A site that runs AI sentiment analysis on real Hacker News developer comments, rather than benchmark scores — Claude comes out on top for engineering sense, GPT close behind, with DeepSeek a notable dark horse.
Too many English finance and tech RSS feeds to skim by hand — TrendRadar, an open-source aggregator, filters, translates, and summarizes them via AI, then pushes the results to Telegram.
One podcast pipeline’s cover art broke when its external image URL died; the fix was routing RSS cover URLs through a single local-file method, plus a couple of quality-of-life tweaks for a language-learning audio feed.
Investing talk kept leaking into work notes and vice versa — the fix was splitting Hermes into a router plus dedicated stock, work, diary, and life sub-assistants, each with its own knowledge base partition and a strict no-cross-referencing prompt.
Podsync started throwing YouTube bot-verification and rate-limit errors — the fix was exporting browser cookies into the container’s mounted data directory and wiring them into yt-dlp’s args.
Four quick tests of the new ChatGPT Images 2.0: visualizing my current life as the AI understands it, an imagined retirement, a comic-strip take on a 2013 journal entry about orchids, and a PowerPoint-generation test.
A personal assistant on the home NAS, a home assistant tied into an investment system, and an intelligence-gathering agent on a VPS — three modules starting to work together as a real perception-cognition-action loop, not just a pile of tools.
A full writeup of the LiteLLM gateway config behind my AI stack — free NVIDIA NIM and OpenRouter models as the safety net, cheap paid models like DeepSeek V3 for daily use, and Claude Sonnet as an on-demand top-tier fallback.
After getting Hermes running on a VPS, I moved it to my Synology NAS instead so it could gradually gain access to household journal and investment data as an internal AI hub — and redesigned the model routing along the way.
OpenClaw’s buzz is fading and everyone’s talking about Hermes — its token compression alone (50%+, reportedly up to 70% in some cases) was reason enough to spin up a VPS and try it.
Blind-testing two Ollama models against real CIPLE A2 Portuguese-exam questions, the well-documented, benchmark-topping model lost badly to an obscure community fine-tune — because the benchmarks were measuring Brazilian Portuguese, not European.
Writing the plan for an AI-generated audit-rules project, we chose to reverse-engineer rules from years of proven production scripts rather than rebuilding a rule system from raw data and business logic.
