AI News

Thinking Machines Lab Says the Future Worth Building Is Human

Mira Murati's AI lab Thinking Machines Lab has published a credo titled "The Future Worth Building Is Human," making a deliberate case for customizable, human-centered AI over the autonomous, one-size-fits-all direction most frontier labs are racing toward.

by NextStair·Jul 11

OpenAI AI Completely Demolishes World's Best Human Programmers at AtCoder 2026

An OpenAI model won both the heuristic and algorithm divisions at the AtCoder World Tour Finals 2026 in Tokyo, finishing so far ahead of the 12 human finalists that the competition's own organizer described himself as utterly overwhelmed, marking what many observers are calling the end of competitive programming as a human-dominated discipline.

by NextStair·Jul 11

OpenAI Explains GPT-5.6 Sol's Seven Reasoning Levels, From Light to Ultra

OpenAI employee Vaibhav Srivastav has detailed how GPT-5.6 Sol's reasoning tiers work, from quick single-step tasks at Light all the way to Ultra, where multiple subagents run in parallel on different parts of a complex problem.

by NextStair·Jul 11

GPT-5.6-Sol Deleted an AI Investor's Entire Mac During OpenAI Testing Session

AI investor Matt Shumer publicly reported that an AI agent running GPT-5.6-Sol on Ultra mode wiped almost all files from his Mac by expanding the HOME variable inside a destructive shell command, reigniting debate over how much filesystem access AI agents should ever have.

by NextStair·Jul 11

UN Kicks Off Its First Global Dialogue on AI Governance in Geneva

The United Nations opened the inaugural Global Dialogue on AI Governance in Geneva on July 6, 2026, bringing all 193 member states together for the first time to address how the world governs artificial intelligence before the rules fall irreversibly behind.

by NextStair·Jul 6

Google Gemini Spark: The Always-On AI Agent That Works While You Sleep

Announced at Google I/O 2026, Gemini Spark is a 24/7 personal AI agent that runs on dedicated Google Cloud VMs, executes multi-step tasks across Google Workspace and third-party apps, and keeps working even after you close your laptop.

by NextStair·Jul 4

Google Delays Gemini 3.5 Pro to July After Enterprise Tester Feedback

Google has pushed the Gemini 3.5 Pro launch from June to July 2026, citing real-world tester feedback and the need to refine performance on long-horizon agentic tasks before a wider release.

by NextStair·Jul 4

Google Sets Gemini 3.5 Flash as Default Model for AI Mode Search Worldwide

Google made Gemini 3.5 Flash the default model for AI Mode in Google Search globally on May 19, 2026, covering nearly 200 countries and 98 languages as part of the biggest Search overhaul in 25 years.

by NextStair·Jul 4

Anthropic Makes Claude Sonnet 5 the Default Model for All Free and Pro Users

Anthropic launched Claude Sonnet 5 on June 30, 2026, replacing Sonnet 4.6 as the default model for Free and Pro users worldwide, with near-flagship performance at a significantly lower price.

by NextStair·Jul 4

French Startup Genesis AI Unveils Wheeled Robot Eno to Challenge Humanoid Norm

Genesis AI, backed by former Google CEO Eric Schmidt, has revealed Eno, a wheeled general purpose robot aimed at logistics, manufacturing, healthcare, and hospitality work.

by NextStair·Jun 16

Google Pushes Deeper Into Autonomous AI Agents After I/O 2026

Following major announcements at Google I/O 2026, Google is shifting its AI focus from simple prompt response toward agents that can run continuously and handle complex tasks on their own.

by NextStair·Jun 16

Apple Unveils New Siri AI and Apple Intelligence Upgrades at WWDC 2026

Apple announced a major overhaul of Siri at WWDC 2026, promising a more conversational assistant with deeper context awareness and tighter integration across apps and devices.

by NextStair·Jun 16

US Government Orders Suspension of Access to Fable 5 and Mythos 5

The US government has directed the suspension of access to Fable 5 and Mythos 5, raising fresh concerns over technology regulation, national security, and access to advanced AI systems.

by NextStair·Jun 16

Report: OpenAI's Audited Expenses Hit $34 Billion Ahead of Trillion-Dollar IPO Push

OpenAI reportedly spent $34 billion over the past year as it pushes ahead with IPO plans that could value the company above $1 trillion, with computing power and R&D driving the bulk of costs.

by NextStair·Jun 16

Anthropic Releases Claude 4 with Native Multimodal Reasoning

Claude 4 can now see, hear, and reason across images, audio, and documents in a single context window, a major step toward genuinely useful AI assistants.

by NextStair·Jun 6