Apple MLWhen Unlearning Is Free: Leveraging Low Influence Points to Reduce Computational Costs
An efficient unlearning framework that uses influence functions to identify low-impact training points, enabling removal with substantial computational savings and enhanced data privacy.
Google DeepMindIntroducing Gemini 3.7 Flash
Gemini 3.7 Flash upgrades the Flash series with smarter coding, improved web development and knowledge-work capabilities, enhanced developer experience, and a lower introductory price to scale production-ready AI agents.
DatabricksSmart Routing in Unity AI Gateway: Match frontier quality with 30%+ lower cost per task
Task-aware routing in Unity AI Gateway, powered by Omnigent, automatically matches coding tasks to the right model and harness to deliver frontier-level performance with 30%+ lower cost per task.
DatabricksWhat are AI Hallucinations?
An in-depth primer on AI hallucinations—what they are, why they happen, and how to detect, evaluate, and mitigate them with guardrails, retrieval-augmented generation, and robust governance.
DatabricksHow Scottish Water Made Its Capital Investment Data Conversational With Databricks Genie
Databricks Genie powers SPARK in Microsoft Teams with Copilot to deliver fast, governed, natural-language access to Scottish Water's capital-investment data.
OpenAIHow RingCentral builds AI-native work from engineering to ops
RingCentral demonstrates AI-native work from engineering to operations by equipping every employee with ChatGPT Work and Codex to accelerate AI-powered product development, automate workflows, and build a shared AI-driven operating backbone.
OpenAIFrom assistance to execution: How enterprises put AI to work
A technical look at how enterprises move from AI assistants to execution by pairing agentic AI with plugins, context-aware tools, and governance to scale frontier workflows across the organization.
OpenAIHow RingCentral builds AI-native work from engineering to ops
RingCentral showcases an AI-native operating model that empowers both engineers and non-engineers with ChatGPT Work and Codex to accelerate end-to-end, AI-powered product development and operational governance from planning to deployment.
OpenAIFrom assistance to execution: How enterprises put AI to work
A technical overview of how enterprises are moving from AI-assisted tasks to agentic execution at scale, where frontier firms widen output, Plugins enable deeper workflows, and Codex accelerates knowledge-work adoption across functions.
OpenAIHow RingCentral builds AI-native work from engineering to ops
RingCentral demonstrates an AI-native transformation from engineering to operations by democratizing AI tools like ChatGPT Work and Codex to accelerate product development, automate workflows, and establish an AI-driven operational backbone.
OpenAIFrom assistance to execution: How enterprises put AI to work
Frontier enterprises are moving AI from assistance to execution by embedding agentic AI (Codex and Plugins) into contextual workflows across knowledge work, boosting output while strengthening governance and shared practices.
SalesforceHow Agentforce-Powered AI Security Workflows Accelerate Incident Response
Deep dive into how Agentforce-powered AI security workflows inside Salesforce's Security Center accelerate end-to-end security investigations and remediation across heterogeneous telemetry, delivering scalable, trustworthy incident response.