October 07, 2026

Decision models are suddenly everywhere. Open AI’s is now public.

This week, OpenAI opened its Decisions API to public beta and added image inputs to make fast decisions from visual The post Decision models are suddenly everywhere. OpenAI’s is now public. appeared first on The New Stack.

Open AI publishes 722 AI-generated math discoveries in major scientific milestone

OpenAI Group PBC has published 722 math papers that it generated using an unreleased artificial intelligence model. Some of the papers, which were posted to GitHub late Tuesday, prove long-running hypotheses. Others disapprove proposed explanations of mathematical phenomena or narrow down the possible answers to complex, yet-unsolved puzzles. Overall, OpenAI’s papers span about 20 mathematical […] The post OpenAI publishes 722 AI-generated math discoveries in major scientific milestone appeared first on SiliconANGLE.

Microsoft releases new Nvidia-chip AI PCs with revamped Windows 11

Microsoft revealed the specs and price for its Surface Laptop Ultra, AI PCs that run on Nvidia chips that are designed to run AI models and agents.

Open AI says teen Chat GPT use limited but research finds it an 'unacceptable risk' - BBC

OpenAI says teen ChatGPT use limited but research finds it an 'unacceptable risk' BBC

Microsoft Turns Windows PCs Into Command Centers for AI Agents - PYMNTS.com

Microsoft Turns Windows PCs Into Command Centers for AI Agents PYMNTS.com

Study Shows AI Chatbots Offer the Rich Higher Price Recommendations - Bloomberg.com

Study Shows AI Chatbots Offer the Rich Higher Price Recommendations Bloomberg.com

Claude Haiku 5.5 Scores 43 On Artificial Analysis Index, Just Behind Kimi K3 While Ahead Of GPT-6 Luna

Artificial Analysis has put Anthropic’s newly released Claude Haiku 5.5 through its Intelligence Index, and the small model lands at 43 at max... The post Claude Haiku 5.5 Scores 43 On Artificial Analysis Index, Just Behind Kimi K3 While Ahead Of GPT-6 Luna appeared first on OfficeChai.

Google Opens Its Synth ID Watermark Detector To The Public - Search Engine Journal

Google Opens Its SynthID Watermark Detector To The Public Search Engine Journal

Anthropic Launches Claude 5.5 Haiku, Beats GPT-6 Luna Across Most Benchmarks While Priced at 1/20th Of Sonnet 5.5

Anthropic has launched Claude Haiku 5.5, the small-model entry in its 5.5 family, and is pitching it as the cheapest, fastest and most... The post Anthropic Launches Claude 5.5 Haiku, Beats GPT-6 Luna Across Most Benchmarks While Priced at 1/20th Of Sonnet 5.5 appeared first on OfficeChai.

Chat GPT with GPT-6 ditches mostly text output for interactive UI with charts, buttons, and mini apps

OpenAI is rolling out GPT-6 with "Intelligent UI," a feature that turns answers into interactive interfaces with charts, buttons, and forms. The model can now respond while still thinking, cutting wait times by 44 percent. Paying customers get GPT-6 Sol, free users get GPT-6 Luna. The article ChatGPT with GPT-6 ditches mostly text output for interactive UI with charts, buttons, and mini apps appeared first on The Decoder.

Introducing Claude Haiku 5.5 on AWS

Claude Haiku 5.5 is now available on Amazon Bedrock and Claude Platform on AWS. According to Anthropic, it is the fastest, most efficient model in the Claude 5.5 family, built for subagents and high-volume, cost-sensitive work, and costs around 75% less than Claude Haiku 4.5 for most tasks. This post covers its improvements and how to get started.

Claude Haiku 5.5 arrives with massive price cuts proving the AI pricing arms race is far from over

Anthropic's new Claude Haiku 5.5 crushes its predecessor in benchmarks, jumping from 15.7 to 72.4 percent on the OSWorld computer use test. Token prices drop by up to 90 percent, though a new tokenizer eats into some of those savings by consuming more tokens per task. The article Claude Haiku 5.5 arrives with massive price cuts proving the AI pricing arms race is far from over appeared first on The Decoder.

These Researchers Made AI Drive a Toyota Corolla to Get In-N-Out

Three engineers put GPT, Claude, and Grok in charge of a real car. Only one of them was successful.

Microsoft Says New Laptop With Nvidia Chip Outperforms Mac Books

Microsoft Corp. revealed pricing and availability for a new flagship laptop built on Nvidia Corp. chips, saying that the machine outperforms a comparable Apple Inc. model on some AI tasks.

Rethinking access control for RAG with Amazon Quick and Amazon Bedrock

Enterprise RAG unlocks insights from knowledge sources like SharePoint, Google Drive, and Confluence, but those sources carry complex permissions. Learn how Amazon Quick and Amazon Bedrock Knowledge Bases enforce document-level access controls in real time, verifying permissions directly with authoritative sources at query time.

Meta’s Muse launches on i Pad just a month after its mobile debut

Meta’s AI agent Muse is now available on iPad, just a month after its mobile debut, as the company rapidly expands the assistant’s reach and integrations.

Microsoft to sell $2,599 Surface Laptop Ultra containing Nvidia AI chip - CNBC

Microsoft to sell $2,599 Surface Laptop Ultra containing Nvidia AI chip CNBC

The Machines that Make the Machines

How we taught robots to assemble GB300 tester trays and what it taught us about robot learning, mechanical intelligence, and good old-fashioned engineering The...

Google got you found. Now real estate agents need to win over AI - Housing Wire

Google got you found. Now real estate agents need to win over AI HousingWire

Open AI’s former safety lead says it’s shipping ‘new capability and risk every Tuesday’

When David Robinson joined OpenAI in May 2023, the day after Sam Altman first testified before the Senate, the company’s The post OpenAI’s former safety lead says it’s shipping ‘new capability and risk every Tuesday’ appeared first on The New Stack.

Sequoia, Nvidia Back Mecka AI’s Robotics Push

Mecka AI has raised $60 million in Series B funding led by Sequoia Capital, with backing from Nvidia, Microsoft’s M12 and Qualcomm, as it builds technology to help robots understand and operate in the physical world. CEO and cofounder Josh Gao discusses the enormous data challenge facing robotics, how Mecka captures everything from physical interactions to force and pressure, and why scaling physical AI will require major investments across sensors, infrastructure and real-world data collection. He joins Ed Ludlow on "Bloomberg Tech." (Source: Bloomberg)

Rising Oil Puts Wall Street on Edge

Get a jump start on the US trading day with Dani Burger on "Bloomberg Open Interest." SpaceX looks to raise $40 billion for Nvidia chips, Apple makes a bigger smart-home push with LG, and rising oil and rates cool Wall Street’s record rally. Plus, Apollo’s John Cortese on how the company is pushing to make private credit more transparent. New York Comptroller Thomas DiNapoli joins us on record banker bonuses and how the AI boom is impacting New York's bottom line and Alvarez & Marsal’s Paul Aversano on private credit, record profits and the M&A slowdown. Correction: "This video was edited to correct the spelling of "Paul Aversano" that aired in the program at 10:10 am ET." (Source: Bloomberg)

Bof A Sees Strong Demand for Open AI, Anthropic IPOs

Bank of America’s JD Moriarty says public markets have the capacity to absorb massive potential IPOs from companies including OpenAI and Anthropic, pointing to SpaceX’s listing as evidence of strong institutional and retail demand for transformational technology companies. He discusses how abundant private capital has changed the traditional IPO playbook and what the pipeline could mean for tech listings next year. He joins Ed Ludlow on "Bloomberg Tech." (Source: Bloomberg)

Chat GPT is getting a lot more visual, with the launch of a new interface

OpenAI is launching a new user interface that will bring interactive visuals to ChatGPT.

Anthropic launches Haiku 5.5 at a much lower price

Anthropic on Wednesday launched Claude Haiku 5.5, the first new version of its smallest and most affordable model in nearly The post Anthropic launches Haiku 5.5 at a much lower price appeared first on The New Stack.

The New Chat GPT Is More Show Than Tell

OpenAI is updating ChatGPT for all users with an “Intelligent UI” that’s more visual—generating interactive elements as part of the chatbot’s outputs.

Zuckerberg's Biohub leads a $1.8 billion push to build AI models that predict cell behavior

Biohub, the research organization backed by Mark Zuckerberg and Priscilla Chan, is coordinating a $1.8 billion initiative to train AI models that predict cell behavior. Meta, Google DeepMind, Isomorphic Labs, and the US Department of Energy are funding data, lab equipment, and compute. A first dataset should be ready in about a year. The article Zuckerberg's Biohub leads a $1.8 billion push to build AI models that predict cell behavior appeared first on The Decoder.

Google says 180 billion images and videos now carry Synth ID watermarks as detector goes public

Google's AI watermark detector SynthID is now public. Anyone can check whether images, videos, or audio were created by Google's AI or partners like OpenAI and Nvidia. More than 180 billion pieces of content already carry the watermark, but the tool only identifies SynthID-tagged content, not AI-generated media in general. The article Google says 180 billion images and videos now carry SynthID watermarks as detector goes public appeared first on The Decoder.

Space X Reportedly Wants $40 Billion to Buy Artificial Intelligence (AI) Chips. Here's Why That's Huge News for Nvidia Investors. - Yahoo Finance

SpaceX Reportedly Wants $40 Billion to Buy Artificial Intelligence (AI) Chips. Here's Why That's Huge News for Nvidia Investors. Yahoo Finance

Space X Seeks $40 B for Nvidia Chips as AI Megadeals Loom

Bloomberg’s Ed Ludlow takes a look at SpaceX in talks with banks and investors to borrow $40 billion to buy chips from Nvidia. Plus, Apple plans its entry into the smart home space with a doorbell, thermostat, and other accessories developed with LG; and JD Moriarty, Vice Chair and head of Global Tech Equity Capital Markets at Bank of America Securities, discusses the upcoming mega AI IPO pipeline. (Source: Bloomberg)

Brief remark on Open AI’s giant math drop - Marcus on AI | Substack

Brief remark on OpenAI’s giant math drop Marcus on AI | Substack

Multimodal open d1 decision models for the edge

Meta rolls out new AI tools to detect ads that secretly lead to child sexual abuse material

Meta launches new AI tools after discovering ads on its platforms that may look normal but direct users to harmful content elsewhere online.

Microsoft or Amazon: If I Could Only Own One for the Next 5 Years, It Would Be This One - 24/7 Wall St.

Microsoft or Amazon: If I Could Only Own One for the Next 5 Years, It Would Be This One 24/7 Wall St.

Now GA: Building permission-aware Databricks Apps with on-behalf-of-user authorization

Databricks Apps lets developers build and deploy data and AI applications directly on the Databricks platform...

Introducing Google Cloud’s U4 compute: Enabling ultra-low latency trading

In capital markets, operational success is built on predictable performance and low latency, where even tiny variations in packet timing alter queue priority and trade execution. At the same time, both exchange venues and their participants face mounting constraints in traditional on-premises data centers — from power and physical rack space limits to lengthy hardware procurement cycles. Increasingly, global capital markets seek the speed and determinism of physical co-location combined with the dynamic scalability and automation of the cloud. At Google Cloud Next ‘26, we announced the Ultra Low Latency (ULL) Solution, now generally available, providing high-frequency trading workflows that run in the cloud with an ultra-low latency network and agility. The ULL Solution includes the new U4 machine family, also generally available today. The ULL Solution is based on three infrastructure pillars: Hardware-level networking: Scaleable, hardware-based multicast data distribution for reliable market-data feeds Advanced networking observability: Dynamic network traffic capture with hardware-level timing accuracy, facilitating consistent auditing, market replays, and absolute trade validation without impacting primary traffic performance Deterministic high-performance compute: Processing of latency-critical execution tiers in a highly predictable, consistent amount of time, every single time Taken together, the ULL Solution’s compute, storage, networking, and observability provide financial exchanges and market participants with a number of technical capabilities: Bare metal performance: Dedicated bare metal compute provides direct access to physical host resources, minimizing jitter and delivering predictable, low-latency execution for market-data feeds and order routing. High-performance storage options: Local Titanium SSDs handle real-time transaction and tick logging on the host, complemented by scalable Google Cloud Hyperdisk for persistent market data archives and analytics. Physical traffic isolation: The ULL trading fabric’s redundant A/B multicast market feeds are accessed through two independent and dedicated Titanium adapters, while an independent third Titanium adapter offloads telemetry, management, and provides access to Google Cloud services. Hardware-accelerated multicast distribution: The ULL network architecture supports hardware-level multicast feed ingestion. This allows participants to stream high-throughput market data directly to low-latency trading applications, bypassing traditional hypervisor-level virtual switches. Accelerated packet processing: Support for OpenOnload and DPDK enables Linux user-space networking to deliver predictable unicast and multicast packet handling while minimizing application code changes. Precision timing and UTC synchronization: Integration with Google Cloud’s Firefly clock synchronization system allows the solution to consistently achieve sub-10 nanosecond network-level timestamping and better synchronization to UTC than the sub-100 microsecond regulatory requirement for financial exchanges. Built-in telemetry: 24/7 low-latency packet capture and seamless out-of-band packet brokering for regulatory compliance and real-time analytics helps ensure deep visibility without impacting primary traffic performance. 24-7 market ready: The solution is designed for continuous, round-the-clock trading readiness by isolating production workloads in a dedicated primary zone for live trading, while routine cloud maintenance and qualification testing occur in a secondary zone for updates and testing. Compute in the ULL Solution is delivered by the new U4 machine family, which brings predictable performance and ultra-low latency compute in three specialized machine series: dual-socket bare metal instances with three physical NICs — U4P for exchange operators and U4C for market participants — alongside U4S high-performance VMs for operators, participants, and service providers. Machine Series CPU Platform CPU Cores Memory NICs Storage Options U4P (Bare Metal) Intel 5th Gen Xeon (Emerald Rapids) 120 Physical Cores 512 GB or 768 GB 3 Physical (1 Standard + 2 ULL) Local Titanium SSD (12 TiB), Hyperdisk U4C (Bare Metal) Intel 5th Gen Xeon (Emerald Rapids) 120 Physical Cores 512 GB or 768 GB 3 Physical (1 Standard + 2 ULL) Local Titanium SSD (12 TiB), Hyperdisk U4S (VM) Intel 6th Gen Xeon (Granite Rapids) 2 to 288 vCPUs Up to 2,232 GB Multi-vNIC (Up to 200 Gbps) Local Titanium SSD (18 TiB), Hyperdisk We developed the U4 machine family to enable the world’s most technically demanding markets to run within Google Cloud and benefit from cloud services and scale. An example of this is our ongoing collaboration with CME Group, through which we are migrating listed derivatives markets to Google Cloud. Here, the U4C and U4P bare metal instances deliver direct physical co-location latency parity, providing predictable low-latency clock precision, and native hardware-multicast feed ingestion. This architecture demonstrates that core exchange systems and trading strategies can run in the cloud with the speed, consistency, and control that financial markets require. "Migrating the world’s leading derivatives marketplace to the cloud requires uncompromising performance, latency, and reliability. Google Cloud’s Ultra Low Latency Solution and the U4 instance families represent a breakthrough for financial market infrastructure. By combining deterministic, ultra-low latency networking and dedicated bare metal compute with the benefits of the cloud, market participants can execute and ingest market data and execute with the speed of traditional co-location and the agility of the cloud." - Pearce Peck-Walden, Managing Director Markets Engineering, CME Group Powered by Google Cloud-native infrastructure The ULL Solution is available in select private Google Cloud regions, where trading teams can leverage core cloud infrastructure benefits such as rapid automated resource provisioning, on-demand capacity scaling, and dynamic fleet management. Trading teams can execute ultra-low latency trades on a dedicated, isolated ULL network, and offload data streams over an independent network interface to Google Cloud services such as BigQuery and Gemini Enterprise. For exchange participants managing tick-to-trade workflows or processing real-time market data, microsecond variations are critical. As bare metal instances, U4P and U4C machine series don’t add any virtualization overhead for latency-sensitive execution tiers. Leveraging Google Cloud's Titanium offload system architecture, these bare metal instances provide direct access to the host resources. Compute Engine instances enable users to provision dedicated hardware through standard Google Cloud APIs, orchestrate deployments with infrastructure as code, apply Cloud Next Generation Firewall (NGFW) policies, and collect rich telemetry. The U4S VM series complements these bare metal offerings for applications such as pre-trade risk validation and real-time analytics. Deployed alongside U4P and U4C instances, these VMs shorten transit hops across the trading architecture while offering elastic scaling from 2 to 288 vCPUs and networking up to 100 Gbps. “Our testing of the U4 instances validated that Google Cloud delivers the low latency and determinism required for our demanding exchange trading workloads. Having both bare metal and VM options allows us to evaluate the value of that performance relative to its cost, helping us align the right infrastructure to each of our trading strategies.” - Corbin Kidd, EVP/CTO, Prime Trading "In our tick-to-trade performance testing, the U4C bare metal instance demonstrated exceptional speed, showing that cloud compute and ultra-low latency networking can comfortably satisfy microsecond trading demands. But, what sets the U4 series apart isn't just raw processing speed, it’s structural determinism. When we stress-tested Google Cloud's U4 bare metal family, the performance curve remained virtually flat, which confirms that these cloud compute layouts can now offer the strict determinism institutional market participants require." - Christopher Wilson, Global Head of Experience Modernization, 28Stone Looking to the future, this low-latency architecture serves as the foundation for algorithmic trading, allowing firms to feed real-time markets directly into AI-driven trading models. Storage and placement for trading workloads The U4 family supports Google Cloud storage options and placement policies, helping trading teams record trades at speed and minimize network latency between machines: Local Titanium SSD for fast logging: For real-time trade logging and data caching without slowing down execution, U4P and U4C provide up to 12 TiB of direct-attached NVMe SSDs, and U4S VMs scale up to 18 TiB. Hyperdisk for data archives: For historical market data, backtesting models, and regulatory archives, you can attach Hyperdisk Balanced and Hyperdisk Extreme volumes across all U4 series, scaling up to 32 volumes and 512 TiB per instance. Placement policies: To reduce physical network hops, compact placement policies keep related machines close together in the data center. Alternatively, spread placement policies place machines on separate racks to improve resiliency. Getting started with the ULL Solution With the ULL Solution, financial exchange operators, exchange participants, and trading service providers can execute trades with the speed and reliability they need today while positioning their businesses to leverage continuously evolving AI and infrastructure capabilities for the trading of tomorrow. The ULL Solution is available in select Google Cloud regions. To learn more, check out the public documentation, and reach out to our team to request access.

Validate AI Factory Changes with Digital Twins and AI Agents

AI factories are some of the most complex operations in the world, combining GPUs, CPUs, switches, DPUs, and SuperNICs alongside schedulers, orchestration...

Scaling Decision Optimization to 100 Million Variables and Beyond with m PDLP in NVIDIA cu Opt

Supply chain problems are expanding across more SKUs, lanes, and constraints than ever before, while energy grids are balancing more distributed sources in real...

Open AI’s release of mathematical findings draws concerns from experts - The Guardian

OpenAI’s release of mathematical findings draws concerns from experts The Guardian

Open AI launches Decisions API that reduces complex evaluations to yes, no, or pick one

OpenAI's new Decisions API classifies text and images about ten times faster than the Responses API, returning yes/no probabilities, category picks, or scale ratings for $0.10 per million input tokens. The company also cut its paid API tiers from five to three. The article OpenAI launches Decisions API that reduces complex evaluations to yes, no, or pick one appeared first on The Decoder.

The Pentagon Hopes to Speed Up ‘Kill Chain’ AI Buys With 5-Minute Videos

The US government’s Tradewinds initiative has made it easier to throw millions of dollars at “nontraditional” defense contractors, including OpenAI, Anthropic, and Google.

Opinion: Can carriers finally move up the stack? AT&T, Cisco and NVIDIA say AI changes the equation - Fierce Network

Opinion: Can carriers finally move up the stack? AT&T, Cisco and NVIDIA say AI changes the equation Fierce Network

Faster Scientific Image Analysis with NVIDIA cu Photon

Observatories and telescopes, lasers and X-ray light sources, and other high-throughput instruments generate image data faster than CPU-bound pipelines can...

Space X vs. Nvidia: Which Is the Better Trillion-Dollar Artificial Intelligence (AI) Stock for 2027? - The Globe and Mail

SpaceX vs. Nvidia: Which Is the Better Trillion-Dollar Artificial Intelligence (AI) Stock for 2027? The Globe and Mail

Open AI Used Just 3 Hours Of Chat GPT Pro Thinking Per Result For Its 300+ Math Proofs, Company Reveals

The average result in the 722-manuscript release took roughly the compute of three hours of ChatGPT Pro thinking, a small fraction of what... The post OpenAI Used Just 3 Hours Of ChatGPT Pro Thinking Per Result For Its 300+ Math Proofs, Company Reveals appeared first on OfficeChai.

AI Agent Adoption Soars on Consumer, Search Trends

Meta's Muse has topped 6.6 million downloads since its Sept. 8 launch and reached 1.8 million daily active users, as consumers show they're ready to hand over their personal information for a chance at some AI assistance. Seema Shah, VP: Insights at Sensor Tower, discusses the growth of personal AI agents, Zuckerberg's Muse, and the broader AI market. (Source: Bloomberg)

Google bets Gemini can turn casual players into game developers with new Playground feature

Google launched Playground, a browser-based platform that lets users create their own games using nothing but text prompts and no coding skills. The platform runs on Gemini, Nano Banana, and Lyria. Alongside Playground, Google and Unity announced Unity Spark, a more professional AI tool with a beta planned for 2026. The article Google bets Gemini can turn casual players into game developers with new Playground feature appeared first on The Decoder.

Google experiments with an AI-powered gaming platform

Google Labs is working on a new AI-powered game-creation platform called Playground for users to build browser-based games using simple text prompts.

Space X Reported to Be in $40 Billion Nvidia Chip-Financing Talks

Elon Musk's SpaceX is in talks with banks and investors to raise $40 billion to buy chips from Nvidia. Sarah Frier has more on "Bloomberg Open Interest." (Source: Bloomberg)

Open AI’s Alexander Embiricos is coming to Tech Crunch Disrupt 2026 — days after the launch of Dots

OpenAI’s Alexander Embiricos is coming to the AI Stage at TechCrunch Disrupt 2026, just days after the launch of Dots. Join this conversation by registering for your pass. Get you pass now to save up to $100 and get a second at 50% off.