Google unveiled Gemini 4 Argon today, its first major new frontier model since Gemini 3. It handles up to one million tokens of context, scores 77.9% on a leading coding benchmark, and launches at intro pricing of $2/$10 per million input/output tokens. This is Google's clearest statement yet that it is back competing at the frontier.
Why it matters: More capable models at lower prices mean better tools for students, professionals, and developers — not just big companies. If you're building anything with AI or choosing an AI assistant for work, you'll feel these upgrades within weeks.
OpenAI's GPT-6.1 Sol: near-flagship coding at one-fifth the cost
OpenAI released GPT-6.1 Sol for ChatGPT Work, its Codex coding tool, and the API. It delivers coding and computer-use performance close to their top-tier model but at roughly one-fifth the price. Developers and power users get serious capability without the premium bill.
Why it matters: The price gap between powerful AI and affordable AI keeps closing. If you've wanted to experiment with AI coding tools but the cost put you off, this is a good moment to try again.
Uber fined €824M for using AI to fire drivers without human review
The Dutch data protection authority issued a record €824 million fine against Uber for using an automated AI system to deactivate driver accounts — effectively ending their livelihoods — without any meaningful human involvement in the decision. It is the largest GDPR fine ever issued specifically for an automated employment decision under Article 22.
Why it matters: This ruling establishes a new standard across Europe: AI cannot make or carry out consequential employment decisions without a human in the loop. Anyone using AI in HR, hiring, or workforce management needs to take note — regulators are no longer waiting.
India makes AI a compulsory school subject starting this academic year
Following last year's policy announcement, India's Ministry of Education is now implementing AI and computational thinking as mandatory subjects from Class 3 across all CBSE, KVS, and NVS schools in the 2026-27 academic year. The Education Ministry outlined a national rollout plan covering teacher training, digital infrastructure, and partnerships with industry at the India AI Impact Summit.
Why it matters: This is no longer a proposal — it is a live curriculum change affecting tens of millions of children in India right now. If you are a parent, teacher, or work in EdTech in India, the shift has already started.
Meta's Neural Band reads your muscles to text without a screen
Meta released a new handwriting feature for its Neural Band wearable that uses surface electromyography — reading electrical signals in your forearm muscles — to detect hand movements and compose text. No phone, no screen, no keyboard required. You write in the air and the message appears.
Why it matters: This is AI moving off the screen and into the body. It sounds futuristic, but wearable AI that works without a device in your hand is the direction consumer AI is heading. Watch this space.
AI cybersecurity startup Armadin raises $255M at a $2.5B valuation
Armadin, an AI-focused cybersecurity company founded by Mandiant's Kevin Mandia, raised $255.5 million in a Series B round just seven months after emerging from stealth. The company uses AI for threat intelligence — the kind of work that catches cyberattacks before they happen. Investors are clearly betting that AI-powered defence keeps pace with AI-powered attacks.
Why it matters: Cybersecurity is one of the fastest-growing career paths in AI. As AI tools create new attack surfaces, the demand for people who understand both AI and security is only going up — worth knowing if you're thinking about where to build skills.
Who controls AI law in America — states or Washington?
California, Colorado, Utah, and Washington have all passed their own AI safety, transparency, and discrimination laws this year. The tech industry is pushing Congress hard for a single federal law that would override all of them. October is the peak of that fight. The outcome shapes how AI is regulated across the world's largest tech economy — and sets the template for other countries watching closely.
Deepfake job fraud hits crisis levels + a new open-source coding agent launches
Deepfake fraud attempts in hiring have grown 1,300% over two years, and new consumer-grade video tools from OpenAI and Google now make convincing fakes possible for anyone with a browser. Voice cloning has crossed the threshold where it's indistinguishable from the real person. Separately, Earendil released Pi 1.0 — a small, open-source coding agent that runs in the terminal — giving developers a free, stable tool for everyday coding tasks.