Quick Note: Why You Can't Count Out Google

Think long term, and be patient.

This is my personal journal about my learning journey in AI and ML for marketing. With over twenty years of experience managing and marketing tech companies, I have deep knowledge of marketing analytics and attribution, always focused on generating measurable and meaningful business results. For the last nine years, I've advised and served marketing leaders in tech companies, enabling me to share my insights on the impacts of AI and ML, including privacy and ethics issues, through this Substack. You can learn more about me, including my policy on using AI for this blog, in the "About" section of this Substack. I welcome your suggestions for further education on AI and ML.

In This Issue:

• Google (Bard) Nails It (at least for me, and at least right now)

• Playing With Midjourney Substack and Bing Images

• Celebrating Humanity

• What’s Next
  

Google(Bard) Nails It, At Least Right Now

Today, Google had its IO conference and announced a lot of stuff regarding AI. Everyone and their brother is writing furiously about it. And, as luck would have it, I’m at Mom’s right now for the first time since cancer, and I’m trying to stay present with her.

But I just had to write about this jump in capability, because it is solving a problem I’ve had for a few weeks now. As far as I know, no other chat bot can do what I’m about to show you, at least not from a graphic user interface. (As always, correct me if I’m wrong). I’m jamming this article out fast to get back to Mom, so please excuse any grammatical issues. This is how you know I write these articles, I guess! :D

My goal: get a chatbot to read a CURRENT article and summarize it for me. Sounds easy enough, but I spent literally hours with ChatGPT-4 trying to get it to do what I’m about to show you that Bard can do right now. Because ChatGPT-4 stops at 2021, it’s not great at taking new information in. You also can’t upload big files or do it in any predictable way.

Here’s what happened to me just a few weeks ago with Chat GPT 4. My goal was to get it to write a LinkedIn post that summarized the Forbes 50 Entrepreneurial Marketers article and called out some themes that were present in the article; some commonalities if you will.

ChatGPT4 did an admirable job of summarizing the document, and with a LOT of prompting actually got to a place where it asked me questions about what I wanted to actually say in the article. I was pretty excited. But when I asked it to use specific examples from the article, it listed marketers who weren’t even in the article!

I only discovered this when I went to post the article to LinkedIn and then realized I couldn’t tag the people listed because they didn’t hold the titles that ChatGPT said they did; and in some cases, they weren't even in the article itself!! ChatGPT kept putting in Fernando Marchado’s name. And while Fenando Merchado is a wonderful and entrepreneurial marketer, he simply isn’t on this particular list. I tried putting the article in a google doc to upload it. I tried copying and pasting the text (it was too long). Truly, I worked on this for hours, wondering where I had gone wrong.

As a result, I had to scrap the entire post, for fear of making some other major blunder.

Meanwhile, literally just now, while on my phone (on LTE at Mom’s, where I’m actually tethering to write this piece, please send sympathy,) I was able to get Google Bard to do an absolutely magnificent job of it in just 3 prompts. I could have done it in one long post, but it stuttered for a second and needed to be led along.

ChatGPT 4’s Output:

Here you can see the fantastic power of ChatGPT, but also its supreme failure and risk. This was after 2 hours of prompting.

Google/Bard’s Output, Tonight: Light Years Ahead

Here you can see Google Bard doing what I ask almost immediately, only my phone as I lazily sipped a vermouth on the rocks and "farted around" with it

A Few Thoughts about Google/Bard, and What’s Next:

1. Surprisingly, Bard was happier being directed to a website link than to my Google Doc, but I bet that will get fixed pretty quickly.

2. I did a bad prompt and it started writing posts about being a better Linkedin marketer. But I closed the app and started again.

3. The speed and quality of this tonight got me thinking: Google’s been crawling the internet for a hell of a lot longer than OpenAI. There are some things Google is going to naturally do better than anyone else for that simple reason alone. I’m not saying they’ll be the winner-anyone who suggests they know who that winner is full of it-but that all the hype around OpenAI doesn’t mean they win.

4. I continue to be so excited about the possibilities. No one can be counted out right now.

5. As always, the usual caveats apply: be smart and check your work, and don’t upload sensitive information to a chatbot until you understand the implications.

Celebrating Humanity: My Family

This tiny hoe was made by my great-grandfather, who ground down a full-size hoe for my great-grandmother. to make it easier for her to weed in between the plants in her flower beds without bending over so much. It’s been handed down through the generations. My mother is an expert gardener and has it sharpened every year. It absolutely murders dandelions. And someday, it will be mine.

Playing With Midjourney Bing and Substack Images

cardinal bird in spring time

Substack’s image generator is now working pretty well so long as I give it a simple prompt. It’s close enough. Here I’m celebrating the beautiful cardinals that show up at Mom’s house. Their song is as identifiable as their bright red plumage.

What’s Next

• Still on my trip, but as the Muse strikes I’ll be here.

• Please keep an eye on my “notes” in Substack for quick tips, links, and more. that’s what will help me keep these blogs from being so huge.

• What do you want me to drill into? what’s interesting to YOU? Let me know!