Mean Time to Adapt

Mean Time to Adapt

Title history
  • Mean Time to Adapt — observed on 20261026-1010 UTC (YouTube)
🎙 Pat Opet 👥 102K 📅 April 16, 2026 ⏱ 42 min 👁 1K 📄 expert opinion 🧭 2026-10-26
Available in: English (current) Français

Keywords

mean time to adaptAI threatssupply chainthird-party riskauthentication standards

Summary

In this fireside chat at RSAC, Pat Opet, Global CISO of JPMorgan Chase, discusses the concept of ‘mean time to adapt’ as a crucial new metric for cybersecurity. He argues that attackers are leveraging AI to exploit vulnerabilities at unprecedented speed, while defenders are often constrained by legacy architecture and slow change cycles. He highlights perimeter devices as a major weak point, with adaptation times potentially spanning a decade. Opet also recounts the impact of his open letter on security-first collaboration, which called out the lack of security in SaaS providers and the risks of OAuth-based third-party integrations. He describes JPMorgan’s response: enhanced third-party oversight, threat model requests, tokenization, bring-your-own-cloud models, and confidential computing pilots. He advocates for stronger authorization models with proof of provenance and context in authentication claims. The conversation concludes with advice for smaller organizations to ask suppliers for threat models and to start collaborative security dialogues.

150 words

Critical Evaluation

Value of the Information & Strength of the Argument

The talk provides valuable strategic insights from a leading industry executive. Opet introduces a novel metric (‘mean time to adapt’) that captures the dynamic between attacker speed and defender adaptation, and supports it with concrete examples such as the 800% increase in perimeter device exploits. His argumentation is logical and well-structured, moving from problem identification (perimeter devices) to solutions (bring-your-own-cloud, confidential computing). However, the claims are predominantly based on personal experience and internal data from JPMorgan, with limited external validation. The discussion of the open letter and its industry impact adds credibility, but the lack of quantitative evidence or rigorous methodology reduces the strength of the argumentation.

Scientific Rigor, Source Quality, Title Accuracy

The speaker references specific examples like the open letter published on JPMorgan’s blog, mentions CISA’s Jenny Sterli (likely a misheard name for Jen Easterly), and discusses standards like OAuth and MCP. However, no direct URLs or external sources are provided in the video description. The title ‘Mean Time to Adapt’ is well-aligned with the core message and is not misleading. The content is logically organized, but the reliance on anecdotal evidence and undocumented statistics (e.g., ‘800% year over year’ or ’top four exploited vulnerabilities’) limits scientific rigor. The conversation is coherent and stays on topic, fulfilling the expectations set by the title.

224 words

Title / Content Match

The title accurately reflects the central concept introduced by the speaker: a new metric measuring the speed of adaptation in cybersecurity vs. attack speed.

Quality & Reliability

7/10

The speaker is a top CISO with credible experience, but the content is largely subjective opinion and strategic advice rather than empirical research. Claims are supported with specific examples and statistics, but not all are substantiated with public data.

Key Moments

Cited Sources

  • Open letter on security-first collaboration — Pat Opet's blog post that sparked widespread discussion on SaaS security and supply chain risks.
  • CISA (Jenny Sterli mention) — Referenced as a source calling for improvements in software supply chain security.

Contribution & Novelties

The video introduces a compelling new metric (‘mean time to adapt’) that could reshape how organizations measure cybersecurity effectiveness. It also provides practical architectural patterns like bring-your-own-cloud and confidential computing as strategies to mitigate third-party risk. The discussion emphasizes the need for a fundamental change in authorization models, moving from implicit trust to explicit, context-aware authentication. This offers a fresh perspective on addressing AI-driven threats in enterprise environments.

Pour aller plus loin :

  • Supply Chain Attack — Relevance: Directly related to the risks discussed regarding SaaS and third-party breaches.
  • Confidential Computing — Relevance: A key technology highlighted for protecting data in third-party environments.
  • OAuth — Relevance: Central to the discussion on authentication and authorization models.
  • Model Context Protocol (MCP) — Relevance: Mentioned as an example of a rapidly adopted standard that could inform authorization improvements.

135 words

Radar Profile

The radar profile shows relatively balanced scores, with a slight dip in technical depth and overall reliability compared to information quantity and quality. This reflects the strategic and experience-based nature of the talk, which delivers valuable insights but lacks formal empirical backing.

Reliability 7/10