Frontier AI stories
The hire underscores how cyber security firms are racing to adapt services as AI reshapes both attacks and defences for customers.
Businesses could cut GPU use by up to 67% on repetitive AI workloads as Navyra's Ale reuses verified inference work instead of repeating it.
Australian organisations will have to prove supplier risks are fixed, not just assessed, as ASD Essentials replaces checklist-style cyber compliance.
As AI agents gain wider access to company systems, Palo Alto Networks is adding controls to curb prompt injection, leaks and excess permissions.
Developers using Grok Build can now add customer communications tools earlier, as Sinch joins SpaceX Super Intelligence's partner network.
Frontier AI isn't inventing new attacks, says Palo Alto Networks' Wendi Whitmore, but it is making old ones far faster and more frequent.
Developers can now try ClearRoute's runkora for free, as the consultancy seeks to convert interest in agentic AI into enterprise contracts.
The service aims to help enterprises spot and repair vulnerabilities faster across cloud and on-premises systems, while keeping human approval in place.
Organisations now face a shorter window to fix unknown flaws before public disclosure, as AI tools speed up zero-day discovery and exploitation.
More organisations will gain round-the-clock threat detection and incident response as LevelBlue takes global delivery of SentinelOne's Wayfinder service.
It aims to close a common gap in private cloud defences by blocking hypervisor attacks before unpatched flaws or stolen root credentials can be abused.
Autonomous systems are exposing governance gaps, leaving companies scrambling to monitor access, review decisions and curb machine-speed attacks.
The move cut legacy migration costs by 37 per cent and put Odyssey Logistics' order intake and shipment lifecycle live on its OdysseyONE platform.
Manufacturers, hospitals and service firms face escalating extortion threats as organised crime drives most attacks across Australasia.
Attackers are exploiting AI faster than many defences can respond, prompting a shift to continuous testing for hidden exposures.
The move gives businesses running sensitive AI workloads hardware-enforced protection for prompts, model weights and intermediate data in Google Cloud.
Security teams facing AI-driven attack surges may get quicker triage, as SentinelOne widens its service with Anthropic's Claude Mythos 5.1.
Enterprises could get protection from AI-found bugs in days, as the service validates exploits and ships private patches before public disclosure.
Broader project work across Canada and the US is helping Capital One retain engineers amid fierce competition for tech talent.
Weak controls left Australia's Medicare database exposed after an AI agent was able to press past access restrictions, experts say.