New episode of Security Now: Security Now #1093 – “Tokens in the Stream” is now available.
Featuring hosts Steve Gibson and Leo Laporte.
Topics covered:
How ‘role’ tokens and end-of-turn markers work inside LLM conversation streams
Why prompt injection may be an unsolvable flaw in current AI architecture
The controversy surrounding AI model distillation
Anthropic expanding access to its most powerful Claude (Mythos 5) model
Bitwarden’s Secrets Manager as protection against agentic and prompt injection abuse
#SecurityNow #Cybersecurity #TWiT