Bilingual medical education for the GCC medical community. Lessons, analysis, and perspectives from the operating room to the digital frontier.
Twenty years in the operating room taught me one thing: not every solution is useful. Same applies to building LLM products. When I decide to operate, I don't use every tool in the room — I choose the proven one.
Most patients think rest means staying still. Here's what actually happens when you lie still too long after an operation — and why walking is the closest thing to medicine you can do for yourself.
معظم المرضى يظنّون أن الراحة تعني البقاء في السرير. إليك ما يحدث عندما تبقى ساكنًا لفترة طويلة بعد الجراحة — ولماذا المشي هو أقرب شيء للدواء يمكنك أن تفعله لنفسك.
When you call an LLM API, you're not just sending a message. You're purchasing context window space — and that space is finite, fast-filling, and billed per token.
Fine-tuning is the most overrated decision in the LLM stack. Most people reach for it too early — and when it actually makes sense.
RAG doesn't hallucinate — but it retrieves lies that look truthy. If you're building a production LLM system, your retrieval pipeline is your actual product.
Most prompt engineering advice is written by people who spent months building systems for edge cases. Stop over-engineering. Start iterating.
Every time someone says 'the model reasoned through this,' a small lie is told. Here's what's actually happening inside the transformer — and why understanding it makes you a better builder.