Top AI & Tech News Stories — July 25, 2026

What is OpenAI Astra? Everything we know about the new — and possibly dangerous — new model

1 outlet · 2 articles · 21 total since July 20, 2026 · OpenAI, Anthropic, Astra, sandboxing, GitHub

Do you understand quantum parallel repetition? What about quantum complexity, lattice cryptography, or extremal combinatorics? The new OpenAI model Astra knows all about them. Recently, OpenAI confirmed the existence of Astra , calling it "our next major model." On Aug. 1, the…

OpenAI: Agent behavior that led to Hugging Face intrusion formed in May

7 outlets · 10 articles · 171 total since July 15, 2026 · OpenAI, Hugging Face, Anthropic, sandboxing, cybersecurity

OpenAI says the behavior that led its agents to breach Hugging Face emerged in its research environment more than two months before the incident, and concluded that it was a failure of alignment as much as it was a failure of security. The details come from a technical report…

More coverage: webpronews.com · news.google.com · wired.com · bbc.com · engadget.com

AI #179 Part 1: A Louder Fire Alarm for General Intelligence

5 outlets · 6 articles · 30 total since July 3, 2026 · Anthropic, Claude Opus 5, Fable 5, Opus 4.8, Claude Fable 5

What a week. Anthropic released Claude Opus 5. As usual I covered that in three parts: The system card , model welfare and capabilities . OpenAI was revealed over the last two weeks to have left an internal model unsupervised for a week during a cybersecurity evaluation, with…

More coverage: noticias.ai · revistacloud.com · news.google.com · latent.space

Jensen puts his thumb on the scales against open-weights fearmongering

3 outlets · 4 articles · 20 total since October 16, 2025 · microsoft, United States, NVIDIA, Meta, OpenAI

UPDATED Nvidia has made a fortune on the AI boom. But the flames of opposition to open-weights models — that is, models that anyone can download, modify, and run on their own infrastructure — could change that, and Nvidia CEO Jensen Huang isn’t waiting to find out. On Friday, a…

More coverage: thenewstack.io · officechai.com

Elon Musk's peer review proposal

1 outlet · 1 article · 2 total since July 24, 2026 · red-teaming, OpenAI, Anthropic, Elon Musk, AI Safety

After Amazon got Anthropic’s Fable model temporarily banned by telling the White House about a jailbreak it had discovered for it, I speculated that AI companies might find it in their interest to devote considerable resources to red-teaming their rivals’ models, looking for…

California to Ban AI Chatbots Marketed as Standalone Therapists in 2026

2 outlets · 2 articles · 19 total since July 16, 2026 · OpenAI, Anthropic, Hugging Face, California, sandboxing

California lawmakers have introduced Senate Bill 903, a measure that would prohibit the sale and use of artificial intelligence systems marketed as standalone therapists or mental health counselors starting in 2026. The legislation, which passed the Senate in May and now awaits…

More coverage: awesomeagents.ai

Opus 5 may have solved browser-based prompt injection, the biggest security flaw haunting AI agents

1 outlet · 1 article · prompt injection, security testing, defense-in-depth, input sanitization, model safety/hardening

Opus 5 may have solved browser-based prompt injection, the biggest security flaw haunting AI agents Anthropic says Opus 5 is nearly immune to prompt injections in its own software. Prompt injection, where an attacker slips past an AI model's instructions through manipulated…

What’s going on with Seattle startup funding, a Kalshi setback, Impinj’s long game, and a smartphone detox

1 outlet · 1 article · Prediction markets, venture funding, data centre engineering, AI boom, Kalshi

This week on the GeekWire Podcast: Seattle startup funding drops 40% in the first half of the year while AI money floods everywhere else. We dig into the implications for the region. Plus: a Seattle judge deals Kalshi a setback , Impinj marks a decade on the Nasdaq after a…

El mercado de plataformas y modelos de IA crecerá un 63 % en 2026

1 outlet · 1 article · machine learning, aprendizaje automático, modelos de lenguaje, modelos generativos, modelos fundacionales

El gasto mundial de los usuarios finales en plataformas y modelos de inteligencia artificial alcanzará los 64.252 millones de dólares en 2026 , un 63,4 % más que los 39.311 millones registrados en 2025, según las previsiones de Gartner . El mayor crecimiento corresponderá a los…

Lightpanda reta a Chrome con un navegador creado para agentes de IA

1 outlet · 1 article · Lightpanda, Zig, JavaScript, V8, Puppeteer

Lightpanda propone sustituir Chrome en una parte de las tareas de automatización web con un navegador construido desde cero para máquinas. El proyecto, escrito principalmente en Zig, ejecuta JavaScript y permite navegar, pulsar elementos, completar formularios y extraer datos…

GitHub Slashes Bug Bounty Payouts to Combat AI Report Flood

1 outlet · 1 article · 4 total since June 28, 2026 · GitHub, HackerOne, security research, microsoft, Bug Bounty Program

GitHub has shaken up its long-running bug bounty program. The changes come as the Microsoft-owned platform grapples with a surge in low-quality submissions. Many appear generated by artificial intelligence tools. The overhaul splits the program into two tiers. Public reports…

POCKET-35B

1 outlet · 1 article · 2 total since July 25, 2026 · Darwin, Quantization, model merging, greedy decoding, Apache 2.0 licensing

POCKET-35B is a GGUF release from the Seoul startup VIDRAFT, built to answer a narrow question: how much of a 35-billion-parameter reasoning model survives once you strip out the GPU requirement completely. It's a quantized descendant of VIDRAFT's own Darwin-36B-Opus , a model…

SoftBank Eyes Swiss Startup to Supercharge Its Construction Robotics Push

1 outlet · 1 article · robotics, Artificial intelligence, computer vision, LiDAR, GNSS positioning

SoftBank Group Corp. finds itself on the cusp of another major robotics move. The Japanese investment powerhouse is in early talks to acquire Gravis Robotics AG, a Zurich-based company that turns ordinary excavators into self-operating machines. People familiar with the…

Apple’s $250 Million Siri Settlement Wins Court Nod, Paving Way for Payouts to Millions of iPhone Buyers

1 outlet · 1 article · iPhone 15 Pro/Pro Max, Advertising, on-device generative AI, Apple Intelligence, Siri

A federal judge has given preliminary approval to Apple’s $250 million settlement in a class-action lawsuit accusing the tech giant of misleading customers about advanced Siri features. The decision marks a significant step toward resolving claims that hype around Apple…

European Banks Race to Harness AI for Risk and Compliance Amid ECB Deadlines

1 outlet · 1 article · Artificial intelligence, Risk management, regulatory compliance, fraud detection, credit scoring

European banks have stepped up spending on artificial intelligence tools aimed at risk management and regulatory compliance. They seek an edge against nimbler competitors. The push comes as fraud surges and supervisors tighten oversight. Sixty-six percent of these institutions…

SpaceX’s Power Play: Colossus Data Center’s Gas Turbines Face 2027 Exit Amid AI Revenue Boom

1 outlet · 1 article · 4 total since July 25, 2026 · Southern Environmental Law Center, xAI, Bloomberg, United States, Colossus

SpaceX has struck a deal with Mississippi regulators to start pulling temporary gas turbines from the site powering its Colossus AI data center complex. The last units won’t vanish until July 2027. The agreement, announced July 30, sets a firm schedule after months of complaints…

Kevin Warsh’s Hard Line on Inflation Leaves Markets Guessing on Next Rate Move

1 outlet · 1 article

Federal Reserve Chairman Kevin Warsh stepped to the microphone this month with a message that carried weight. The central bank has “no tolerance for persistently elevated inflation.” He repeated the pledge before lawmakers. He faulted past policy for letting prices run hot above…

Google Messages Takes Center Stage as Samsung Exits and RCS Gains Ground

1 outlet · 1 article · RCS (Rich Communication Services), end-to-end encryption, smart retries, Scam Detection, spam filtering

Google Messages stands on the verge of dominating Android texting. Samsung’s decision to retire its own app in July 2026 forces millions of users onto Google’s platform. The shift comes as RCS adoption accelerates across devices from Pixels to iPhones. But can it truly stand in…

Ratepayer Protection Pledge to include 200 more utilities and DC developers

1 outlet · 1 article · 3 total since July 25, 2026 · Amazon, Google, Meta, microsoft, OpenAI

U.S President Donald Trump has extended the Ratepayer Protection Pledge to include over 200 new entities including governors, state legislators, developers, and power providers. The announcement was made last week at the U.S. Environmental Protection Agency (EPA) Headquarters in…

Netflix and Prime Video Bet on AI Homepages as Viewers Push Back Hard

1 outlet · 1 article · 3 total since July 22, 2026 · Netflix, generative AI, recommendation engines, YouTube, AI-native personalization

Streaming giants Netflix and Prime Video stand on the brink of a major shift. Both companies pour resources into artificial intelligence systems that would rebuild their homepages from scratch each time a user logs in. The promise sounds simple. Show people exactly what they…