Picking a Model

For anyone serious about leveraging AI, you've got three main players in the race (let's continue the unofficial F1 movie sponsorship here):

Anthropic's Claude, Google's Gemini, and OpenAI's ChatGPT.

Seriously, you can't go wrong with any of them. Think of it like choosing between a Ford, a Chevy, or a Ram – they all get the job done, just with slightly different bells and whistles.

What do you get with these models?

Fast and powerful models, a voice mode for when your fingers are tired (or you’re biking), the ability to process images and documents, code execution, slick mobile apps, and (mainly) the ability to churn out images and even video (sorry, Claude, you're sitting out the image party on this one).

Most of the premium features are behind a paywall, so budget around $20 per month to access the whole experience.

Consider it an investment in your future self, who can relax while AI handles some of the work.

Now, for those of you thinking, "But what about my other tech toys?"

Sure, there are specialized AI tools, and some folks swear by them.

But for general-purpose AI, stick to the three main.

Elon Musk's Grok is available if you're a heavy X user, but they're somewhat secretive about how their AI works. Given its recent “way off the rails” suggestion, I would suggest HR folks stay away for now.

Microsoft's Copilot is integrated into Windows, which is convenient, but it can be finicky.

And DeepSeek r1, a Chinese model, is capable and free, but its long-term future is hazy. So, yeah, back to the big three.

Important Note: these recommendations are made from experience, not sponsorship.

I Picked One. Now What?

This is where people get tripped up. It's not enough to just have the tool; you need to know how to use it effectively.

Model Management 101 - Speed vs. Brainpower

Think of your chosen AI like a toolbox with different wrenches.

You wouldn't use a tiny Allen key to fix a car engine, right? The same goes for AI models. Each system offers different tiers:

• Fast models (Claude Sonnet, GPT-4o, Gemini Flash) - these are your quick-chat, brainstorm buddies. Good for a rapid fire of ideas or answering simple questions.
• Powerful models (Claude Opus, o3, Gemini Pro) - this is where you go for the heavy lifting. Analysis, writing, research, coding – if it's high-stakes, use these.
• Ultra-powerful models (o3-pro) - for those truly gnarly problems that need serious processing power (and time – we're talking 20+ minutes of "thinking").

Important note: most AI systems default to the fast model to save on computing power.

So, if you're doing anything important, you need to manually switch to the powerful model using the dropdown selector.

If you don't see these options, you're likely on a free version, which often limits access to the most robust models.

So, fork over the $20 if you want the real deal!

I lean on Gemini 2.5 Pro for my serious work.

Deep Research: Your New Secret Weapon

This is the most underutilized feature, and it's a game-changer.

Deep Research tools spit out high-quality reports that even seasoned professionals (HR, lawyers, accountants, consultants) find impressive.

But here are some other fun ways to use it:

• Internal communications - use and prompt these tools to provide key communication documents based on personas (leaders, employees or executives)
• Travel Guides - "I'm going to Wisconsin and want a guide focused on unique cheese-related sites."
• Second Opinions - recently, I used Gemini to analyze images from a CT scan and the accompanying report to better understand the medical report, well before the doctor called to inform me. (covering the sludge in the medical industry here soon).

AI models are becoming surprisingly adept at diagnosis, with low hallucination rates, offering an interesting second perspective.

To activate this, simply turn on the web search option in Claude and o3 for a mini-research session. Google's Gemini offers additional features, allowing you to transform reports into infographics, quizzes, or even podcasts.

The Magic of Voice Mode

Seriously, if you're not using voice mode, you're missing out on the easiest way to interact with AI.

I wrote about how voice mode can be a thought partner ​here​.

Gemini and ChatGPT have the best implementations, allowing you to have natural conversations while driving, walking, or simply being too lazy to type.

Claude's voice mode is weaker than the other two.

But here's the real killer feature: screen and camera sharing.

Point your phone at a broken gadget, a complex math problem, a recipe you're following, or a sign in a foreign language.

The AI sees what you see and responds in real-time.

I've used it to identify dogs, trees, troubleshoot computer issues, and even get advice on fixing broken appliances.

This is the futuristic stuff, folks, not just talking to Siri.

Let the AI Create: Images, Video, and More

Want an image of an otter piloting a tiny 747?

Just ask! ChatGPT and Gemini can whip up images for you (Claude cannot).

(ChatGPT's image tool is generally more controllable, while Gemini offers some cool multimodal options).

And if video is your jam, Gemini's Veo 3 is seriously impressive, and you get some free uses daily.

These systems can also churn out documents, statistical analyses, interactive tools, simulations, and even simple games.

Simply select the "Canvas" option in Gemini or ChatGPT to obtain these structured outputs, or ask Claude – it's quite capable on its own.

So, what tool will you choose? Let me know.