Hacker News: datakazkn

New comment by datakazkn in "AI is not a coworker, it's an exoskeleton"

datakazkn — Thu, 19 Feb 2026 22:35:45 +0000

The exoskeleton framing resonates, especially for repetitive data work. Parts where AI consistently delivers: pattern recognition, format normalization, first-draft generation. Parts where human judgment is still irreplaceable: knowing when the data is wrong, deciding what 'correct' even means in context, and knowing when to stop iterating.

The exoskeleton doesn't replace instinct. It just removes friction from execution so more cycles go toward the judgment calls that actually matter.

New comment by datakazkn in "Show HN: Micasa – track your house from the terminal"

datakazkn — Thu, 19 Feb 2026 22:35:02 +0000

The hallucination-in-analysis problem is real and often undersold. Pattern that works well: use the LLM only to structure already-extracted data (parse fields, normalize formats), then apply deterministic logic for anything numerical. That way the LLM is doing classification/extraction where it's reliable, and you're not trusting it to compute or compare values where it isn't.

New comment by datakazkn in "Gemini 3.1 Pro"

datakazkn — Thu, 19 Feb 2026 22:34:16 +0000

One underappreciated reason for the agentic gap: Gemini tends to over-explain its reasoning mid-tool-call in a way that breaks structured output expectations. Claude and GPT-4o have both gotten better at treating tool calls as first-class operations. Gemini still feels like it's narrating its way through them rather than just executing.

Show HN: Vinted MCP Server – Compare prices across 6 EU countries via AI

datakazkn — Sat, 14 Feb 2026 13:18:17 +0000

I kept manually checking Vinted prices across EU countries and noticed the same Nike AF1 could be 45 EUR in France but 73 EUR in Germany. So I built an MCP server that lets Claude (or any MCP client) search Vinted across FR, DE, ES, IT, NL, and BE simultaneously.

The killer feature is cross-border price comparison. Ask "compare PS5 prices across EU" and get a table showing where to buy cheap and sell high. Real spreads from my testing: PS5 46%, iPhone 162%, Louis Vuitton bags 61%.

Tech stack: TypeScript, got-scraping for TLS fingerprinting, residential proxies for Cloudflare bypass. Runs as stdio MCP server locally or as HTTP endpoint on Apify.

5 tools: search_items, compare_prices, get_trending, get_seller, get_item.

npm: https://www.npmjs.com/package/vinted-mcp-server GitHub: https://github.com/KinderCN/vinted-mcp-server Hosted version (no install): https://apify.com/kazkn/vinted-mcp-server

Free and open source. Would love feedback on what features to add next.

Comments URL: https://news.ycombinator.com/item?id=47014318

Points: 1

# Comments: 0