'Keep AI on the leash' because it's far from perfect, says OpenAI's cofounder Andrej Karpathy

OpenAI's cofounder Andrej Karpathy said AI will still make mistakes "no human will make."

Michael Macor/The San Francisco Chronicle via Getty Images

• Andrej Karpathy said people need to "keep AI on the leash."
• The OpenAI cofounder highlighted AI's tendency to hallucinate and make mistakes "no human will make."
• Karparthy urged developers to slow down, focus on concrete prompts, and be the "bottleneck" for AI.

Andrej Karpathy thinks we're getting way too excited about AI...