SN 1072: LiteLLM

TWiT Shows Security Now
1

New Security Now! is live:

We’re diving into a supply chain attack that almost slipped past everyone—until one developer caught the malicious code red-handed. Scary stuff.

In this episode:

:police_car_light: LiteLLM’s close call—how dangerous code nearly hit millions of AI systems
:locked_with_key: Apple’s new age verification headache in the UK (and what it means for California)
:globe_with_meridians: Russia’s homegrown 5G encryption and Ukraine’s discovery of spy webcams
:alarm_clock: Google pushes quantum computing threats to 2029 (more breathing room than we thought)
:robot: AI-generated spam is taking over Reddit, plus the latest on ClickFix malware

This one’s a solid mix of near-misses, international intrigue, and the AI mess we’re all watching unfold.

Check it out and let us know what you think.

#SecurityNow #Cybersecurity #TWiT

2

It’ll be interesting to see what happens with future LLMs that are trained on public discussion sites as those sites are largely taken over by the very LLMs that train on them! I think the feedback loop will cause some wild hallucinations. Personally, I can envision this phenomena as a net-positive. People will realize that anonymous, public comment sections are not great sources of information, and hopefully they’ll turn to bona-fide sources. Maybe even a resurgence in journalism.