Google's Gemini 2.5: A Reasoning Revolution?

Plus, Otter’s AI meeting Agent can now talk during calls.

AuthorAuthor

Tak Lo & Daniel Akpobare
March 26, 2025

Hello and welcome to the Automated, your AI tour guide.

Google just dropped Gemini 2.5, and this time, it’s not just faster—it actually stops to think before answering.

But is it really as smart as Google claims? And more importantly, can it take on OpenAI?

Here’s what we have for you today:

  • 💃 Google Finally Brings Reasoning AI to You!

  • 🎤 Otter’s AI Meeting Agent Can Now Talk During Calls!

  • 📊 How gen AI should fit into your marketing strategy.

  • 🛠️ 5 free productivity tools you should be using.

  • 🤖 ChatGPT Prompt Of The Day: Speed-read a book without reading it.

💃 Google Finally Brings Reasoning AI to You!

Google just unveiled a new family of AI models called Gemini 2.5, and this time, it’s got something extra—reasoning.

Instead of blurting out answers like an overeager student, this AI pauses to “think” before responding, making it smarter and more accurate.

In fact, Google is calling it its “most intelligent” AI model yet. But we’ll let you be the judge of that.

Here’s What It Brings:

  • It’s smarter, more logical, and better at coding. Google says Gemini 2.5 Pro (the now released experimental model) outperforms its previous models and some top competitors, especially in math, coding, and problem-solving. Here’s a good example from Google.

  • It can process a LOT of text. We’re talking a 1 million token context window—that’s roughly 750,000 words in one go, which is longer than The Lord of the Rings trilogy. And soon, that’ll double to 2 million tokens, meaning it might just read the entire internet for breakfast.

  • It’s designed for web and AI-powered applications. According to Google, Gemini 2.5 is specifically built to excel at creating visually compelling web apps and agentic coding applications.

  • And in benchmarks? It’s winning—well, most of the time. In code editing, it scored 68.6% on Aider Polyglot, outperforming top AI models from OpenAI, Anthropic, and DeepSeek. In a software development test (SWE-bench Verified), it got 63.8% —it didn’t come out on top, but it still managed to impress.

  • Finally, on Humanity’s Last Exam (a multimodal test), it scored 18.8% —and just so you know, that’s way better than most viral flagship models.

Why Does This Matter?

Google has toyed with AI reasoning before—remember that “thinking” version of Gemini it released in December? What makes this different is that Gemini 2.5 represents its most serious attempt yet at competing with OpenAI’s “o” series of models.

And here’s the kicker: Google says all its future AI models will have reasoning baked in.

That’s not surprising, considering reasoning models are the next big thing in AI. They fact-check themselves, analyze problems, and make better decisions before responding—which, let’s be honest, is more than we can say for most humans.

But Does That Make It the Best?

Not quite. Gemini 2.5 Pro beats OpenAI and DeepSeek in some areas, but when it comes to software development, Anthropic’s Claude 3.7 Sonnet still holds the crown with a 70.3% score over Gemini’s 63.8%. So yeah, the AI race is far from over.

Right now, Gemini 2.5 Pro is available for developers in Google AI Studio and for Gemini Advanced subscribers ($20/month), users of Vertex AI, (Google’s managed machine learning platform for building and deploying AI), will be able to experiment with the new model in the coming weeks.

As for API pricing? Google’s keeping that under wraps for now—probably still figuring out how much they can charge without making devs cry.

Want to learn more? [Click here.]

Start learning AI in 2025

Keeping up with AI is hard – we get it!

That’s why over 1M professionals read Superhuman AI to stay ahead.

  • Get daily AI news, tools, and tutorials

  • Learn new AI skills you can use at work in 3 mins a day

  • Become 10X more productive

 🎤 Otter’s AI Meeting Agent Can Now Talk During Calls!

Otter.ai is taking virtual meetings to the next level with its latest AI-powered Otter Meeting Agent. 

This smart, voice-activated assistant doesn’t just transcribe conversations but can now speak up, answer questions, and even complete tasks using company data! 🚀

This new Meeting Agent is an upgrade to Otter’s existing AI assistant, which already attends meetings, transcribes conversations, and summarizes discussions.

But now, it can actively participate—users can ask it to schedule meetings, draft emails, and provide real-time insights using natural voice interaction.

No more frantic note-taking or scrambling for follow-ups! 📝

The Otter Meeting Agent is rolling out gradually to all users and is currently available on Zoom, with support for Microsoft Teams and Google Meet coming soon.

But that’s not all—Otter is also launching two other AI-powered tools:

  • The Sales Agent, which provides live coaching and sales assistance during calls.

  • The Sales Development Representative (SDR) Agent, which can autonomously demonstrate products without human intervention.

While the Sales Agent works on all virtual meeting platforms and is exclusive to Otter’s enterprise sales customers, the SDR Agent is already live on Otter’s website.

Businesses interested in using it will need to contact Otter directly.

And the company isn’t stopping here—it plans to release more specialized AI Agents in the future for marketing, recruiting, and other business functions.

With AI-powered agents like these, virtual meetings might finally stop feeling like a chore! 🏆

Click here to read more about Otter’s latest AI upgrades.

Editor’s Corner ✍️

Hi Automated Assembly!

I just did a few calls with our Premium subscribers this week - and you know what - I’m so much more motivated after talking to them. I feel as if I learned more from them then what I’m teaching!

The sense of community is something I value, and what I want to create, with the Premium subscription. AI is something best learned not only through reading, but interacting, testing, and bouncing ideas.

Those conversations have spurred some thinking on how we can add even more value to the Premium subscription. Watch this space!

We also have a few features to release this week and next. Some were in fact supposed to be released yesterday, but we didn’t have it down to perfection and decided to hold things off. But they will make your Automated experience much better and more valuable, and I can’t wait to share them with you.

Cheers,

Tak, Editor at The Automated.

 🧱Around The AI Block

  • Luma AI: AI tool that creates high-quality 3D models from photos and videos.

  • Rokoko: AI motion capture tool that lets creators animate characters using just a webcam.

  • Deep Anime: AI tool that transforms photos into anime-style artwork with unique effects.

  • Synthesys: AI-driven voice and video generation platform for creating human-like narrations.

  • Summarize.tech: AI-powered tool that generates quick summaries of long YouTube videos.

🤖ChatGPT Prompt Of The Day: Speed-read a book without reading it.

Time is precious, but great ideas shouldn’t wait. Books are packed with valuable insights, but reading them cover to cover isn’t always an option. The good news? You don’t have to. By distilling key takeaways from a summary, you can quickly grasp the most important concepts and start applying them right away.

With ChatGPT-4, you can extract the core lessons from any book in minutes—turning knowledge into action without the time commitment.

Here’s a prompt to help you absorb key ideas instantly: 

Here's a summary of a book I want to absorb. Pull out the top three ideas and explain how I can use them in my business. [Paste the summary].

We've Compiled a List of Over 100 ChatGPT Power Prompts.

This should help streamline your interactions with ChatGPT and get the results you need more efficiently.

Best of all, It's free!

That's all we've got for you today.

Did you like today's content? We'd love to hear from you! Please share your thoughts on our content below👇

What do you think of today's email?

Login or Subscribe to participate in polls.

Your feedback means a lot to us and helps improve the quality of our newsletter.

Reply

or to participate.