Technology News
Daily Tech Reader - Technology Edition
AI Software & Platforms
AI Software & Platforms
OpenAI Launches GPT-6 Sol and Luna at Half the Cost of 5.6 — New Models Target Coding and High-Volume Clerical Work
OpenAI launched GPT-6 Sol and Luna on Tuesday, priced at half the GPT-5.6 tier — $2/$10 per million tokens for Sol versus $4/$20 previously — with Sol positioned for complex coding work and Luna aimed at high-volume clerical tasks; OpenAI says Sol makes roughly half as many factual errors as its predecessor and matches flagship Astra on reliability, with the drop landing approximately 90 minutes after Anthropic released Claude Opus 5.5.
Sources: OpenAI · TechCrunch · AI Weekly
Anthropic Ships Claude Opus 5.5 at 40% Lower Cost — Faster Output, Stronger Safety Scores, Available Across Major Cloud Platforms
Anthropic launched Claude Opus 5.5 on September 22 at $4/M input and $20/M output tokens — roughly 40% cheaper than Opus 5 on typical workloads — reporting 66.4% on Terminal-Bench 4.0 versus Opus 5's 52.3%, 81.8% on OSWorld 2.0 computer use, and output 30% faster; the release is the first since Dario Amodei's "pace the frontier" essay and ships across AWS, Google Cloud, Azure, and the Claude Platform.
Sources: Anthropic · AI Weekly · TradingKey
OpenAI Fires Contractors for Using AI to Train Its AI — Project Lily Workers Caught Violating Human-Only Review Rules
404 Media reports OpenAI has terminated multiple contractors under Project Lily — its human-feedback program for rating ChatGPT responses — after catching them using AI tools including GPTZero, Grammarly, and other LLMs; termination letters cited "authenticity" failures, and reviewers flag AI use by watching for repetitive language, excessive em dashes, and unusually fast completion rates.
Sources: 404 Media · Tom's Guide · AI Weekly
AI Devices & Hardware
AMD Crosses $1 Trillion Market Cap for the First Time — Five-Day AI Chip Rally Puts It in the Club with Nvidia, Broadcom, and Micron
Advanced Micro Devices crossed $1 trillion in market capitalization on Monday as shares hit a record intraday high of $613.92, up roughly 9.6% on the session, making AMD the fourth U.S. chipmaker to reach the milestone; the five-session rally has added approximately 24% to the stock and reflects surging demand for AMD's full AI server systems that bundle processors, networking gear, and accelerators.
Sources: Reuters · CNBC · Bloomberg · Qz.com
Alibaba Unveils Zhenwu V900 — Claims China's Most Powerful AI Chip with 3x Performance Gain Over Previous Generation
At its annual Hangzhou Apsara Conference, Alibaba CEO Eddie Wu unveiled the Zhenwu V900 accelerator, calling it the most powerful AI chip in China today and citing triple the performance of the prior Zhenwu M890; Alibaba also announced plans to train a next-generation model at five to ten trillion parameters and targets 20 gigawatts of computing capacity by 2032, framing the push as mobilizing "every resource" to meet AI demand.
Sources: Associated Press · Bloomberg · BNN Bloomberg
Qualcomm Kicks Off Snapdragon Summit 2026 — Snapdragon 8 Elite Gen 6 on TSMC 2nm Leads Agentic AI PC Push
Qualcomm opened its three-day Snapdragon Summit on Tuesday with the announcement of the Snapdragon 8 Elite Gen 6 and 8 Elite Extreme Gen 6, both built on TSMC's 2nm process with two 5GHz prime cores and six 4GHz performance cores; CEO Cristiano Amon is using the event to detail Snapdragon's strategy in the era of agentic AI, with sessions covering mobile AI, personal AI, and AI PC workloads.
Sources: TechRadar · TradingKey · AI Weekly
Enterprise & Cloud
Snorkel AI Raises $350M at $3.5B Valuation — ARR Up 17x in a Year After Pivot to Finished Datasets and RL Environments
Snorkel AI closed a $350 million round led by Insight Partners and S32 with Addition, Greylock, and Wells Fargo, valuing the company at $3.5 billion — nearly triple its $1.3 billion May 2025 mark; CEO Alex Ratner said annualized revenue crossed $350 million, up from roughly $20 million a year earlier, driven by a shift from selling software to supplying finished training datasets and reinforcement learning environments to frontier labs, hyperscalers, and federal customers.
Sources: The Star · AI Weekly
Alphabet's Intrinsic Open-Sources Full Robotics Stack Under Apache 2.0 at ROSCon 2026 in Toronto
Alphabet's Intrinsic released Intrinsic Core, a ROS-compatible robotics environment, at ROSCon 2026, bundling hardware-agnostic real-time control, Nvidia FoundationPose-based pose estimation, motion and grasp planning, simulation and calibration services, and Intrinsic-ROS drivers supporting Universal Robots and FANUC arms under the Open Machine Tending reference design.
Sources: SiliconAngle · AI Weekly
Cybersecurity
Cisco Talos Discloses CLOSEDQUORUM — First Reported Fully Autonomous Multi-Model AI Command-and-Control Implant
Cisco Talos released CAIRN, a metadata-first toolkit for hunting AI-integrated malware by extracting prompt templates, API keys, and jailbreak strings without executing binaries, and simultaneously disclosed CLOSEDQUORUM — what it calls the first reported fully autonomous multi-model AI C2 implant that operates with no human operator; Talos says AI-integrated malware evolved from optional LLM helpers to autonomous C2 orchestration in a single year.
Sources: Cisco Talos · AI Weekly
ShinyHunters Hijacks Clop's Dark Web Leak Site — Extortion Demand Rises Every 24 Hours as Gang War Goes Public
The ShinyHunters group defaced the Clop ransomware gang's dark web leak site with an eight-figure extortion demand that escalates every 24 hours, threatening to expose which companies paid Clop ransoms, the amounts paid, and the Bitcoin addresses used; ShinyHunters says the feud began after Clop exploited an Oracle E-Business Suite vulnerability without permission and threatened one of its members.
Sources: The Record · CISO Series
Regulation & Policy
British Columbia Sues OpenAI Over Tumbler Ridge School Shooting — Province Says ChatGPT Safety Team Flagged Shooter, Leadership Overruled
British Columbia Attorney General Niki Sharma filed suit in San Francisco federal court against OpenAI and Sam Altman, seeking damages tied to the February 10 Tumbler Ridge shooting that killed eight; the province alleges OpenAI's safety team flagged shooter Jesse Van Rootselaar's ChatGPT conversations about gun violence and recommended police contact, but leadership overruled the recommendation, with OpenAI subsequently refusing to share the shooter's chat history with provincial investigators.
Sources: Al Jazeera · CTV News · iPhone in Canada
OpenAI and Anthropic to Brief UN Security Council Wednesday — First Time Frontier US and Chinese AI Labs Face Council Together
France, holding the September Security Council presidency, convened a 15-member session on AI and international security for Wednesday, September 23; OpenAI CEO Sam Altman and senior Anthropic representatives are expected alongside DeepSeek and Moonshot AI, marking the first time the Council has directly hosted frontier Chinese and U.S. AI developers together on shared safety concerns, as Xi Jinping arrives in Washington for a state visit with President Trump.
Sources: TechStartups · AI Weekly
Bessent Emerges as Frontrunner for Trump's AI Czar Role — White House OSTP Director and OPM Director Also in Contention
Semafor reports Treasury Secretary Scott Bessent has emerged as the leading candidate for the newly created White House AI czar position, per three sources, having already led AI policy work including recent talks with Chinese officials on an AI national security incident notification mechanism; White House OSTP Director Michael Kratsios and OPM Director Scott Kupor are also under consideration.
Sources: Semafor · AI Weekly
Media & Entertainment
Microsoft Cuts 268 at Halo Studios, Hands Franchise to Activision — Broader Xbox Overhaul Cuts Total Fewer Than 600
Microsoft announced 268 job cuts across Halo Studios and Xbox Game Studios management, with Xbox EVP Matt Booty saying the reset is now "three-quarters" complete; Activision will take over development of the next Halo title with a separate team distinct from Call of Duty, while a small Halo Studios team continues supporting released games, with broader Microsoft cuts across cloud, AI, marketing, and research totaling fewer than 600 positions.
Sources: GeekWire · AI Weekly
On the Horizon
Xi Jinping Arrives in Washington Wednesday — AI, Trade, and Tariffs Top the Agenda at Trump State Visit
Chinese President Xi Jinping is set to arrive in Washington on Wednesday for a state visit with President Trump, with AI competition, trade policy, and tariffs expected to headline discussions; the visit comes as Alibaba unveiled its most powerful AI chip and U.S. AI leaders including Anthropic CEO Dario Amodei have been publicly warning that China's AI development poses a strategic threat to the United States.
Sources: Associated Press · Bloomberg
Qualcomm Snapdragon Summit Runs Through September 24 — Agentic AI PC Strategy and CEO Keynote Still to Come
Qualcomm's Snapdragon Summit 2026 continues through September 24, with upcoming sessions covering personal AI, agentic PC workloads, and a unified AI computing software layer following Qualcomm's acquisition of Modular; investors are watching for detail on how Snapdragon can expand beyond smartphones and grow its revenue contribution from non-mobile markets.
Sources: TradingKey · TechRadar