Tech news, October 22, 2026

LeRobot trains directly from Lance datasets in object storage, a hybrid scan mines robot fleet data for edge cases, and distributed search reaches 1,066 QPS across 10B vectors, pl…

SR 26-2 left generative AI and AI agents out of bank model risk guidance.

In March and April 2023, the Python Software Foundation (PSF) received three (3) subpoenas for PyPI user data.

We are experiencing an impairment on the Bandwidth network which may affect a portion of calls.

Posted by Christian Fischer on Oct 21 Hello, > On 21.10.2026 16:44 CEST Jakub Zelenka <bukka () php net> wrote: > > Affected versions are PHP 8.2 before 8.2.35, PHP 8.3 before 8.3…

Replacement: gemini-omni-1.1-flash.

PyTorch Conference North America 2026 brought 3,500+ attendees to San Jose on October 20–21 for two days focused on accelerating open intelligence and exploring the impact of PyTo…

Preliminary report detailing Railway's October 21st security incident.

Every Pro team can claim a free domain for one year.

Copilot usage metrics now show the names of the skills, custom agents, Model Context Protocol (MCP) servers, slash commands, and plugins that your organization creates, instead of…

Samartha Chandrashekar & Nick Parker You click play on a Friday night, and in the second that follows, dozens of services answer in milliseconds.

Google Vault now supports eDiscovery filtered exports for Gemini in Workspace.

See how public sector data modernization helps agencies replace legacy systems, unify data and build a secure foundation for AI readiness.

Safari Technology Preview Release 255 is now available for download for macOS Golden Gate and macOS Tahoe.

Amazon Connect Customer Profiles now allows administrators to customize which customer data that agents see in the agent workspace, and how it is displayed, using layouts they bui…

EDB runs its own revenue reviews, renewals, and executive metrics on the same data and AI platform it sells.

Every organization has unique requirements for building and scaling generative AI.

Today, AWS announces the general availability of AWS Continuum for code vulnerabilities.

WarpStream questions answered — architecture, BYOC vs.

Cloud block storage costs 24x or more per GiB than S3.

Amazon OpenSearch Serverless now offers intelligent ingestion, a fully managed feature that turns logs and documents stored in Amazon S3 into searchable content in Amazon OpenSear…

In September, we announced a global default policy for generally available GitHub Copilot features and supported client capabilities on Copilot Business and Copilot Enterprise pla…

Explore how Cohere Model Vault secures AI inference with confidential CPU and GPU computing, end-to-end encryption, and remote attestation.

AWS Continuum for code vulnerabilities is a new security service that uses AI to automatically find, validate, and prioritize code vulnerabilities in your environment, then delive…

Anyone who has published an app knows that dreaded message from a user: “Doesn’t work for me”.

New open weights from NVIDIA on the Hugging Face Hub.

New open weights from Tencent Hunyuan on the Hugging Face Hub.

New open weights from Tencent Hunyuan on the Hugging Face Hub.

New open weights from Tencent Hunyuan on the Hugging Face Hub.

New open weights from Tencent Hunyuan on the Hugging Face Hub.

Posted by Abhinav Agarwal on Oct 22 CISA has published four CVE records and CSAF advisory VA-26-294-01 for vulnerabilities in HAPI FHIR, an open-source Java implementation of HL7…

Reduce triage times and focus on the risks that matter most by combining context across code, infrastructure, and runtime

Posted by Alan Coopersmith on Oct 22 -------- Forwarded Message -------- Subject: [Security-announce][CVE-2026-114583] tarfile.tar_filter path traversal bypass allows writing outs…

Posted by Maurits van Rees (Plone) on Oct 22 On behalf of the Plone/Zope Security Team I announce an information disclosure vulnerability in `plone.app.content`.

On October 21st, 2026, the PHP project released emergency updates for two critical vulnerabilities tracked as CVE-2026-109618 and CVE-2026-109619.

Posted by Valtteri Vuorikoski on Oct 22 Having spent the day whacking this particular molehill, my strong opinion is that these packages (minus certifi) should stop vendoring root…

Overview Two memory-related vulnerabilities have been identified in PHP versions prior to 8.2.35, 8.3.36, 8.4.27, and 8.5.12.

The Beta channel has been updated to 157.0.8103.15 for Windows, Mac and Linux.

Bulletin ID: 2026-147-AWS Scope: AWS Content Type: Important (requires attention) Publication Date: 10/22/2026 13:30 PM PDT Description: aws-c-mqtt is a C99 implementation of the…

There is a MEDIUM severity vulnerability affecting CPython.

The post Microsoft and the Public Library Association expand free AI literacy training appeared first on Source.

1.

The Amazon Polly Generative engine gets a major upgrade: more expressive, more accurate speech with low latency, plus 14 new voices across 9 locales and general availability of Hi…

Learn how IskayPet, a leading pet care company in Iberia, built a multilingual corporate knowledge assistant on Amazon Bedrock Managed Knowledge Bases.

Amazon SageMaker AI serverless model customization now includes training sessions, generally available today.

Amazon SageMaker AI training session APIs let you write your own LLM training loop (SFT, RL, or custom objectives) in a Python script while the service runs the GPUs.

Amazon Quick is a HIPAA-eligible service, but eligibility is not the same as compliance.

Voice agents often wait for an agent's full response before starting speech synthesis, adding latency users feel directly.

GPU applications are processing ever-growing datasets with increasingly powerful accelerators.

Teams deploying custom text-to-speech models on Amazon SageMaker AI often guess at autoscaling thresholds, and a wrong guess either wastes GPU spend or breaches a latency SLO.

“X-Hinges” helps design and 3D print items like lamps and gloves with an interactive twist: automatically embedded sensors that perceive and adaptively respond to human contact.

Following the People's AI Assembly, acclaimed author Arundhati Roy joins Amy Goodman and Nermeen Shaikh to discuss the rise of artificial intelligence, which she calls "capitalism…

The challenge Last month, I sat down with a CIO for a manufacturing company in Johannesburg.

The Department of the Treasury, as part of its continuing effort to reduce paperwork and respondent burden, invites the general public and other federal agencies to comment on the…

npj Digital Medicine, Published online: 22 October 2026; doi:10.1038/s41746-026-03349-z Large language models for multimodal clinical predictive modeling: a systematic review

Elastic GPUs, sandbox growth, and the economics of the agent cloud

It was the probably worst day in cybersecurity history.

What I learned from touring Intel’s most advanced fab and talking to its leadership.

A month’s worth of security breaches in a single day – will things get worse before they improve?

How Moshi turns market data into actionable insights, and why AI tools have to fit the trader’s workflow.

Looking more closely at the function's behavior and requirements.

Controllable 3D part generation is pivotal for digital asset creation, yet existing methods lack the flexibility to accommodate diverse user workflows requiring varying levels of…

Robotic foundation models are typically evaluated through costly closed-loop rollouts, which motivates cheap offline proxies for model capability.

Video foundation models have advanced rapidly in understanding and generating visual sequences, while video world models seek to capture the underlying dynamics of the world they…

Test-time scaling has emerged as a key inference strategy for reasoning models, where a model generates a large number of candidate solutions in the hope of finding at least one c…

Constructing accurate and computationally tractable approximation spaces is a fundamental task in scientific computing.

We consider online monitoring of an AI system that produces a sequence of decisions.

Auditing whether a text-to-image (T2I) model was fine-tuned on a particular image collection is difficult: a model can absorb collection-level visual characteristics without repro…

Wrong automatic routing of an IT support ticket can cost more than human triage.

Meta-reinforcement learning aims to train agents that adapt quickly to unseen tasks from limited interaction, but this requires representing task evidence over long interaction hi…

Frontier coding agents are becoming increasingly adept at 3D modeling.