engblogs

An in-depth look at how Agentforce-powered AI security workflows in Security Center accelerate end-to-end incident response, from telemetry-driven investigations to scalable remediation playbooks.

Explains how WhatsApp's Scam Alert performs on-device, privacy-preserving scam detection within end-to-end encryption, using confidential federated analytics, differential privacy, and a verifiable transparency ledger to ensure model integrity and user control.

Explains Scam Alert, WhatsApp’s on-device, end-to-end encrypted scam-detection feature that warns users about likely scams with user control and privacy-preserving, verifiable analytics and transparent model delivery.

Lambda becomes the first private neocloud to secure an investment-grade-rated Term Loan B facility to fund GPU infrastructure for AI deployments, signaling strong institutional demand and a scalable asset-backed SPV financing model.

Lambda prices a $926 million senior secured term loan B to fund GPU infrastructure for its AI cloud platform, becoming the first private neocloud to achieve investment-grade financing in the term loan B market amid strong investor demand.

Reimagining Airbnb authentication with a server-driven, Identify first, then Challenge approach that uses a policy engine and a Challenge Picker to tailor login flows across regions, devices, and user contexts.

Airbnb's Flexible Authentication rethinks login for millions by applying an Identify first then Challenge flow powered by a server-driven UI, a policy engine, and a per-region Challenge Picker to adapt challenges to context.

Shows how Livelymerge introduces per-replica, $-prefixed local properties to keep ephemeral UI and per-session state out of the shared Automerge document, while preserving object persistence and GC semantics.

Introducing SL2T, a massively multilingual, on-device sign-language-to-text translation model that translates ASL to English using landmark-based vision, enabling real-world consumer applications while prioritizing privacy.

Deploy Daybreak frontier cyber models on AWS Bedrock via Daybreak Blue and Daybreak Red to accelerate vulnerability research, exploit validation, and incident response within existing AWS security and governance workflows.

Daybreak frontier models, including Daybreak Blue and Daybreak Red, are now available on AWS via Amazon Bedrock, enabling production-ready cybersecurity workflows within existing AWS environments.

Daybreak models are now available on AWS via Amazon Bedrock, with Daybreak Blue and Daybreak Red through Daybreak Access for authorized defensive security work—integrated into existing AWS environments to accelerate vulnerability research, exploit validation, incident response, and security testing under governance.