As a part of my AI Product Design course with MIT xPro, I just lately found one thing about how we work together with Massive Language Fashions (LLMs). Throughout a immediate engineering train, I made a decision to check ChatGPT’s data in an space I do know effectively — blockchain expertise. What adopted was an eye-opening lesson within the artwork of asking questions.
I began with a broad question: “Ought to enterprises use public permissionless or personal permissioned blockchains?” It appeared like an affordable query on the time, however I rapidly discovered why this method not often yields helpful insights.
“It’s like asking ‘What automotive ought to I purchase?’ with out mentioning in the event you want it for household highway journeys or System 1 racing.”
Right here’s the place it will get attention-grabbing. As quickly as I requested particularly about enterprises exploring healthcare options. The LLM grow to be rather more insightful. As a substitute of generic professionals and cons, I obtained focused insights about dealing with delicate affected person information and assembly regulatory necessities.
Nevertheless, I observed one thing essential: the AI rapidly defaulted to recommending personal permissioned blockchains — the “protected” reply — with out totally…
