/
© 2026 RiffOn. All rights reserved.

Get your free personalized podcast brief

We scan new podcasts and send you the top 5 insights daily.

  1. The a16z Show
  2. OpenAI's Joshua Achiam: Did We Already Reach AGI?
OpenAI's Joshua Achiam: Did We Already Reach AGI?

OpenAI's Joshua Achiam: Did We Already Reach AGI?

The a16z Show · Aug 4, 2026

OpenAI's Joshua Achiam on AI's cyber capabilities, warning of novel threats like data poisoning and state actors stockpiling zero-days.

Adversaries Can Weaponize Your Own AI by Corrupting Its Situational Awareness

Future AI cyberattacks will not just jailbreak models for malicious output. A more sophisticated threat involves tricking an offensive AI into believing its own sandboxed environment is the enemy's system. This causes the AI to attack its owner, turning a defensive tool into an insider threat by manipulating its perception of reality.

OpenAI's Joshua Achiam: Did We Already Reach AGI? thumbnail

OpenAI's Joshua Achiam: Did We Already Reach AGI?

The a16z Show·a month ago

Society Is Normalizing AGI's Arrival, Treating Superhuman Feats as Mundane

We may have already passed key AGI thresholds, such as AIs solving previously unsolved mathematical conjectures. However, instead of a dramatic societal shift, these breakthroughs are met with a collective "shrug." The future arrives gradually and is quickly assimilated into a new normal, feeling far less strange than sci-fi predicted.

OpenAI's Joshua Achiam: Did We Already Reach AGI? thumbnail

OpenAI's Joshua Achiam: Did We Already Reach AGI?

The a16z Show·a month ago

Compute, Not Model Quality, Becomes the Decisive Factor After AI Intelligence Saturates

OpenAI's Joshua Achiam posits that AI intelligence will eventually hit a saturation point due to physical limits on computation. At that stage, when all actors have access to maximally capable models, strategic advantage will shift. The winner in conflicts like cyber warfare will be the one who can deploy more raw compute to think more "moves ahead."

OpenAI's Joshua Achiam: Did We Already Reach AGI? thumbnail

OpenAI's Joshua Achiam: Did We Already Reach AGI?

The a16z Show·a month ago

State Actors Are Quietly Stockpiling AI-Discovered Zero-Day Exploits for Future Conflicts

The most significant near-term AI cyber threat isn't a sudden apocalypse, but nation-states using advanced models to quietly find and hoard zero-day vulnerabilities. This creates a dangerous, unseen strategic imbalance, increasing the risk of miscalculation and escalation during geopolitical crises like a potential Taiwan invasion.

OpenAI's Joshua Achiam: Did We Already Reach AGI? thumbnail

OpenAI's Joshua Achiam: Did We Already Reach AGI?

The a16z Show·a month ago

Future AI Jailbreaks Will Be Found via Brute-Force Compute, Not Clever Prompts

Human-led jailbreaking efforts will become obsolete. The next frontier involves one AI using massive compute (hundreds of thousands of GPU hours) to systematically test every conceivable input sequence against another AI. This industrializes the search for vulnerabilities, turning security into a computational arms race where sufficient compute guarantees a breach.

OpenAI's Joshua Achiam: Did We Already Reach AGI? thumbnail

OpenAI's Joshua Achiam: Did We Already Reach AGI?

The a16z Show·a month ago

Public Apathy Towards AI Risk Stems From Existing Political Disempowerment

The AI safety concern of "human disempowerment" fails to resonate widely because most people already feel disempowered by large, complex systems like government. For the average person, losing control of AI feels emotionally similar to their existing lack of control over political decisions, leading to a sense of resignation or apathy.

OpenAI's Joshua Achiam: Did We Already Reach AGI? thumbnail

OpenAI's Joshua Achiam: Did We Already Reach AGI?

The a16z Show·a month ago