AI
Anthropic and Google's what-passes-for-standards in agentic AI has everyone scrambling.
Large Language Models can be backdoored by introducing just a limited number of “poisoned” documents during their training, a team of researchers from the UK’s Alan Turing Institute and partners found. “Injecting backdoors through data poisoning may be easier for large models than previously believed as the number of
Other vendors "are handing you the pieces, not the platform", says Google, as it launches one chat to rule them all.
Pinning down Huawei's supply chain puts you right up there with those selling weapons to Taiwan, it seems.
Updated at 2:26 p.m. on October 15, 2025 to remove an incorrect reference by Chris Wyosopal to kernel structs and verifier hooks in the presentation example code. In August, a relatively unknown security researcher named Agostino “Van1sh” Panico gave a talk at hacking conference Defcon. The 45-slide deck
Atlassian is giving JSM agents free customer service management, cloud-only of course
Data scientists at AWS have created a “novel algorithm” to help tackle a sticky industry problem: how to pinpoint errors in multi-agent systems. They have turned, in part, to “regex pattern matching,” to help do this; saying doing this via LLMs themselves was “computationally intractable.”