Daily 'AI for Work' Pulse: 8th of November

Daily AI for Work Pulse - 2026-10-02

Latest AI Tools

All AI News from Yesterday

24 News Articles published.

Latest AI News for Creatives

Box partners with Nvidia to build faster, cheaper AI agents into its platform

Box embeds Nvidia Nemotron models into its platform, promising queries run four times faster with 20% fewer errors and 90% less token use. The conceptual search goes beyond keywords to find relevant document chunks across millions of files, starting early 2024 for Enterprise Advance customers.

WeLM-Image ranks first among evaluated open-source models on three image generation benchmarks

Tencent built a 31.7-billion-parameter image model that runs on a single GPU and tops three major benchmarks. It generates a 1024×1024 image in under 5 seconds on one NVIDIA H20 GPU.

VFlash speeds up robot motion forecasting by 25x with a new visual planning method

NVIDIA and Johns Hopkins researchers released VFlash, a method that generates spatial predictions in one pass instead of token-by-token. It cuts 3D point motion forecasting time by roughly 25x while raising task success from 62.6% to 70.2%.

Ai2 introduces ToyBlocks, a controlled simulation framework for robot learning research

ToyBlocks generates 85,336 robot demonstrations and 144 million frames on a single GPU without human teleoperation. The open framework lets researchers trace exactly why a model succeeds or fails, a capability still rare in robot learning.

Meta Extends Active Memory and Recovery Models to All Muse Users

Meta says its new memory models for Muse help with long chats and changing tasks, but it hasn’t shared full test results, so companies should verify the claims themselves.

AI spending flatlines despite rising token volumes as price wars and thriftier customers undercut revenue growth

AI token spending grew only 1.5% across August and September, a sharp drop from July's 42% jump. Price cuts from OpenAI and Anthropic, including a roughly $600 million revenue hit for Anthropic, are capping revenue even as usage rises.

SageMaker inference adds Nvidia multi-instance GPU support to pack more models onto each endpoint

Amazon SageMaker Inference now supports NVIDIA MIG GPU partitioning, letting you carve a single GPU into up to seven isolated instances. Testing showed a 3/7 slice of an H100 delivered up to 89% more output tokens per second and cost 18-34% less per million tokens than a whole A100.

Future Secure AI raises $80 million at $1.2 billion valuation to expand globally

Future Secure AI raised US$80 million at a $1.2 billion valuation to expand its AI agent teams for investment banking, consulting, healthcare, and law. The company expects annual recurring revenue to near US$100 million by end of 2026 ahead of a planned 2027 ASX listing.

Anthropic's AI r&d pace doubles since late 2025, external evaluators find

Anthropic has roughly doubled its AI capability improvement rate since October 2025, with benchmark growth surging to 1.9 times the historical pace. METR estimates about 80% of that acceleration comes from the company's own AI assisting its R&D.

Foundry Local brings agentic coding to developer hardware with Qwen3.8-27B

Microsoft released Foundry Local, letting developers run the Qwen3.8-27B coding model on their own hardware with no cloud uploads. The 4-bit quantized model fits on 24 GB GPUs, eliminating per-token API costs and keeping proprietary code on-device.

South Portland Schools Consider Temporary AI Ban for Younger Students

The South Portland school board may temporarily ban AI tools for students up to eighth grade for the rest of the school year due to safety and cheating concerns.

Aaronson-Ambainis conjecture becomes a theorem after AI research system finds proof

A decades-old quantum computing conjecture has been proved by an AI system, confirming that quantum computers cannot achieve exponential speedups over classical machines on random, unstructured data.

Google DeepMind finds two weaknesses in legacy 1024-bit RSA encryption

DeepMind's AI found two weaknesses in 1024-bit RSA that cut the cost of forging digital signatures and batch-factoring related moduli. Modern 2048-bit RSA remains secure, but the research signals AI-assisted cryptanalysis is accelerating the obsolescence of older keys.

Global stocks plunge as AI panic and US election uncertainty wipe out $3.2tn

Over $3.2tn was erased from the US stock market as investors fled AI and tech stocks after the midterm elections. Nvidia dropped more than 9% and Japan's Nikkei fell nearly 6% in a rout fueled by fears of Democrat-led regulatory slowdowns.

US consumer sentiment rises in November as worries over AI bubble surface

Consumer sentiment rose to 51.0 in early November, but it fell 5% for those with the largest stock portfolios amid growing anxiety over a possible AI bubble.

Google acquires Crusoe's Wyoming data center site after pre-development work

Google bought a shovel-ready Wyoming data center site from Crusoe, transferring a project with zoning, power agreements, and interconnection studies already in place. The deal can compress pre-construction timelines by 18 to 36 months for hyperscalers racing to secure powered land for AI workloads.

Latest AI News for Government

YMTC targets 70% domestic equipment for fourth chip factory

YMTC plans to source 70% of equipment for its next Wuhan NAND factory from Chinese suppliers, up from just over half at its last plant. The facility will stack more than 300 memory-cell layers and begin construction next year.

Axon invests $1.2 billion in Skydio and becomes its exclusive public safety distributor for the United States and Canada

Axon is investing $1.2 billion in Skydio, raising its stake to about 37% and becoming the exclusive distributor of Skydio drones to U.S. and Canadian public safety agencies.

Meta pauses plan to require Facebook business pages to allow AI training on external content

Meta indefinitely paused a terms update that would have forced Facebook business page owners to allow scraping of external websites for AI training. Courts in Brazil and Norway issued injunctions and the UK's competition authority opened a formal investigation.

Ro Launches AI Healthcare Research Unit for Drugmakers

Ro is launching an AI lab to study patient outcomes and sell the findings to drugmakers, raising concerns about consent and data privacy.

A humanoid robot cleans a San Francisco apartment with remote human help

Tau Robotics charges $60 per weekly session with two humanoid cleaning robots in San Francisco, though the machines still rely on remote human pilots for most tasks. Every counter wiped and toilet scrubbed generates training data the company needs to make its AI fully autonomous.

Latest AI News for Human Resources

Latest AI News for Insurance

Latest AI News for IT and Development

OpenAI partners with Hacktron to find vulnerabilities in its systems

OpenAI tapped security firm Hacktron to hunt vulnerabilities in how its services connect with third-party integrations, granting researchers expanded access to frontier models.

Latest AI News for Management

Latest AI News for Marketing

Latest AI News for Operations

Physicl Raises $10 Million for Physical-AI Synthetic Data

Physicl built a library of 10,000 simulated 3D environments and 100,000 assets for robot training, raised $10 million, and early tests show reduced reconstruction error, but real-world transfer remains unproven.

Latest AI News for Product Development

Latest AI News for Real Estate and Construction

Latest AI News for Sales

Latest AI News for Science and Research

Nvidia open source 6G tools surpass 1 million downloads as researchers build AI-native networks

NVIDIA's open-source wireless research tools passed 1 million downloads as academic adoption grew tenfold to over 8,000 institutions. One project cut robot task times by 52.6%, while others boosted 5G throughput 60% and slashed RAN power use 45.6%.