The memory layer is a retrieval contract, not a storage featureMost teams evaluating agent memory compare storage options: vector database, graph, hosted service. The choice matters less than what you learn after wiring anything up, which is that storing is easy 11SOct 26 · 5 min readjoecien.hashnode.dev
I thought fine-tuning was the hard partI write this as someone who does not write code. A friend runs a pet supplies shop. Their support agents call customers 铲屎官 and themselves 小毛, and they never say "dear." Customers get answers in underOct 23 · 6 min readjoecien.hashnode.dev
I expected the translation model to winI came into this test with a lazy assumption: for translation, use the model that was built for translation. Twenty Japanese and Korean slang terms later, that assumption was gone. qwen-mt-uni got twoOct 23 · 6 min readjoecien.hashnode.dev
Deep research is usually evaluated by reading the report and nodding. The more useful question is structural: how does the model divide the work, where does the output become unusable, and what does a report actually cost in tokens and minutes? I ran Bailian's qwen-deepOct 23 · 5 min readjoecien.hashnode.dev
Real-time voice is usually sold as a demoReal-time voice is usually sold as a demo: a person talks, a model talks back, everyone claps. The engineering question is less glamorous. Where does the latency come from, what does an interrupt actuOct 19 · 5 min readjoecien.hashnode.dev
