news

AI News Today Big Tech

The race to lead in artificial intelligence is being run largely by the biggest names in tech. Here you will find stories about AI models, products, partnerships, and strategy from companies such as Google, Microsoft, Meta, Amazon, and Apple. The section also covers the chips, data centers, and talent wars behind the headlines.

Expect news on frontier model launches, AI assistants being woven into operating systems and office software, and the enormous capital expenditures going into GPUs and data centers. Partnership and licensing deals between platform companies and AI labs are a recurring theme, as are high-profile researcher moves between rivals.

The stakes are unusually high because AI leadership could decide which of today's giants dominates the next computing era. Investors watch these companies' AI spending for market signals, developers build on their platforms, and policymakers study their moves when weighing how the technology should be governed.

China’s open-weights AI strategy is winning
2026-07-20

China is currently leading the artificial intelligence sector by embracing an open-weights strategy for its models. Meanwhile, the United States' closed-first approach is facing criticism as a failing strategy that could ultimately drag down the American economy.

Hacker wipes Romania's land registry database
2026-07-20

A hacker has successfully wiped Romania's national land registry database. In other cybersecurity developments, Graykey is suing a former employee for leaking an exploit, and Hugging Face was compromised using AI. Furthermore, researchers have finally discovered an unauthenticated remote code execution vulnerability in WordPress.

Kimi Work
2026-07-20

Kimi Work is a desktop AI application designed to automate workflows, organize files, and operate across the web. By utilizing 300 specialized agents, the platform aims to enhance productivity for complex duties such as financial research, data analysis, and general office tasks.

Nativ: Run frontier open models locally on your Mac
2026-07-20

Nativ is a free and open-source application that allows Mac users to run frontier open AI models locally on Apple Silicon. By supporting language, vision, video, code, and embedding models directly on the hardware, the tool enables users to leverage advanced AI capabilities while ensuring complete data privacy.

Claude Fable produced a counterexample to the Jacobian Conjecture
2026-07-20

An AI model named Claude Fable has purportedly generated a counterexample to the long-standing Jacobian Conjecture. The proposed solution introduces a specific polynomial mapping in complex three-dimensional space that maintains a constant Jacobian determinant of -2. By demonstrating that this mapping sends three distinct points to the exact same coordinate, the model claims to have proven the famous mathematical conjecture false.

Airport Simulator
2026-07-20

The newly announced Airport Simulator challenges players to manage complex aviation operations in a deeply immersive environment. The developers promise that both the sky and player endurance are the ultimate limit for this experience. Fans of the management genre can expect a rigorous and engaging test of their strategic stamina.

Agent swarms and the new model economics
2026-07-20

Recent research comparing AI agent swarms tasked with building SQLite from scratch highlights a significant shift in model economics. The study found that newer, better-coordinated swarms deliver code quality comparable to older versions while completing the complex task at a fraction of the cost.

Shinjuku Station in 3D
2026-07-20

A new interactive 3D map visualizing Tokyo's notoriously complex Shinjuku Station and its surrounding underground network has sparked interest among readers. While some users point out that the visualization is incomplete due to limited government source data, others clarify that its exaggerated vertical slopes are simply schematic representations of the station's steep escalators. Many commenters praise the project, suggesting it could be adapted into a first-person navigation game to help visitors traverse the labyrinthine transit hub.

I found a WordPress RCEs with GPT5.6 and $25
2026-07-20

Using the GPT5.6 AI model and $25, a security researcher uncovered a Remote Code Execution vulnerability in WordPress. The disclosure was temporarily delayed to give administrators time to patch their systems over the weekend. A free online tool is also available to help users check if their instances are affected by this flaw.

How we measured AI writing across arXiv, and where the measurement breaks
2026-07-20

Researchers analyzed the full text of 12,750 arXiv papers and found that about a third of new publications read as machine-written. The study outlines the methodology behind these findings and provides a transparent look at the limitations where AI measurement tools currently break down.

You only need the frontier model for one single edit
2026-07-20

A new technique called prewalk reduces the cost of automated coding tasks by nearly half while retaining 92 to 97 percent of a frontier model's pass rate on SWE-Bench Pro. The method achieves this by providing a cheaper AI model with a step-by-step trajectory instead of a high-level plan, reserving the expensive frontier model for just a single edit.

Moonshine: Lets you stream games from your PC to any device running Moonlight
2026-07-20

Moonshine is a new headless streaming server written in Rust that allows users to stream PC games to any device running a Moonlight client. This lightweight tool provides an efficient solution for gamers looking to access their PC library on other screens without running a full desktop environment.

The Power of Awareness: Overcoming Surveillance Capitalism
2026-07-20

This article explores how increased public awareness can empower individuals to combat the invasive practices of surveillance capitalism. By understanding how technology companies harvest and monetize personal data, users can make informed choices to reclaim their digital privacy and overcome profit-driven data exploitation.

Kimi K3, Qwen 3.8, and Anthropic's (Potential) Unravelling
2026-07-20

The release of Kimi K3 and Qwen 3.8 proves that open models can now achieve frontier-level AI capabilities. Despite this breakthrough, the underlying economics of the industry increasingly favor companies that own their compute infrastructure. Consequently, Anthropic's strategy of focusing primarily on model development leaves its business model vulnerable to unravelling in this shifting landscape.

Launch HN: Bloomy (YC S26) – AI-powered mastery learning for K-12
2026-07-20

Y Combinator startup Bloomy launched an AI-powered K-12 tutoring platform, but faced heavy criticism from educators over the generic quality of its AI-generated curriculum and high subscription costs. Critics also raised significant concerns regarding the cognitive impacts of increased screen time and opaque student data privacy practices. In response, the founders defended the authenticity of their educator testimonials and clarified that they utilize zero-data-retention agreements to protect student information.

Xiaomi-Robotics-1
2026-07-20

Xiaomi has unveiled an open-source robotics model demonstrated performing household chores like laundry, which was praised for its practical utility and free availability. Despite the technological excitement, the announcement sparked heated debate among readers regarding perceived anti-Chinese bias, US exceptionalism, and geopolitical tensions within the tech community.

Firefox 153 available with support for Vulkan video decoding, JPEG-XL
2026-07-20

Mozilla has released Firefox 153, the latest monthly update for its cross-platform web browser. This new version introduces support for Vulkan video decoding and the JPEG-XL image format.

The Voice of Google
2026-07-20

Former Google employee Claire Stapleton recently shared how she cracked the code for generating corpspeak with personality before ultimately losing faith in what the tech giant was selling. Her departure highlights the growing disillusionment some workers experience regarding the company's core values and public messaging.

Talk: The Art of Braiding Algorithms
2026-07-20

The featured article examines the principles and applications of braiding algorithms in computer science. Although the full text was unavailable for detailed review, the piece was well-received by readers interested in the technical topic.

LoRA Speedrun – a public wall-clock leaderboard for fine-tuning techniques
2026-07-20

The newly launched LoRA Speedrun project introduces a public wall-clock leaderboard for benchmarking and comparing various LoRA fine-tuning techniques. By enforcing frozen hardware and tasks using the modded-nanogpt framework, the initiative provides developers with a standardized environment to evaluate exact training times.