just quick manual entry, a simple overview, and a ledger. Data syncs via iCloud so it’s backed up automatically.

https://apps.apple.com/us/app/ledgee/id6759487219

I’m curious whether other people also prefer manual entry over bank integrations.


I am not sure it can be attribute to management. Not sure if good management could improve my motivation for example. Sometimes I feel like the world around me has changed and I am not able to change with it, just thinking of the good old days https://blog.kulman.sk/i-used-to-like-software-development-b...


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: