Hey AI Enthusiasts!

Today’s roundup highlights two major breakthroughs. from OpenAI’s newest model upgrade to DeepMind pushing agents closer to real AGI-level gameplay.

Let’s dive in!

In today’s insights:

🤖 OpenAI launches GPT-5.1 with sharper reasoning and a warmer tone
🎮 DeepMind debuts SIMA 2, an AI agent that learns games like a real player

_{Read time: 3 minutes.}

OpenAI
 🔥 1. OpenAI Releases GPT-5.1 With Warmer Tone and Smarter Reasoning

The AI Field: OpenAI has rolled out GPT-5.1, a major update designed to improve reasoning, conversation stability and overall “human warmth” compared to GPT-5.

Details:

• Two versions: GPT-5.1 Instant and GPT-5.1 Thinking.

• New personality presets in chat: Default, Professional, Friendly, Candid, Quirky, Efficient, Nerdy, and Cynical.

• More adaptive reasoning: the model now slows down for complex questions and speeds up for simple ones.

• Developer updates include a larger 24-hour prompt cache and new tools like apply_patch and shell for coding workflows.

Why This Matters:
GPT-5.1 isn’t just more capable, it’s more usable. OpenAI is shifting toward personalization, tone control, and real reliability, which signals how future models will balance raw power with human-like interaction.

👉 Read more here

From Hype to Production: Voice AI in 2025

Voice AI has crossed into production. Deepgram’s 2025 State of Voice AI Report with Opus Research quantifies how 400 senior leaders - many at $100M+ enterprises - are budgeting, shipping, and measuring results.

Adoption is near-universal (97%), budgets are rising (84%), yet only 21% are very satisfied with legacy agents. And that gap is the opportunity: using human-like agents that handle real tasks, reduce wait times, and lift CSAT.

Get benchmarks to compare your roadmap, the first use cases breaking through (customer service, order capture, task automation), and the capabilities that separate leaders from laggards - latency, accuracy, tooling, and integration. Use the findings to prioritize quick wins now and build a scalable plan for 2026.

Get the Full Report

google
🎮 2. DeepMind Unveils SIMA 2, an AI Agent That Learns Video Games Like a Real Player

The AI Field: Google DeepMind has released SIMA 2, a next-generation general-purpose AI agent that can learn and play a wide range of video games. from No Man’s Sky to Valheim and Goat Simulator 3, using only high-level instructions.

Details:

• Powered by Google Gemini, enabling the agent to understand goals, reason about environments, and adapt to new games.

• SIMA 2 isn’t hard-coded for specific titles, it learns mechanics, physics, and tasks directly from gameplay.

• Researchers see gaming as a gateway to building real-world AGI agents that can navigate physical environments and perform complex tasks.

Why This Matters: This is one of the clearest steps toward AGI-style agents. If an AI can understand new worlds, learn skills on the fly, and accomplish open-ended goals, the leap to real-world robotics becomes much smaller.

👉 Read more here

Disney tests AI video tools for Disney+ users
Disney Plus may let users create and consume AI-generated videos: CEO Bob Iger says the platform could introduce short-form AI video creation tools directly inside the app.

NotebookLM gets Gemini “Deep Research” mode
Google’s NotebookLM now includes Deep Research: Users can choose between fast or detailed analysis, pull data from Sheets, Drive, and Docs, and generate full research plans with Gemini.

Firefox adds new “AI Window” browsing mode
Mozilla introduces an opt-in AI Window for Firefox: A privacy-focused AI sidebar with chatbot tools and user-selectable models baked directly into the browser.

Anthropic invests $50B in new US data centers
Anthropic commits $50 billion to AI infrastructure: Massive new data center builds in Texas and New York to support future frontier-model development.

Golden Nuggets

🤖 OpenAI launches GPT-5.1
🎮 DeepMind debuts SIMA 2

Until next time!
Olle | Founder of The AI Field