- One more Thing in AI
- Posts
- Gemini 3 and Grok 4.1 are out!
Gemini 3 and Grok 4.1 are out!
Plus: I tested Nano Banana Pro
Renjit Philip
November 23, 2025 • Estimated Reading Time: 12 minutes
Date: 23-Nov-2025
Hey AI enthusiast,
And just when you think you’ve seen it all…there’s always One More Thing in AI.
Table of Contents
I hope you enjoy this edition!
Best,
Renjit
PS: If you want to check out how to implement AI agents in your business and get more revenue with the same number of employees, speak to me:
Meet Gemini 3: Google’s Next-Gen AI for Builders
Imagine an AI that not only answers questions but becomes your go-to partner for planning, building and creating. That’s the direction Gemini 3 is taking.
Summary Highlights
• 📊 Google announced Gemini 3 as its most intelligent model so far, with upgraded reasoning and multimodal capabilities.
• 🔍 The model outperforms prior versions across key benchmarks, including reasoning tests, math challenges and tool-use scenarios.
• 🧠 Gemini 3 introduces “Deep Think” mode, intended for more complex, long-horizon tasks (available soon via Google AI Ultra subscribers).
• 🛠 It supports three main value drivers: learn anything (text, images, video, code), build anything (zero-shot generation, agentic workflows), plan anything (multi-step, long-context reasoning).
• 📦 For builders: it’s accessible through the Gemini app, AI Mode in Search, AI Studio, Vertex AI, and a new dev platform called Google Antigravity, designed for agent-first workflows. If you are confused by this proliferation of tools, I am with you! I have found that for practical purposes, Google AI studio is useful.
• ✅ On the safety front: Google says Gemini 3 underwent its most rigorous safety evaluation yet; it includes improved protection against misuse, prompt injection and sycophantic responses.
With Gemini 3 Google is staking a claim on the future of AI-enabled work. For you as a founder or advisor that means the time to explore agentic architectures, long‐context workflows and AI-first development is now. Right now, it is clear that Google is back on top in the AI game. Your move, Open AI.
🎙️ Meet Grok 4.1: The Chatbot That Learns to Feel and Act
Imagine chatting with an AI that actually gets you, not just the words, but the mood behind them. That’s the promise behind the latest release of Grok 4.1 from xAI. It’s not just smarter; it’s more human-aware.
✨ What’s new:
• The model is described as “exceptionally capable in creative, emotional, and collaborative interactions” while keeping its core intelligence intact.
• Developers used the same large-scale reinforcement-learning infrastructure as Grok 4 but added new methods to optimise style, personality, helpfulness and alignment.
🔧 Enterprise & tool-focus variant:
• There’s a version called Grok 4.1 Fast designed for tool-calling and enterprise use. It supports a 2 million token context window and features the Agent Tools API (real-time data, code execution, web search).
• In tests (τ²-bench Telecom), Grok 4.1 Fast achieved high marks for real-world agent workflows. You can see that XAi is targeting the enterprise market with this release. I think this is unlikely to work given Microsoft 365’s strangled hold of the enterprise market with its Copilot products.
👥 Why it matters:
This model suggests a shift from purely “smart-answers” to AI that understands context, tone and even emotion.
That means better customer support bots, smarter virtual advisors, and richer conversational UI. Grok 4.1 signals that conversational AI is evolving; less “robotic answer machine” and more “empathetic partner in dialogue”.
How LangChain Built an Agent Engineering Platform with Big Ambitions
The team at LangChain just raised a monster round and shifted gears from library to full platform.
Summary Highlights
• 📈 They raised $125 million at a $1.25 billion valuation. Phenomenal amount but, is it too late?
• 🧠 Their goal: build a complete platform for “agent engineering”, turning LLMs into systems with access to data, APIs, memory and long-running tasks.
• 🔧 Originally they focused on the open-source langchain library. Now they’ve rewired the library (v1.0) and introduced LangGraph for orchestration and memory.
•👀 They released major updates:
• A rewritten langchain package emphasizing stability and no breaking changes until 2.0.
• LangGraph runtime for low-level control of custom agents.
• Insights Agent in LangSmith: automatically categorises agent behaviour patterns.
• Agent Builder (private preview) for no-code text-to-agent experiences aimed at business users.
• 🌍 Growth indicators: 90 million combined monthly downloads of langchain + langgraph. 35% of Fortune 500 companies use their services. LangSmith’s monthly trace volume grew 12× year-over-year.
• 🤝 They named a set of investors that are also customers or partners: IVP (led the round), existing investors like Sequoia, Benchmark, Amplify, and new ones such as CapitalG and Sapphire Ventures.
Why this matters
• From a SaaS and M&A perspective, this tells us: open-source tools + enterprise platform = strong GTM play. For buyers or investors in this space, platforms that bridge OSS dev communities with enterprise-grade offerings are high value. They’re betting that “agents” are the next architecture in generative AI.
Master ChatGPT for Work Success
ChatGPT is revolutionizing how we work, but most people barely scratch the surface. Subscribe to Mindstream for free and unlock 5 essential resources including templates, workflows, and expert strategies for 2025. Whether you're writing emails, analyzing data, or streamlining tasks, this bundle shows you exactly how to save hours every week.
Learn AI: I tested out Google Nano Banana Pro
Google Nano Banana Pro Prompts that I tested out in the video:
1. High-quality flat lay photography creating a DIY infographic that simply explains how digital fintechs works, arranged on a clean, light gray textured background. The visual story flows from left to right in clear steps: Content is based on this: https://www.renjitphilip.com/root-insurance-business-model-teardown
Simple, clean black arrows are hand-drawn onto the background to guide the viewer's eye, clearly marking the flow of payments and information. The overall mood is educational, modern, and easy to understand. The image is shot from a top-down, bird's-eye view with soft, even lighting that minimizes shadows and keeps the focus on the process. Format 16:9
2. Change the Text on the can to read Fresh Ai Juice, then Translate all the English text on the three yellow, black and blue cans into Arabic and Hindi, while keeping everything else the same. The font color of the text should be the most visible as per the color wheel complement
3. Create a smooth logo in a graphic style is a vibrant and playful form of typographic illustration, deeply rooted in the retro aesthetics of the 1960s and 1970s loosely based on the sketch Its defining feature is a groovy, psychedelic-inspired typeface characterized by soft, rounded, and fluid letterforms. Don't exactly follow the sketch, get inspired from it. The letters are skillfully distorted, stretched, and compressed, abandoning rigid structure to flow together and form a cohesive, recognizable shape. This technique, known as a calligram, masterfully merges text and image, where the word's form visually embodies its meaning. The word "FS Brew Podcast" is artfully arranged into the fluid silhouette of a wave. The design is a clever visual pun, making the message instantly accessible and memorable. The color palette reinforces the vintage feel, employing a simple two-toned scheme with warm, often muted or earthy colors light blue background and deep blue logo. This choice enhances the nostalgic charm of the artwork. The overall effect is one of whimsical nostalgia and clever graphic design. It's a bold yet approachable style that communicates a simple, positive message through the seamless integration of shape and word, creating an immediate and delightful visual impact.
4. Now create brand identity system one by one, use 10 high quality mockups with variety of relevant products, ads, billboards, bus stop, etc. generate one at a time, 16:9 each
5. A medium shot of the 14 fluffy characters sitting squeezed together side-by-side on a worn beige fabric sofa and on the floor. They are all facing forwards, watching a vintage, wooden-boxed television set placed on a low wooden table in front of the sofa. The room is dimly lit, with warm light from a window on the left and the glow from the TV illuminating the creatures' faces and fluffy textures. The background is a cozy, slightly cluttered living room with a braided rug, a bookshelf with old books, and rustic kitchen elements in the background. The overall atmosphere is warm, cozy, and amused
6. This is a high-angle, almost top-down photograph of a crowd of people dispersed across a vast, pale, sun-drenched surface, likely a concrete plaza or a salt flat.
The composition feels spacious and uncentered, with figures scattered unevenly throughout the frame. Similar to its companion piece, the most striking artistic element is the interplay of light and shadow. A low, bright sun casts incredibly long, dark, and sharply defined shadows from each person, making the shadow often appear more prominent than the individual casting it.
These elongated shadows create a powerful graphic pattern, a rhythmic series of dark lines that slice across the pale ground, all pointing in the same direction. The people themselves are small figures, distinguished by the varied colors of their attire, which appear as small dots of red, black, and blue in the expansive scene.
The minimalist, almost bleached-out color palette of the ground starkly contrasts with the dark shadows and colorful specks of humanity, creating a sophisticated and highly stylized composition that explores themes of scale, individuality, and the collective
Reply