New research shows frontier models outperform human scientists in troubleshooting virology procedures—lowering barriers to the development of biological weapons.

For years, people have cautioned we wait to do anything about AI until it starts demonstrating “dangerous capabilities.” Those capabilities may be arriving now.

Virology knowledge has been limited to a small number of experts. Expertise in dual-use fields like virology is difficult to attain, with people completing multiple degrees and dedicating their careers to reaching the forefront of research. Where knowledge is publicly available, the jargon-heavy literature is largely indecipherable to most people outside the field. To perform research involving biosafety level 3 (BSL-3) pathogens—such as SARS, anthrax, or H5N1 influenza—researchers must clear a series of approvals, including facility certification, security clearances, specialized training, and ongoing medical surveillance. Only then can they get access to these pathogens and begin acquiring the tacit skills needed to work with them

These high barriers to entry have limited the pool of people with access to powerful dual-use knowledge, keeping the chances [...]

---

Outline:

(07:19) Friction as a key factor

(10:47) A ticking time bomb

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

Corporate capture of AI research—echoing the days of Big Tobacco—thwarts sensible policymaking.

Make a testable claim.

Test it through observation and experiment.

Update the claim to match reality.

The common-sense notion that we should assess claims about the world through observation appears in texts dating back to ancient civilizations. But it was the Scientific Revolution of the 16th and 17th centuries that formalized and popularized what we now call the scientific method.

We are all steeped in this paradigm. And, at first glance, it seems entirely reasonable to expect that AI policy be firmly grounded in evidence. Many proponents of this view argue there simply isn’t enough proof to justify alarm over various AI risks: from AI systems perpetuating bias and discrimination, to enabling the development of biological weapons by malicious actors. Precautionary measures, they claim, are often premature, overly restrictive, and even unscientific or unethical — likely to stifle innovation and block the benefits AI could bring to society.

Yet [...]

---

Outline:

(01:29) Black Boxes and Benchmarks

(05:02) Addressing Corporate Incentives

(09:21) The Path to Better Evidence

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

The U.S. social insurance system remains stuck in the past, unprepared for the scale and speed of AI-driven displacement. We need a new social safety net, one designed to support workers whose jobs are displaced by AI technologies.

Rapid advances in artificial intelligence promise to reshape industries, redefine job roles, and fundamentally alter the landscape of the global economy. No longer confined to manufacturing, AI-driven tools are encroaching on white-collar professions once thought immune to automation — law, accounting, marketing, finance, design, insurance, human resources, and even software engineering.

Yet, the United States’ social insurance framework remains outdated and ill-equipped to handle this disruption. Workers, their families, and entire communities stand on the edge of an economic cliff, vulnerable to sudden AI-driven job displacement. Rather than scramble to respond after mass layoffs occur, policymakers must begin modernizing social insurance now. The Social Security Act of 1935 was designed for a different era—one that bears little resemblance to the economic realities of 2024.

Even if AI ultimately spurs job creation, AI's rapid acceleration risks leaving displaced workers in a prolonged period of unemployment before new opportunities emerge—if they [...]

---

Outline:

(03:37) The Inadequacy of Current Social Insurance

(06:42) The Challenges of AI-Driven Unemployment

(08:50) AI Displacement Insurance: A New Social Safety Net

(09:40) Increasing Participation

(10:24) Tailored Benefits

(11:15) Innovative Funding Mechanisms

(13:50) Toward a Resilient Social Insurance System in the Age of AI

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

Continued sales of advanced GPUs to China raise concerns over America's position in the global AI race.

Chips are the backbone of modern artificial intelligence. One chip in particular, Nvidia's H20, has become a flashpoint. Developed specifically for the Chinese market, the H20 was engineered to avoid existing US export controls while preserving key capabilities that make it highly attractive for advanced AI systems.

National security experts and AI policy leaders have urged the US government to treat chips like the H20 as a strategic asset. And until just days ago, it appeared the Trump administration was prepared to act—moving to block their sale to China.

Following a dinner at Mar-a-Lago attended by Nvidia CEO Jensen Huang, the administration reportedly paused plans to impose further restrictions on H20 chip sales to China. Sales will continue. Experts argue the decision undermines a policy Trump has championed: to ensure America, not China, leads the world in AI.

A Key Shift in AI

Sales of the H20 chip carry [...]

---

Outline:

(01:09) A Key Shift in AI

(02:59) Americas Last Advantage

(05:07) Profits or Policy?

(06:39) The Worst Possible Time

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

AI risk may have unique elements, but there is still a lot to be learned from cybersecurity, enterprise, financial, and environmental risk management.

Malcolm Murray — Apr 9, 2025

Image: MchlSkhrv / iStock

Unlocking AI's potential means prioritizing hope for what can go right over fear for what could go wrong.

Kevin Frazier — Apr 9, 2025

Cover image: The production line at a semiconductor factory. Credit: SweetBunFactory / iStock