Loading
Last discussed
10mo ago
Dec 21, 2025
1 tagged in 12 months
Why securing AI is harder than anyone expected and guardrails are failing | HackAPrompt CEO
Lenny's Podcast·Dec 21, 2025
All time
1
episodes
On record
1
took a position
Where they standWhere they stand
Talking pointsNot enough disagreement in the transcripts to form camps. These are the positions taken by people with demonstrated expertise on this topic first, then by how many people heard them on it, then by VoiceRank. Positions are extracted from transcripts by a model and may misattribute who said what. Every quote links to the episode it came from.
SS
Prompt injection attacks on AI systems are a persistent, unresolved issue that cannot be adequately mitigated by currently available software guardrails.
“I found some major problems with the AI security industry. AI guardrails do not work. I'm gonna say that one more time. Guardrails do not work. If someone is determined enough to trick GP5, they're going to deal with that guardrail. No problem.”
Lenny's Podcast · Dec 2025 · 1 episode · 20K views on this topicExtracted by a model; may misattribute who said what.
In the transcripts