Python stories
Australia's developers are contributing more widely abroad, with GitHub data showing a 16% quarterly rise in cross-border open-source collaboration.
Developers may get patched code faster as Google's preview agent scans repositories, tests exploits and drafts fixes for review.
Security and compliance teams can now patch buildpack-based containers from a single hardened base, rather than chasing Dockerfiles across repositories.
Developers can now use Google Cloud's examples to build and govern Gemini-powered agents for approvals, compliance and data workflows.
Labour shortages and an ageing workforce are pushing Japanese manufacturers to test AI-controlled robots in factories, logistics and hospitals.
Notebook users can now query local pandas data with SQL and pass results back to Python, reducing data shuffling in analytics workflows.
Exposed serverless apps can let attackers steal tokens, read secrets and take over cloud projects if weak code is left unpatched.
The update gives enterprises managed access to Anthropic's models with built-in failover, data residency and compliance controls for production AI use.
Developers are using the queue to smooth AI traffic and protect services from spikes, as AWS adds higher throughput, security and recovery tools.
Enterprises can now patch older open source software without disruptive upgrades, as IBM and Red Hat target stubborn vulnerability backlogs.
Researchers can now analyse data, search literature and draft papers in one place, with outputs kept reproducible on their own infrastructure.
The release aims to ease a key hurdle for firms moving AI agents into production by unifying memory, retrieval and access across environments.
The expansion gives IT teams central control over AI agent permissions, reducing risky static keys and easing reviews as workplace use widens.
The move aims to help defenders turn faster vulnerability discovery into working fixes, as OpenAI broadens access to its cyber tools and partners.
Trusted third-party access has let attackers quietly pull large volumes of Salesforce records from enterprise systems via a Klue integration.
More than half of Vercel deployments are now triggered by coding agents, as monthly AI token traffic has jumped tenfold.
Millions of downloads were exposed to silent code execution as a flaw in Hugging Face Transformers let malicious models run on load.
CrowdStrike said state-backed espionage and extortion are surging as AI assets inside tech groups draw hackers seeking code, models and access.
The compact desktop aims to cut cloud costs for AI developers by letting them fine-tune and run large models locally on Windows.
Enterprises can now vet open-source dependencies before build time, as the catalogue adds independent malware and vulnerability checks for Python and Java.