Risky Bulletin

Tom Uren and Amberleigh Jack talk about new research that shows the Chinese-made DeepSeek-R1 AI model produces insecure code when prompts include topics that the Chinese Communist Party dislikes. It’s interesting research, but the CCP doesn’t have a monopoly on imposing AI bias.

They also discuss the complete doxxing of the Iranian cyber espionage group known as APT35 or Charming Kitten.

This episode is also available on Youtube.

In this edition of Between Two Nerds Tom Uren and The Grugq talk about the differences between telcos and cloud companies. Does the nature of the business force cloud companies to be better at security?

This episode is also available on Youtube.

Tom Uren and Amberleigh Jack talk about Anthropic’s discovery of an “AI-orchestrated” cyber espionage campaign. To Tom, it feels a research project, but it’s pretty clear it will be really useful for threat actors that aren’t focussed on specific high-priority targets. Think ransomware, Chinese intellectual property theft and North Korean hackers. But it won’t be so good for Western intelligence agencies.

They also discuss Google’s legal disruption of the China-based Lighthouse phishing as a service operation. Surprisingly, it seems to be working!

Finally, they talk about why the memory safe Rust language has been a triple win for Android.

This episode is also available on Youtube.

In this edition of Between Two Nerds Tom Uren and The Grugq talk about the strategic “logic” of Russian wiper attacks on the Ukrainian grain sector.

This episode is also available on Youtube.

Europol takes down servers behind three malware operations, the US sanctions another Burmese military group linked to scam compounds, Google backs down from mandatory Android developer registration, and Checkout-dot-com donates its ransom to cybercrime researchers instead of paying hackers.

Tom Uren and Amberleigh Jack talk about a new Reuters’ report that reveals how Meta is knowingly raking in cash from scam advertisements. It’s around $16 billion worth, and in documents Meta calculates that it outweighs the costs of possible regulatory action.

They also discuss recent state-backed supply chain attacks that have, so far, remained targeted and responsible. Finally they look at the UK’s decision to stop sharing intelligence with the US about suspected drug boats in the Caribbean.

This episode is also available on Youtube.

Internal data leaks from another Chinese security firm, a US Congressional Budget Office breach has not been contained, the Cyber infosharing act likely to be extended until January, and we have a new OWASP Top 10.

In this edition of Between Two Nerds Tom Uren and The Grugq discuss how cyber criminals and even state actors are being dumb about using AI.

This episode is also available on Youtube.

Myanmar starts demolishing the KK Park scam compound, the US Congressional Budget Office gets hacked by a foreign APT, Chrome will remove risky X-S-L-T support, and scammers in Singapore will get the cane.

In this sponsored interview Casey Ellis chats to Toni de la Fuente, founder and CEO of Prowler, an open source platform for cloud security. They chat about how and why Prowler selectively applies AI to ensure it adds value rather than just because they can.