---
title: Understanding the Benefits and Drawbacks of ChatGPT and AI
description: A seminar on what large language models are good for and where they fail — origins, practical tips, and the limits worth knowing before you rely on one.
publishDate: 2024-04-24
canonical: https://www.mandalivia.com/talks/silicon-hills-chatgpt/
---

Large language models break from the mental models of software we're used to. For
decades we worked with deterministic programs that explicitly follow directions.
Language models are probabilistic — exciting when you want imaginative creation,
alarming when you need facts.

A seminar on both halves of that: where the models came from, what they are genuinely
good at, and the failure modes worth understanding before you depend on one.