Google’s New AI Tools: A No-Hype Guide to What Actually Matters

Free AI tools that actually work: Google's new releases let you analyze videos, conduct research, and generate incredible images - all available today.

Google made about a dozen releases in as many days. From Gemini 2.0 to Willow's quantum breakthroughs, this is your guide to what came out and what's actually useful.

What Happened

Foundation: Gemini 2.0 Flash

Available now: API for developers, AI Studio for testing, Gemini Advanced for mobile

• Twice the speed of 1.5 Pro, with better performance across benchmarks
• Handles text, images, video, and audio simultaneously - Google’s the best in the biz at image and video understanding
• They just released an experimental version called Flash Thinking that shows its reasoning process explicitly - their competitor to o1 making it easier to understand how it reaches conclusions
• Video analysis tool in AI Studio: Parse and understand video content, enabling applications like automatically tagging objects, identifying key scenes, or summarizing events in recorded footage.
• Powers two experimental agents: Project Mariner which can take over your browser and Project Astra which lets you Facetime the model

Research & Knowledge Tools

• Deep Research: Browses hundreds of sites, delivers cited reports in minutes - Great for exploring a whole new landscape
  • Available now in Gemini Advanced
• NotebookLM Plus: Get a podcast made about anything, chat with sources - Great for going into the weeds but making it fun
  • Available now at notebooklm.google.com
• FACTS Grounding: New benchmark measuring AI's factual accuracy

Visual Creation Suite

• Whisk: Remix images with a subject, scene, and style - Available now & great for mixing ideas together
• Imagen 3: Enhanced image generation across artistic styles - Available now & great for generating new ideas and making variations
• Veo 2: High-quality video generation (waitlist)

Quantum Future

• The Willow quantum chip is a huge breakthrough, it performed calculations that would take supercomputers millennia in minutes.

My Take

Let's be clear: these aren't just demos. I've spent the week testing each tool, and while not everything's perfect, there's genuine utility here.

While OpenAI has dominated headlines, Google's wave of releases shows that the sleeping giant is waking up. Google is finally leveraging its massive advantages in data (think YouTube for visual models) and infrastructure (DeepMind combined with Google Brain into one team).

Gemini 2.0 Flash running faster, cheaper, and better than 1.5 PRO points to major breakthroughs. Not only that, but they’re starting to make actual useful tools! Honestly, I think Logan Kilpatrick, the former head of developer relations at OpenAI who left and became Lead Product for Google AI Studio, has played a massive role in it. He’s worth the X follow.

We're already seeing glimpses of what's possible. Tldraw's creating new ways to compute visually, Volley's reinventing gaming interactions, and Toonsutra's breaking language barriers in comics. All powered by Flash 2.0. These aren't AI companies – they're domain experts using AI to solve real problems they deeply understand.

Yes, there are limitations. API rate limits need work, and monetization remains a question mark (though Google's search revenue gives them plenty of runway). While Claude still leads in pure language tasks, Google's betting on a multimodal future + supporting devs – and they might be right.

I’m personally really excited to see how these tools are starting to bridge the digital-physical divide. When AI can truly understand video or images, we unlock more ways to have AI help not only the digital world but the physical one too.

Anyway, last week I said the next wave of breakthroughs won't come from AI experts. These releases make that even more likely. As these tools become more accessible and capable across text, audio, and visual domains, there are even more ways where they can empower YOU!

What You Can Do Today

• Jump into AI Studio right now: The video analysis tool is wild, and you can test Flash 2.0 for free. If you're building something, the API is ready (just watch the rate limits, you can always test with Flash 2.0 and if you need to run at scale, use 1.5 for now).
• Need to dive deep into a topic? Grab a Gemini Advanced trial. Deep Research is actually useful. I've been using it to research personal knowledge bases and it's saved me hours (working on something I’ll announce next year). Once you’ve got some beefy readings to dive into, use NotebookLM – it’s actually fun.
• For the visual folks: Whisk and ImageFX are both live in Google Labs. Whisk is perfect if you have images or concepts you want to combine. ImageFX is great for bringing new ideas to life. Pro tip: Use Gemini to turn any image into a prompt to start making variations.
• Curious about the quantum? Andy Matuschak’s Quantum.country introduces it in a mnemonic medium, which tries to make it almost effortless to remember what you read. Or I’ll always recommend The Beginning of Infinity by David Deutsch who pioneered the field of quantum computation. It’s a pretty approachable and inspiring read that touches as much on society as quantum. Willow's breakthrough is fascinating but probably won't impact your work tomorrow.

Everything except Veo is available now - and honestly, that's plenty to play with. Pick one tool that matches your current project and start there. The best way to understand these tools is to actually use them. I’m excited to see what you create.