This is The Stepback, a weekly newsletter breaking down one essential story from the tech world. For more on the data center buildout, follow Emma Roth. The Stepback arrives in our subscribers' inboxes on Sunday at 8AM ET. Opt in for The Stepback here. How it started Years before the AI boom threatened local power […]
Researchers cobbled together funding and time to show how quantum computing could aid in the development of drugs to help underserved populations and combat rare diseases.
ChatGPT is hiring a dedicated product manager to build experiences for families, caregivers, and older adults, according to a job posting.
The world’s most valuable assets are stored on rows of servers in giant, anonymous buildings. And they can be stolen.
Anthropic's distillation complaints expose an awkward question: does AI's fair use argument cut both ways?
Beijing’s cybersecurity warning against Anthropic is expected to speed up a shift already under way.
S&P Global has downgraded Oracle's credit rating to "BBB-," one notch above junk status.
Anthropic analyzed 1.2 million Claude Cowork sessions from more than 600,000 organizations.
OpenAI CEO Sam Altman now says he's "pretty sure" AI has created more jobs than it's eliminated.
Microsoft Copilot now covers everything from free AI chat to Office tools and GPT-5.6-powered work features, so here’s how its many versions fit together.
Musk and Altman's latest public feud escalated July 11 with accusations over space data centers and Grok 4.5 adoption at Tesla/SpaceX. The fight reveals the strategic split between orbital compute and terrestrial buildout for AI infrastructure.
Athenry, Ireland became ground zero in 2015 for a growing global conflict between tech giants and local communities over data centers. As AI's hunger for energy grows, this battle is far from over.
S&P Global's recent downgrade of Oracle's credit rating highlights OpenAI's turning point role in Oracle's financial health. A sudden shift could leave Oracle in a precarious position.
The Wall Street Journal printed an outright false headline and heavily misleading story claiming this, which of course was uncritically amplified by the usual suspects.
The Kyutai spinout is opening a Bay Area office less than a year after forming, with valuation and Nvidia's deal terms undisclosed.
Anthropic's analysis reveals that half of the AI's usage is dedicated to mundane office tasks. The focus is on tasks many avoid, such as report compiling and checklist creation.
Instagram's latest AI feature raises privacy eyebrows, as it dives deeper into users' data. Is it innovation or just invasive?
Sam Altman, CEO of OpenAI, now claims AI boosts employment, marking a significant departure from previous warnings of job loss. What's behind this change?
Christian Cantrell's Palo Alto lab is chasing control, typography, and workflow depth against image models from OpenAI, Meta, Google, and Microsoft.
The rapid expansion of AI servers and large language models (LLMs) is driving unprecedented demand for data center computing, making high-speed interconnect technologies one of the next major battlegrounds in AI infrastructure.
Coherent has agreed to prepay AXT US$22.3 million under a three-year agreement for committed 6-inch indium phosphide substrate capacity, showing how buyers are moving to reserve future supply as AI data center investment strains a concentrated photonics materials market.
In this tutorial, we explore NVIDIA tile-based GPU programming with TileGym, building a Colab workflow that runs across different hardware.
Apple has filed a federal lawsuit against OpenAI, accusing the ChatGPT maker of orchestrating a systematic campaign to steal confidential hardware designs, manufacturing processes, and supplier relationships through more than 400 former Apple employees now working at OpenAI.
OpenAI isn't abandoning its browser ambitions. it's recalibrating the strategy. This shift reflects a broader vision than competing directly in the browser space.
Meta has disabled a feature allowing AI deepfakes of any public Instagram account. This raises questions about privacy and AI's role in social media.
In a significant organizational shift, OpenAI is merging its research and safety departments under a single executive role. This change raises questions about the future prioritization of AI safety.
Phoebe Gates and Sophia Kianni's AI shopping app faces allegations of inflating affiliate sales with fake clicks. The incident raises questions about the ethical oversight in AI-driven commerce.
Apple sued OpenAI on Friday, accusing the AI startup of orchestrating a campaign to steal trade secrets from iPhone maker.
OpenAI's GPT-5.6 Sol Ultra produced a proof of the Cycle Double Cover Conjecture in under an hour, using 64 subagents working in parallel.
Dubbed “Freeze AI on Slushy Day,” the march wound from OpenAI’s Mission Bay HQ to Anthropic and Google, with a detour to jeer at a16z.
Apple's lawsuit against OpenAI over alleged trade secret theft and engineer poaching is heating up Silicon Valley's talent war. With both companies having been close partners before, this legal battle is a masterclass in the risks when partnerships sour.
A Cambridge study found that Boko Haram uses AI chatbots like ChatGPT, Claude, and Gemini to plan attacks, build explosives, and maintain weapons.
OpenAI GPT-5.6 Sol, Terra, and Luna tested: benchmarks, $1 to $30 pricing, 1M context, and where Claude still wins. Full hands-on review.
Apple sues OpenAI, Gemini 3.5 Pro gets a July 17 date, SK Hynix lands the largest ADR debut ever. The 15 biggest AI stories for July 12, 2026.
Apple lawsuit against Sam Altman's OpenAI explained, including trade secret theft, former employees, AI hardware, and supplier claims.
OpenAI's GPT-5.6 Sol Ultra has cracked the Cycle Double Cover Conjecture, unsolved for half a century, in less than an hour. Is this computational prowess a leap in AI creativity or just a high-speed remix of existing knowledge?
The defense-AI company asked current and former employees in May to move from ESOP rights to a plan without direct equity.
This wasn’t a contest so much as a sweep. Kimi-K2.7-Code dominated the matchup on both score and consistency, beating mistral-medium-3-5 across nearly every task type with a decisive 100% confidence verdict.
Johannes Heidecke’s departure comes as OpenAI tries to further integrate its research and safety teams.
Business Insider's Stephen Council used OpenAI Codex to build a custom Mac clipboard app in minutes—no coding required.
OpenAI’s head of safety systems, Johannes Heidecke, is leaving the company following an internal restructuring that merges its safety and research teams under a single leader, Wired reported on Friday.
OpenAI's merger of Codex with ChatGPT offers a new way to manage text snippets on your computer. It's a simple yet powerful tool for anyone juggling words.
OpenAI has responded after being sued by Apple over alleged theft of the iPhone maker's trade secrets.
Anthropic's AI model, Fable 5, is back online after the Trump administration lifted an export control. This heavyweight AI tool, now available again, highlights the tension between innovation and regulation.
OpenAI contemplates a 5% stake to the US government, aiming to align AI benefits with national interest. Sam Altman leads this unexpected strategic play.
NASA's agentic search system leverages large language models to speed up geoscience data discovery, offering a new benchmark for efficient retrieval and analysis.
Apple has sued OpenAI, alleging that former employees that now work for the AI company have stolen Apple's trade secrets "for the benefit of OpenAI." In its complaint, Apple alleges that it has uncovered "a pattern of theft of Apple's trade secrets by OpenAI employees who were formerly at Apple," and it names IO...
Studying for a test, but not sure where to start? Study notebooks, a new feature in the Gemini app, can help you get organized and learn more efficiently.Think of study …
Google’s newly surfaced Magic Pointer listing reveals a Gemini-powered cursor for Googlebooks, but incompatible Android devices mean the feature remains an early preview rather than something consumers can try today.
Apple has sued OpenAI in a California federal court, accusing the ChatGPT maker of using current and former employees to steal hardware designs as it prepares to launch AI-focused consumer devices.
Following the launch of ChatGPT Work and GPT-5.6 Sol, OpenAI has acknowledged significant issues: excessive compute usage, a confusing transition to the desktop interface for chats and projects, an unclear distinction between Codex and ChatGPT Work, and regressions in existing workflows.
Ant Group's Robbyant has released the LingBot-VA 2.0 technical report — a Physical AI video-action foundation model built from scratch for embodiment rather than fine-tuned from a video generator.
The iPhone-maker claims OpenAI encouraged poached employees to bring over confidential presentations, secret prototypes, and key supplier details.
Apple alleges the misconduct was directed by OpenAi's senior leadership, including a long-time former employee.
Apple is suing OpenAI over systematic employee poaching and the alleged theft of trade secrets tied to unreleased products.
Anthropic's clash with the Trump administration over AI model security raises questions about international AI governance and the future of model deployments.
Anthropic has introduced Claude Science, an AI workbench designed to unify scientific tools and data. This move could accelerate discoveries in biotech and pharma.
Apple says OpenAI stole its trade secrets. Here's how Apple has faced similar accusations from rivals in the past.
A large Digital Realty data center in Ashburn, Virginia turned on diesel generators over the July 4 weekend to deal with a strained grid.
Relay supplier Song Chuan Precision reported record June 2026 revenue as customer orders rose and its acquisition of solenoid valve maker Tai Hsing Electronics was folded into its books. The company said the second half should stay supported by demand in new energy and AI power applications.
As the AI boom boosts demand for RAM, SK Hynix - one of the world's biggest suppliers of memory chips - launched on Wall Street Friday.
The AI chip boom just produced its biggest Wall Street moment yet. Now SK Hynix and Samsung are being asked to build US factories.
Agentic AI is putting new pressure on enterprise networks and cybersecurity infrastructure as companies accelerate internal deployment, with Cisco and Foundry warning that more than 70% of enterprises see their current networks as unprepared for future AI demand.
AI models like GPT-4o show potential in identifying antisemitic events but face significant challenges. Clear definitions and examples can aid their performance.
Apple Inc. today sued OpenAI Group PBC for allegedly stealing intellectual property related to its consumer devices.
TRIAGE refines agentic reinforcement learning by introducing role-typed credit, enhancing success rates and reducing inefficiencies.
Apple's decision to sue OpenAI for trade-secret theft does more than escalate a soured partnership between two of technology's most influential companies.
Plus: The Pentagon is training amateurs to become part of its hacker army, a Flock license plate reader error led to cops surrounding a car reviewer, and more.
The lawsuit targets OpenAI's Jony Ive-led device push and names io Products, Tang Tan and Chang Liu as defendants.
Apple accused OpenAI of stealing trade secrets in a lawsuit filed Friday. OpenAI's recruiting techniques were put on blast.
RuntimeWire testing found an undocumented mini-game inside xAI's terminal coding agent
One model showed up as a complete video generator; the other mostly showed flashes of promise without converting them into task wins. Across four prompts, Gemini Omni Flash separated itself on prompt fidelity and scene construction, and the margin wasn’t subtle.
OpenAI has issued a formal statement in response to Apple’s lawsuit accusing the company of trade secret theft. Read it below.
Apple said in a Friday lawsuit that OpenAI’s nascent hardware business is “rotten to its core.”
According to OpenAI, GPT-5.6 Sol independently fine-tuned the smaller Luna model, triggered by a single "fairly under-specified prompt." In OpenAI's internal RSI benchmark for recursive self-improvement, Sol scores 16.2 points higher than GPT-5.5.
In a filing, Apple says former employees who left for OpenAI had access to its systems after leaving.
Large language model (LLM) training workloads increasingly run into GPU memory limits before compute is fully used. Model weights, gradients, optimizer states,...
Dynamic Unsloth NVFP4 Quants. Faster and more accurate.
How Deutsche Telekom is becoming an AI-native telco with OpenAI-transforming customer service, employee workflows, network operations, and the future of voice.
A writer left LA for San Francisco to work at the AI startup, Corgi, after she received a cold message from its chief of staff.