“Keep Thinking” was Great, is OpenAI Thinking About Marketing at All?

A little over a month ago, Anthropic, the maker of Claude, started running their “Keep Thinking” ad. Personally, it’s the best ad I’ve seen in a long time (for those of us who keep track of that sort of thing). I remember seeing it in bits and pieces as it started popping up around Twitter and I noticed it on a commercial or two, but it was when I went to see One Battle After Another in IMAX that I got fully hooked.

The beautifully shot, colored, edited, everythinged ad in 4:3 (the superior aspect ratio), looked even better and felt even more impactful on that huge screen. It was so effective that I watched the trailers, 3 hour movie, drove home, and still remembered the ad and downloaded Claude that night.

The ad has already gotten plenty of praise, deservingly so, but what I haven’t seen explored as much is OpenAI’s very poor response attempt, how it highlights the larger issue they seem to have with marketing, and how the two brands appear to be positioning themselves in a way that has a lot of similarities to two other companies from the days of the computer boom.

Visually Correct, Emotionally Soulless

The Keep Thinking ad worked so well and connected with so many people because it did something we hadn’t really seen an AI company try to do yet, it positioned Claude as a true helper to people. Not something to handle mundane work, or plan a vacation, an actual helper.

Throughout the ad you see people working on a problem. Classic pacing around, laying on the floor staring at the ceiling, paper and tools spread all over the place, actually trying to solve something, and while you see Claude in all these scenes, it’s always secondary. Someone’s brainstorming with it, someone’s using it to run calculations, someone else to research, you get it. Claude is a character, but the main characters are the people trying to solve the problem, all with the undertone of hope for the future.

Meanwhile, in the series of responses OpenAI has attempted, take the “Road Trip” ad for instance, ChatGPT is positioned as a replacer.

The ad looks similar to the Claude ad. Natural colors, people, warm-looking, but the actual emotion of the ad is cold. It opens on a brother and sister in a car on what’s revealed to be a Blue Ridge Mountains trip, planned, by ChatGPT… The prompt, and then ChatGPT’s response, literally takes over the humans interacting in its plain white text and starts scrolling with the full plan for the trip as the camera shifts to the car driving off in the distance as the OpenAI and ChatGPT logo roll.

The full plan. It doesn’t show a discussion, interaction, brainstorming session, or conversation between the brother and ChatGPT. It doesn’t flashback to him using the product and having a dialogue to help plan the trip. All the helping and hummanity shown in the Claude ad is missing, and instead, the sell is “Let ChatGPT do everything for you.”

The other variations of this ad are equally soulless and have the same core message of “Well why would you want to think? We’ll handle that.” With the ChatGPT response always covering the person basically acting on the plan ChatGPT gave them. It’s not collaborative, it’s instructive, and it’s not assistance, it’s replacement.

Back to Marketing 101

It wouldn’t shock me at all if OpenAI had ChatGPT/Sora (their video generation model) write, direct, and compose these ads. It’d actually surprise me if a real human touched these ads at all because there’s zero signs of it in the ads.

Selling to businesses is one thing; selling to consumers is another, and something OpenAI seems to be really bad at. This ad campaign missed what is a classic Marketing 101 lesson I find myself quoting all the time, the Betty Crocker Instant Cake Mix example.

The story goes that in the late 1940s General Mills launched their Betty Crocker Instant Cake Mix where all you had to do was just add water. Despite the convenience, the mix didn’t sell very well, because it was too easy. The theory was that people felt guilty using a mix so simple because there wasn’t any work required to make it. The solution that was landed on was to reformulate the mix to require a fresh egg in the recipe instead of the powdered eggs that were already included in the mix, so that the people making the cake felt like they were actually taking part in the creation process.

The truth of that story is is iffy, but I think the core lesson is still important. Consumers don’t want to feel completely replaced, they want convenience, but there still needs to be friction.

Anthropic Looks Like Apple, OpenAI Looks Like Microsoft

The two different tones that Anthropic and OpenAI are striking reminds me a lot of Apple and Microsoft. Both companies passed the $4 trillion mark this week, so they’re both doing fine, but they are very different companies selling very similar products, but one is much more consumer, or individual, focused while the other is aiming for organizations.

Apple put their own spin on the Keep Thinking campaign a couple of weeks ago with their “Great Ideas Start on Mac” ad. It’s no 1984 or the memorable iPod ad, but it’s still amazing.

The ad opens, and stays, on a Mac on a desk looking out a window with a blank document open and the cursor flashing. Someone sits down at the desk while the narration by the late Dr. Jane Goodall starts.

“Every story you love. Every invention that moves you. Every idea you wished was yours. All began as nothing.” Dr. Goodall says as the camera zooms in more and more on the flashing cursor. It’s here you really notice that the kick drum in the music is aligned with the cursor’s flashing and sounds, and is paced, a lot like a heartbeat. “Just a flicker on a screen. Asking a simple question, what do you see?” Then a collage of people working on a Mac in every creative situation imaginable rolls ending with “Great Ideas Start Here” typed out.

That ad highlights perfectly who Apple is for, and the angle I think Anthropic is going for with its Keep Thinking campaign. It makes you imagine, it then asks you “What do you see?”, finally it rolls on people using a Mac to create their vision, offering a tool to do it with, but not to do it for you.

Now try for a minute to remember a Microsoft ad. Could you? I couldn’t. Have you ever actually wanted a Microsoft product, or was it kind of just there? Has the design of anything Microsoft has ever done ever inspired something for you? Microsoft, funny enough a large shareholder in OpenAI, doesn’t really have much of a personality to speak of. It does a good job of existing and getting out of your way (kind of), but their physical products and their software aren’t exactly awe inspiring. You don’t want to use them, you just kind of have to along with everyone else you work with.

Microsoft makes products that work well for a business in a business setting for people who need to do normal, mundane, business tasks. Clearly there’s a large ($4 trillion+ apparently) market for that, and I think this is more the lane OpenAI is trying to exist in anyways, but at a time where the battle for the go-to consumer, or individual, AI platform is still ongoing, OpenAI’s bland, corporate, lifeless style isn’t going to help them.

One Miss After Another

Have you ever subjected yourself to an OpenAI product demo? You haven’t missed much if not. I remember watching the GPT-5 announcement and struggling to engage. The product, and what they had achieved in general, was incredible, but the people delivering the message, and the way they were delivering it, were not doing it well.

The point of this product demo was to show off what was supposed to be the next great step in AI. Meanwhile, all of the people who talked during the demo either did no prep at all, or don’t have the public speaking skills required to make something like GPT-5 feel as magical as it needed to feel. They seemed to be trying to deliver a highly-technical presentation at the same time as something that the general public would understand. They did not do a great job of that.

That’s without mentioning the several times in the demo where charts in slides (clearly/hopefully? generated by ChatGPT) were incoherent, labeled in a way that made no sense, and I guess were never looked over by anyone, because who would internally use a chart this bad, much less in your most important product demo to date.

How did this possibly make a presentation?

Then there was the Sora 2 launch video, which, while it did show off the product’s impressive abilities, was also just flat-out lazy. I’m not sure that wasn’t the point though. Another term for lazy, at least in this case, is cost-effective. I’m sure businesses saw the launch video and went “Well this great. I don’t need people creating videos for me anymore.”, but the larger general public’s response to Sora 2 is summed up well in the term “SlopTok”.

For all I know, OpenAI has no desire to compete in the consumer AI market, and the goal of their more public-facing advertising is to be palatable enough for the average person that they won’t flat out refuse to use an OpenAI product when it’s pre-installed on their boring, work-provided, Microsoft laptop. If that’s the case, they’re doing a decent enough job, but if OpenAI wants to be people’s go-to AI platform outside of work, they’re missing the mark about as badly as their chart suggesting 52 is larger than 69 which is equal to 30.