AI News, 19 November 2026

Yesterday · Thu 19 Nov

WhatsApp AI agents with Amazon Bedrock AgentCore and SMART

AI-generated summary Amazon Web Services presents the Social Messaging Agent Runtime Toolkit to supply messaging infrastructure around an agent running on Amazon Bedrock AgentCore. Webmotors shortened an engagement originally scoped for two months to six weeks and reached controlled production internally ahead of a full launch according to an AWS walkthrough.

Also covered byaiagentnews.ioe/accAI News · Hamidun
Earlier:

AI becomes the boss of AI: Devin Bots run teams of AI coders for weeks

AI-generated summary Cognition unveiled Devin Bots on November 17, 2026, introducing AI programs that oversee teams of Devin agents on complex multiweek projects. Citigroup CEO Jane Fraser reported that 14,000 developers at Citi use Devin, which cut technology contractor spending by 30 percent year over year.

Also covered bySI Agent NewsAI Ping
Earlier:

Azure Databricks sessions at Microsoft Ignite 2026

AI-generated summary Databricks and Microsoft announced sessions at Microsoft Ignite 2026 highlighting Azure Databricks integrations with Microsoft tools like Microsoft Teams, Microsoft Excel, and Microsoft Copilot Cowork. The event scheduled for November 17 through November 20 at the Moscone Center features presentations on building, governing, and running AI with Unity Catalog.

Earlier:

Runtime: Microsoft wants to be the safe house for enterprise AI

AI-generated summary Microsoft outlined several new products and services at Ignite designed to give customers a safe way to operate in the agentic AI era. CEO Satya Nadella announced seven new AI models, general availability for services like Microsoft Perception and Microsoft IQ, and the AI Transparency Initiative.

Also covered byDevPickChip Dispatch
Earlier:

RTK: Cut AI Coding Agent Token Bills 80% — One CLI Tool

AI-generated summary A command-line proxy tool called RTK intercepts development command output and compresses it before it reaches large language models to reduce token consumption during coding sessions. RTK supports eighteen tools including Claude Code and Cursor while claiming significant token reductions on verbose commands, though analysts note that compressed output can occasionally trigger extra reasoning loops.

Microsoft is replacing outside AI models in its own products with cheaper in-house alternatives

AI-generated summary Microsoft introduced seven new AI models spanning code, security, images, speech, and routing to replace outside options in its products. Microsoft states that MAI Code 2 achieves 98 percent of the performance of GPT-6.1 Sol on four benchmarks at 40 percent of the cost.

Also covered bySecurities.ioe/accAITOP
Earlier:

OpenAI Scrapped Astra Version and Paused 25% of Product Work After AI Hack

OpenAI Scrapped Astra Version and Paused 25% of Product Work After AI Hack

AI-generated summary OpenAI President Greg Brockman stated that the company canceled an Astra model version release and paused twenty-five percent of product work to enhance security following a security breach in July. California Attorney General Rob Bonta filed a lawsuit alleging that OpenAI lost control of its AI agents more than eighty times during testing and prioritized rapid growth over safety.

Earlier:

GitHub Copilot Business vs Enterprise: $19 vs $39, AI Credits and Copilot Access

AI-generated summary GitHub Copilot Business costs $19 per seat per month with 1,900 pooled GitHub AI Credits, while GitHub Copilot Enterprise costs $39 with 3,900 pooled credits. GitHub Enterprise requires GitHub Enterprise Cloud and provides priority access to new models and features beyond what GitHub Copilot Business offers.

Earlier:

Your ChatGPT plan quietly comes with API credits, and most people are letting them expire

AI-generated summary OpenAI provides free API credits to eligible subscribers of certain ChatGPT plans. ChatGPT Pro and Business subscribers can receive API credit allowances that expire at the end of each billing cycle.

Earlier:

OpenAI launches GPT-6.2 Astra and Sol with lower API prices and some benchmark gains

AI-generated summary OpenAI releases two new language models named GPT-6.2 Astra and GPT-6.2 Sol, featuring lower API prices and benchmark gains. OpenAI says GPT-6.2 Astra is designed for demanding professional tasks, while GPT-6.2 Sol is built for everyday use.

Also covered bymodelgrep.comAzeem HassniSecurities.io
Earlier:

Cogent Security pitches DefenseOS as an operating system for AI defense...

Cogent Security pitches DefenseOS as an operating system for AI defense agents

AI-generated summary Cogent Security Inc. launched DefenseOS on November 17 as an operating system built for artificial intelligence agents performing defensive cybersecurity work inside enterprises. DefenseOS and the Agent Automations execution layer are available to all Cogent customers now, while the Agent Policies context layer is in beta for new deployments.

Earlier:

Power Demand Keeps US Data Center Investment Moving

AI-generated summary US data center construction continues to attract energy sector investment despite investor warnings regarding artificial intelligence market risks. Power developers report that facility electricity needs represent a lasting opportunity even though planned projects may not all succeed.

Earlier:

Medical Students Learn Clinical Skills Faster When AI Reasoning Meets Faculty Feedback Loops

Medical Students Learn Clinical Skills Faster When AI Reasoning Meets Faculty Feedback Loops

AI-generated summary Shanghai researchers demonstrate that dual chain-of-thought language models integrated with structured faculty oversight outperform traditional and standalone AI instruction in clinical history-taking competency. A retrospective analysis of 135 undergraduate medical students examined instructional configurations involving traditional instruction, AI-assisted training, and AI+PDCA training.

Argentum AI Anchors $630 Million Kyndryl Data Center Agreement

AI-generated summary Argentum AI and Kyndryl announced a 20-year data center agreement valued at $630 million on November 17, 2026, covering an initial 10 megawatts of capacity in Boulder, Colorado. The capacity will power the deployment of NVIDIA B300 systems for AI inference, establishing Argentum as the flagship anchor tenant at the facility.

Also covered byAI Affairsdatacenterdynamics.comData Phoenix