Podcasts

In this podcast episode James Wilson chats with Tinfoil co-founder Tanya Verma about how you can run a powerful LLM in the cloud without the inference provider seeing your prompts.

Tanya talks James through how private inference works, from trusted execution environments and hardware attestation, to TLS termination and GPU isolation. Customers can verify the exact code and model processing their data, while Tinfoil and its infrastructure providers remain locked out.

That’s clever engineering… but who really needs it? Is private inference only useful if you’re doing something bad, or will it become a privacy baseline like TLS? James and Tanya discuss the costs and trade-offs, and how open weights make private inference more transparent and trustworthy.

A Meta AI model also escaped a testing sandbox, a cyberattack disrupts ports in North Carolina, the Philippines will establish a cybersecurity agency, and a Ransom Cartel admin gets 16 years in prison.

Tom Uren and James Wilson talk about North Korea losing control over some of its hacker workforce. Expect some tightening of controls and oversight, and perhaps even a reduction in the country’s ransomware operations.

They also discuss escalating attacks on American water infrastructure. Although the impact of these attacks is relatively minor so far, the US government has been slow to respond from a political perspective.

This episode is also available on YouTube

On this week’s show Patrick Gray, and James Wilson are joined by bearded man of leisure Adam Boileau to discuss the week’s cybersecurity news, including:

• Accidental AI agent hacking sprees have the world’s media freaking out, but we think it’s all pretty funny
• The bugpocalypse is so chaotic, Microsoft can’t patch fast enough
• A ColdCard wallet flaw led to millions in Bitcoin theft, but the back story behind the bug is bonkers
• Iran hacks and disrupts water infrastructure in multiple American states
• North Korea’s state-backed hackers turn criminal. Or their criminals turn into state-backed hackers. Or something. It’s all a bit confusing, actually.
• Much, much more!

This week’s show is brought to you by Sondera. Co-founder Josh Devon joins Patrick and James to talk through some absolutely hilarious LLM horror stories.

This episode is also available on YouTube

A hacker breached Hungary’s State Treasury, Russia will mandate 40 apps on all smartphones next year, hackers steal Liechtenstein’s business database, and an AI agent got real CVEs for hallucinated vulnerability reports.

In this edition of Between Two Nerds Tom Uren and The Grugq talk about whether hacker culture is inherently anti-authoritarian and how different states get their country’s hackers to work for the state.

This episode is also available on YouTube.

Anthropic models also did the hacky-hacks, Coldcard was hacked for $70 million in Bitcoin, npm adds publish-time malware scanning, and Russia is behind the recent hotel WiFi hacks.

In this sponsored interview James Wilson chats with Permiso CTO Ian Ahl about detecting ShinyHunters-style attackers as they move through cloud and SaaS environments.

Ian explains how ordinary-looking events such as a password reset, a new MFA device, unusual searches and a first-time AWS role assumption can combine to reveal an intrusion. Permiso’s platform connects these signals across identity providers, cloud platforms and SaaS applications. They also discuss how AI is helping attackers move from initial access to extortion in just four hours.

A non-profit puts a $22,000 bounty on the INC ransomware group, hackers breach the UK Department for Education, Russia charges Telegram founder Pavel Durov, and the FCC bans foreign robots and power inverters.

Tom Uren and James Wilson talk about open-weight AI models and distillation. These topics have been subject to a lot of US government attention in recent weeks, but let’s not forget that America’s overriding goal is to remain ahead of China in the AI race. There are better ways to do that than overindexing on distillation.

They also discuss Iranian attacks on US critical infrastructure. Given that the war in Iran is unpopular, incidents that make headlines without causing serious impact are perfectly calibrated.

This episode is also available on YouTube