There’s just no way I’m going to tip a robot.
AI how-to: 10 prompts to try with ChatGPT or your favorite bot
© Korakot Taechakan | Dreamstime.com, ChatGPT.com
Twenty-five-odd years ago, CBS and ABC execs told me tech was just a fad and passed on my national radio show. So, I syndicated it myself. The rest is history.
Now, I hear the same thing about AI: It’s a phase, and it won’t last. Hate to break it to you, but AI is here to stay, and it’s already changing everything. Start using it now or risk getting left behind.
So, what chatbot should you use?
There are so many options, but OpenAI’s ChatGPT is the most reliable I’ve found. I pay $20 a month for Pro, but a free account is fine for most folks. (Call me, Sam Altman. I’d love a freebie.)
If you want to use Chinese-built DeepSeek’s snazzy new R1 model, try it through Perplexity. It’s hosted in the U.S. and doesn’t have all the Communist China guardrails you find on the DeepSeek site.
The big caveat: You only get five free R1 searches per day. For unlimited, upgrade to Perplexity Pro for $20. Or use it for free and you’ll just get answers from a different AI model.
Now, what should you ask? The world is your oyster.
10 prompts to save and try later
1. “How can I make this better?”
Then, add in anything you’ve written. This prompt works wonders for polishing your work and pointing out improvements. Make it better by adding some context, such as “for my team or customers.”
2. “You’re an English professor. Find any typos or grammatical errors in my blog post.”
Change up the persona you’re asking it to assume and the specific instructions based on what you need. You can use this formula for just about anything. Try “You’re a life coach. Help me create a plan to achieve my goal of meditating every day for a month.”
3. “How do you know?”
This one is powerful if you’re not sure AI made something up or you just want to know the original source. Try it next time you get a suspicious answer to force AI to reveal its sources … or admit when it’s guessing. I often add to that, “I think your numbers are wrong. Double-check everything.”
4. “Remember that …”
In ChatGPT, go to Settings > Personalization and toggle on Memory. The same thing is available on Google Gemini for Advanced (paid) subscribers. Tell it something like, “Remember, I’m a tea drinker, not a coffee drinker,” and it’ll save that info.
DeepSeek told me some weird things
chat.deepseek.com
DeepSeek, the blockbuster AI chatbot from Communist China, told me that 95% of global internet traffic flows through undersea cables. That wasn’t surprising, but what came next was:
This is from the state-approved chatbot that’s gone viral for blocking anything Communist China deems inappropriate.
What are you tackling in 2025? Even if all you have is a vague idea, AI can help you with an action plan. Here are a few prompts to try with ChatGPT, Gemini or Claude.
- “I have a goal for 2025 to [fill in the blank]. Can you help me make it SMART?” (SMART is an acronym for Specific, Measurable, Assignable, Realistic and Time-related.)
- “You’re a life coach who wants to help me have my best year yet. Help me figure out some goals for 2025. Ask me questions one at a time to help me decide what to focus on.”
- “My goal is [fill in the blank]. What are some obstacles that might come up and ways I can overcome them? Give me specific examples for each obstacle.”
- “I want to [fill in the blank], but it feels overwhelming, and I don’t know where to start. Can you help me by breaking it down into more manageable tasks?”
⭐ AI isn’t going anywhere: Instead of sticking your head in the sand, make this the year you embrace AI. You’ll get daily tips here in this newsletter and, coming soon, my AI 101 guide for total newbies.
In the meantime, give this list of AI trends for 2025 a read and grab NetSuite’s free knowledge drop, “The CFO’s Guide to AI and Machine Learning.”* You’ll sound smart (and know what you’re talking about!) when it comes up in conversation.
Ship happens: Google’s shiny new AI assistant, Mariner, will take over Chrome to do stuff like book flights, fill out forms and do your shopping without you doing a thing. It can’t fill out credit card or billing info “for privacy reasons.” (Google actually had the gall to say that!) Watch the prototype in action here.
“Find the weaknesses in your last response”: Try that next time you’re feeling iffy about an AI chatbot’s response. It’ll reanalyze and look for any major holes or biases. Finish up with, “Try again, but this time …” based on whatever it told you.
Looking for a job? How to get past ATS and hiring AI
I was talking with Joey, our new podcast producer, and he mentioned the tricks I pass along to get past the AI-powered applicant tracking systems (ATS) help a ton of people.
Humans no longer filter resumes at most companies. It’s inefficient when AI can review over 100 resumes per minute.
ChatGPT-4o can see everything
ChatGPT-4o dropped some of its newest features that are free to everyone! The odds that they charge for them in the future is high. They want to get you hooked now.
Be the smart one at work: Download NetSuite’s free knowledge e-book, “The CFO’s Guide to AI and Machine Learning.” No matter what you do, you should know more about AI. It’s not going anywhere.
Try again: If ChatGPT totally swings and misses responding to your question, give it another go with a model that “thinks” differently. Click the switch icon. GPT-4o is my go to, even though o3-mini is newer.
🤖 Try Apple Intelligence first: Test some of Apple’s AI features with the iOS 18.1 beta on an iPhone 15 Pro or 15 Pro Max. Go to Settings > General > Software Update > Beta Updates > iOS 18 Developer Beta > Back. Then select Upgrade to iOS 18 and hit Update Now.
Bowser, meet your new robot friend
OpenAI’s ChatGPT-4o can “think,” write and talk in real-time. Check out this video to see it in action. Listen to that laugh! Crazy stuff.
Is this the year AI totally changes music? Take Udio, founded by ex-Google employees and sitting on $10 million in funding. They’re cranking out tracks at insane speeds — 10 songs a second, or roughly 864,000 songs a day. If you want to try it, it’s open to the public, and you can create 1,200 free songs each month.
Why AI is so bad at creating hands
AI can write essays, diagnose illnesses and even develop software. But there’s one thing AI can’t seem to get right: Human hands. Ask an AI art program to create a photo of hands, and the results are usually pretty darn weird. Why is this?
Deepfake Tom Cruise
Deepfake videos are all over the web and this one of Tom Cruise is particularly convincing. You won’t believe your eyes. It looks just like the Top Gun star!
How to tell if a video is made by AI
Look for unnatural body language and strange shadows, blurs or light flickers. And listen closely for choppy sentences or weird word inflection.
You won't believe how prolific AI art can be
Artwork created by artificial intelligence (AI) has taken off over the last year. It can generate virtually any image you can think of. But it doesn’t always match exactly what they were picturing when artists make a request. One artist took it upon himself to change that by using his artistic skills to work with AI and create some incredibly stunning images.
Will AI eventually replace greenscreens?
People have been using greenscreen technology for a long time. Meteorologists use it to explain the weather, movie directors use it to create incredible scenes, and online influencers use it to make fun videos. But now that artificial intelligence is becoming more advanced, some wonder if greenscreens will become obsolete.
Warren Buffett’s take on AI, upload dead people to a computer & our guide to earning $ online
Plus, it’s garden season! I’ve got the best smart sprinklers to keep your lawn green. But that’s not all – Matt’s got a genius email trick to track down data-selling companies. And if you’re sharing a web doc, don’t miss Allie’s tip to avoid embarrassing edits. Also, I tell the tale of an unbelievable Jeep underwater rescue. And can you believe “The Simpsons” predict the future? Matt shares some bone-chilling moments that prove they might.
Why AI doesn't speak every language
When you ask ChatGPT a question, it’s good at quickly returning a valid answer. But what if you speak a language that isn’t as common as English, Spanish, Chinese or French? Well, the AI model doesn’t have quite the same abilities. So why is this, and how will it improve in the future?