Shared Security Podcast 2025 Predictions: Hits, Misses & What We Learned
Dec 29, 2025
The hosts reflect on their 2025 predictions, analyzing what came true and what didn't. They dive into AI-themed phishing trends and the effectiveness of AGI guardrails. TikTok remains unbanned while federal privacy laws stagnate. Predictions for 2026 promise bold forecasts, from rising pushback against AI to local privacy legislation growth. The team also celebrates viewer milestones and the impact of engaging content, like shorts, on their audience. Tune in for insights on cybersecurity, AI, and geopolitical changes ahead!
AI Snips
Chapters
Transcript
Episode notes
AI Guardrails Feel Fragile
- Guardrails around powerful AI remained uncertain and felt fragile by end of 2025.
- High-profile warnings (e.g., Dario Amodei) left hosts skeptical about timely effective controls.
Fix Fundamentals Before Chasing Headlines
- Do prioritize basic hygiene over chasing exotic vulnerabilities with headlines.
- Patch legacy systems and remove Windows 7 devices from domains before chasing fancy attack research.
Agents And Edge AI Expand Attack Surface
- AI agents rose as a distinct category: autonomous tools that perform tasks rather than chat interfaces.
- These agents plus edge AI create new attack surfaces and social-engineering vectors.