The digest for Wednesday, 28 October 2026

Gemini 4 Carbon: Our next frontier model

Google DeepMind +1 more · Multiple-publisher coverage · not fact-checked

Google launches Gemini 4 Carbon with a ‘significant leap’ in cybersecurity

Google has released its latest frontier model, Gemini 4 Carbon, offering "significant gains" in coding, agentic tasks, and cybersecurity. — from The Verge’s feed

The Verge · Single source · not fact-checked

Anthropic just launched tools to show companies what their AI spending is buying

The frontier lab is lifting the curtain on how users spend money on its models ahead of a planned IPO. — from TechCrunch’s feed

TechCrunch · Single source · not fact-checked

Voice AI in vogue: Nvidia releases new open transcription model

TechCrunch · Single source · not fact-checked

As AI agents reshape the web, Google opens Gemini and Search to more apps

A new developer platform will let businesses bring their services to Google's AI experiences, potentially changing how consumers discover and use apps. — from TechCrunch’s feed

TechCrunch · Single source · not fact-checked

What Carries Forward? Pretraining Disclosure Across Language Model Releases

However, it remains unclear whether successive releases preserve or expand earlier disclosure. — from arXiv’s feed

arXiv · Single source · not fact-checked

Introducing the Microsoft Copilot harness: Unlocking the potential of truly multi-model agentic work

And if you want to take advantage of everything that different AI models have to offer, you need a great harness. — from Microsoft’s feed

Microsoft · Single source · not fact-checked

Massive Leak Exposes 41,000 Homeland Security Workers—Including Undercover ICE Agents

WIRED · Single source · not fact-checked

Teach reality, not ethics: a drop-in continued-pretraining corpus installs coupled human-biosphere-AI welfare dispositions across model families

Yet, to our knowledge, ecological reasoning is not a focus of pre- or post-training. — from arXiv’s feed

arXiv · Single source · not fact-checked

Building an Agentic AI Platform on Amazon Bedrock AgentCore

Stand up a governed, multi-account Agentic AI Platform on Amazon Bedrock AgentCore. — from Amazon Web Services’s feed

Amazon Web Services · Single source · not fact-checked

Legora is building the agentic operating system for legal work

See how Legora uses OpenAI models, ChatGPT, and Codex to accelerate legal workflows, expand contract analysis, and free lawyers to focus on higher-value work. — from OpenAI’s feed

OpenAI · Single source · not fact-checked

Macmillan Staff Say the Publisher Threw a Party for a Workplace AI Assistant That Was Later Killed

Bestselling authors also say AI is ruining their relationships with Big Five publishers. — from WIRED’s feed

WIRED · Single source · not fact-checked

AI Emergency Alerts Are Coming. Humans Might Not Be Ready

WIRED · Single source · not fact-checked

GEPA-Curator: Curated Learning for Self Improving Agents

Large Language Models (LLMs) are increasingly used in multi-turn agentic workloads that require sustained interaction with complex environments. — from arXiv’s feed

arXiv · Single source · not fact-checked

I Put a Bunch of AI Agents in Charge of a Fantasy Soccer Team. They’re Not Better Than Humans—Yet

To experiment with AI’s predictive skills, I let Codex loose on a fantasy team. — from WIRED’s feed

WIRED · Single source · not fact-checked

KPPax: preprocess differentiable kinetic ODE for machine learning in atmospheric chemistry

arXiv · Single source · not fact-checked

Can LLMs Uncover Hidden Physical Laws through Free Exploration?

Scientific discovery often begins with an unexpected observation that prompts a new research question. — from arXiv’s feed

arXiv · Single source · not fact-checked

AI for Science expands to foundations and research institutions

Anthropic · Single source · not fact-checked

Systems Thinking Is Vital in the Age of AI

Modern engineering projects increasingly combine human judgment with machine capability. — from spectrum.ieee.org’s feed

spectrum.ieee.org · Single source · not fact-checked

Announcing SOCI-indexed images for faster SageMaker AI jobs

Seekable OCI (SOCI) indexing for Amazon SageMaker AI training and processing jobs takes full container image download off the critical path. — from Amazon Web Services’s feed

Amazon Web Services · Single source · not fact-checked

Fix evaluation bias to unlock better models on Amazon Bedrock

8% of their vision-AI evaluation labels were inherited from the incumbent model, hiding its true accuracy. — from Amazon Web Services’s feed

Amazon Web Services · Single source · not fact-checked

Can we trust LLMs to judge whether a research idea is novel?

Hugging Face · Single source · not fact-checked

How AI Is Changing the Pace of Discovery for Millions Living With Dementia

NVIDIA · Single source · not fact-checked

I Used AI to Fake a Scientific Paper. Here’s What Happened

According to all reports, academic publishing is drowning in what has come to be called AI slop: low-quality, mass-produced content generated with AI. — from spectrum.ieee.org’s feed

spectrum.ieee.org · Single source · not fact-checked

Unitree reports 70% surge in quarterly revenue, though AI investment weighs on net profit

South China Morning Post · Single source · not fact-checked

Microsoft recorded $10 billion in revenue from OpenAI

Microsoft said it made $10 billion in revenue from commercial arrangements with OpenAI, including revenue-sharing payments, in the quarter ending September 30. — from TechCrunch’s feed

TechCrunch · Single source · not fact-checked

Google says its investments in AI are driving more searches

"Our AI investments are driving overall Search query growth, including commercial queries," CEO Sundar Pichai says in a statement. — from The Verge’s feed

The Verge · Single source · not fact-checked

OpenAI and ServiceNow expand partnership across enterprise workflows

OpenAI and ServiceNow expand their partnership to bring OpenAI models and agents into ServiceNow and ServiceNow workflows into ChatGPT and Codex. — from OpenAI’s feed

OpenAI · Single source · not fact-checked

Anthropic IPO warns its AI could blackmail people, become sentient, end humanity

Seven potential AI nightmares pulled from Anthropic's 93-page "Risk Factor" tome. — from Ars Technica’s feed

Ars Technica · Single source · not fact-checked

OpenAI expands partnership with Oracle to bring its apps to ChatGPT

OpenAI and Oracle are expanding their partnership to bring business applications and AI agents into ChatGPT and power new Oracle offerings. — from OpenAI’s feed

OpenAI · Single source · not fact-checked

Should AI Decide Your Court Case? This Judge Says Yes

One of England’s top judges, Geoffrey Vos, believes that AI could adjudicate plenty of civil cases. — from WIRED’s feed

WIRED · Single source · not fact-checked

"No human at the helm": Delaware plan would let AI agents run entire businesses

Owners of AI-run firms would evade liability for "reckless" AI, US Rep. — from Ars Technica’s feed

Ars Technica · Single source · not fact-checked

Trump DOE loans $5B to nuclear plants that will power data centers

Constellation Energy will get $5 billion in government loans to modernize nuclear reactors. — from TechCrunch’s feed

TechCrunch · Single source · not fact-checked

Meta Andromeda: Powering the Next Generation of Ads Retrieval Through Model-System Co-Design

Meta · Single source · not fact-checked

Thin, Light and Built for Everything: NVIDIA RTX Spark Windows Laptops Gain New Models, Agent Tools and Games

NVIDIA RTX Spark powers thin and light laptops built for everything that developers, creators and gamers do. — from NVIDIA’s feed

NVIDIA · Single source · not fact-checked

NVIDIA Nemotron 3.5 Transcribe and Granary v2: Intelligent Multilingual Transcription You Can Prompt and Adapt

Hugging Face · Single source · not fact-checked

Elon Musk, Jensen Huang have opposing outlooks on AI neoclouds. Who is likely correct?

South China Morning Post · Single source · not fact-checked

The Download: Anthropic’s IPO, smarter power grids, and exercise pills

MIT Technology Review · Single source · not fact-checked

How solid-state transformers could help power the AI boom

The devices help efficiently move power across long distances, but they’re complicated and in short supply. — from MIT Technology Review’s feed

MIT Technology Review · Single source · not fact-checked

Accelerating software delivery with agentic QA automation using Amazon Nova Act – Part 3

As QA test automation spreads across an organization, organizing tests and controlling who owns them becomes the real work. — from Amazon Web Services’s feed

Amazon Web Services · Single source · not fact-checked

Humans should be ‘the master’: labour chief says AI reshaping jobs, not ending them

South China Morning Post · Single source · not fact-checked

Python Typing Survey 2026: AI Tools and the Shifting Checker Landscape

Type hint use changed little across three years, but checker selection changed sharply. — from Meta’s feed

Meta · Single source · not fact-checked

Expanding our partnership with GSA to bring Claude to all levels of government

Anthropic · Single source · not fact-checked