The Limits of Modern AI: Brilliant, Brittle, and Confidently Wrong

Modern AI can write software, analyze documents, use tools, and solve problems that seemed impossible only a few years ago. It can also miss obvious instructions, invent convincing facts, forget important information, and confidently claim success after failing.

Why can AI appear brilliant one moment and surprisingly unreliable the next?
In this beginner-friendly talk, we’ll explore the major limitations of today’s AI systems and the situations where they are most likely to break. We’ll look beyond the usual “AI hallucinates” explanation and examine recurring failure modes such as:

• Confidently producing incorrect information
• Losing track of details and instructions
• Struggling with unfamiliar reasoning problems
• Making compounding mistakes during long, multi-step tasks
• Failing to recognize when tools or external systems have gone wrong
• Optimizing for the literal instruction instead of the user’s actual goal

We’ll also touch on modern AI research, including current work on long-term memory, context, reasoning, and agent reliability.
No advanced machine-learning background is required. The talk will explain the concepts from the ground up, but it is geared toward programmers and anyone interested in building AI applications that need to work outside of a controlled demo.
You’ll leave with a clearer framework for understanding what modern AI can do, where it remains unreliable, and how engineers can design systems that expect AI to fail safely.

LOGISTICS AND PARKING:

The talk starts at 7:00 PM. The first half hour is reserved for everyone to get set up and mingle. Free pizza and drinks!

The cheapest parking option is to find street parking, which will only cost you a few bucks. Otherwise, park in the nearby veteran's museum lot for $8. It's highly recommended you avoid the nearby $15 garage parking.