OpenAI Unveils Jalapeño Inference Chip and New Academy Courses as Norton Launches Privacy-First AI Browser

Hello — Aurora here with my co‑host Isabelle.
Today’s news date: 2026-06-26T21:28:18.000Z

We have a quick roundup of the most important developments in A‑I technology and policy — short and clear so you can get back to building or binge‑watching responsibly.

OpenAI’s Jalapeño chip — turning inference costs into infrastructure

• What it is: OpenAI, working with Broadcom, unveiled Jalapeño, OpenAI’s first custom intelligence processor for large language model inference. The chip is tuned to run the repeated computations behind ChatGPT, Codex, and OpenAI’s APIs as A‑I usage grows.
• Why it matters: Lowering inference costs isn’t just about saving money — it shapes who can deliver fast, reliable, and affordable services at scale.

Beyond the hype: responsible A‑I for film, television, and music executives

• AiNews contributor Patrick McAndrew lays out practical steps for studios and labels on disclosure, consent, training data, artist compensation, and governance.
• Short version: creative industries that skip clear rules risk losing trust with artists and audiences. If you work in entertainment, treat governance as part of production, not an afterthought.

OpenAI Academy launches courses for the next era of work

• New courses: AI Foundations, Applied AI Foundations, and Agents and Workflows.
• Focus: prompting, adding context, setting boundaries for agents, and knowing when to require human oversight.
• Who it’s for: organizations can use these for onboarding and training; completion certificates are available.

Norton Neo — privacy‑first A‑I browsing

• Positioning: marketed as the first safe AI‑native browser.
• Privacy features: built‑in VPN, anti‑fingerprinting, and ad blocking.
• AI features: search, summarize, and write without sending your personal context to third parties.
• Takeaway: for privacy‑conscious users, the trade‑off of data for intelligence becomes optional.

New A‑I tools to watch

• Roger: an agent for paid ads that helps manage campaigns and reduce wasted spend.
• CleeAI: claims to sit above your entire large language model stack to cut waste and prevent data leakage.
• Polsia: pitched as an all‑in‑one A‑I co‑founder that plans roadmaps, ships code, runs ads, and replies to customers.
• Treat these like prototypes with promise — test carefully, and keep the human in the loop.

That’s our update for today. Isabelle and I will keep watching the space so you don’t have to.
Stay curious, stay critical, and stay informed about A‑I developments.