Did AI Kill Software?

Software stocks got pummeled last week. The S&P 500 Software and Services Index declined 8%, dragging the Nasdaq 100 into its biggest three-day slide since last April. Though stocks recovered on Friday, Salesforce, ServiceNow, and Atlassian are still down 25%, 30%, 40%, respectively, so far this year.

The sell-off was triggered by a wave of new AI tools that spooked investors. Anthropic rolled out industry-specific products for legal, finance, and customer service. OpenAI launched a more powerful, multi-agent version of its coding tool and a new platform to help companies deploy AI agents. 

To top it off, OpenClaw, a new, open-source AI assistant, exploded on social media. Self-dubbed “the AI that actually does things,” it can execute tasks on its own and remember past interactions, allowing it to learn user preferences and perform personalized actions.

The sell-off reflects growing concern that AI agents could disintermediate the entire software-as-a-service industry. If companies can build their own software with AI, why would they keep paying for traditional SaaS products?

There are signs that software company revenues are already softening. So far this earnings season, just 71% of software companies in the S&P 500 have beaten revenue expectations. That compares with 85% for the overall tech sector.

Robert Armstrong, the Financial Times’ U.S. financial commentator and frequent Markets guest, observed that last week’s carnage was part of a structural shift out of the tech sector and into lower-growth, more defensive categories. While the S&P 500 is down almost 1% so far this year, the equal weight S&P 500 is up nearly 4%, and “unsexy” sectors like household and personal care, discount stores, and construction machinery are up 10%, 16%, and 24%, respectively. 

This is panic selling. It’s very similar to what happened when ChatGPT came onto the scene. Everyone decided that search was dead, and Google stock cratered. But since then, Google has risen 280%. They learned from ChatGPT, they invested in AI, and their search revenue is up 50% since ChatGPT was released. 

The same thing is happening here. Software isn’t dead. Nothing is stopping these companies from integrating AI into their existing products. 

The other point that people aren’t recognizing is that switching costs are huge for enterprises. It takes on average more than half a year to find a new service provider. Executives need to sign off on the decision. It also costs money. If you’re going to terminate the contract, then you can expect to pay 100% of the remaining fees. 

The other thing they’re counting out is trust. They don’t have those same relationships with the newer startups. As the old saying goes, “Nobody ever got fired for going with IBM” — in this case, it’s Salesforce or Adobe.

But everyone started selling because they’re nervous, and they can’t tell what’s going to happen in the future. Markets hate uncertainty. That’s when you want to dive in.

You can read more about the opportunities I see in the market after the software sell-off in my latest edition of Simply Put, out tomorrow. Subscribe here.

It was a busy week in the entertainment industry. Disney finally named its next CEO: Josh D’Amaro, the head of its Experiences division. He’ll take over from Bob Iger next month.

The announcement came on the heels of Disney’s Q4 earnings report. Disney beat expectations and reported record revenues in its Experiences division. Streaming had a strong quarter too, with profits surging more than 70%. But the stock still dropped as much as 7% on concerns about softening international tourism, ballooning sports rights costs, and continued weakness in traditional entertainment.

Meanwhile, Netflix and Warner Bros. Discovery executives were on Capitol Hill defending their planned merger. What started as an antitrust hearing quickly devolved into a political circus. Sen. Eric Schmitt accused Netflix of producing “the wokest content in the history of the world.”

Every Disney earnings call follows the same playbook. Parks? Unbelievable business with an incredible moat. Film studio? Doing well, creating IP that feeds into the parks. Disney streaming? Getting leverage with singular positioning around family. All of it’s great. 

Then they have to apologize for their linear platform, which declined 12% last year. 

The moment they shed the linear TV business and focus entirely on parks, experiences, cruises, streaming, and the studio, this company will go up substantially in value. Right now, the overhang of these linear networks creates confusion around the business strategy and forces them to spend the last 20 minutes of every earnings call apologizing.

The antitrust hearing was completely ridiculous. 

These Republican senators default to talking about “wokeness” because they don’t know where to land on the antitrust argument. 

The hearing was supposed to address what happens when an extremely large entertainment company buys another one, creating legitimate monopoly concerns.But that’s boring. They don’t understand it. They’d rather talk about LGBTQ actors in children’s shows because that’s what gets their supporters riled up.

Will any of this affect the transaction or change monopolization in America? No. But it doesn’t matter. It’s all for the clips, the drama, and the show.

Anthropic used the Super Bowl to take a shot at OpenAI, and it worked. The company dropped a series of ads early; one, titled “Violation,” pulled in 5 million views on X in under 48 hours. The tagline: “Ads are coming to AI. But not to Claude.”

It was a direct jab at OpenAI, which recently announced it would begin testing ads on ChatGPT. That news caused a stir online, given that Sam Altman once called ads “a last resort for us as a business model.” 

The irony is that AI companies have been pouring money into traditional advertising while publicly positioning themselves as skeptical of the ad business model. Last year Anthropic, Google, Microsoft, OpenAI, and Perplexity collectively spent more than $330 million on broadcast TV ads for their AI offerings, up 43% from the year before.

They’ve still got a ways to go. Only 17% of American adults think AI’s impact on the country will be positive over the next two decades. 

Prof G Markets wrote about Anthropic’s excellent marketing back in October. Check it out here.

 Sam Altman put out a pathetic, 400-word tweet trying to argue that the ad isn’t accurate, that they’re not going to surreptitiously insert ads while you’re doing therapy. Maybe that’s true. But no one cares.

Anthropic’s ad was good. They nailed the voice of ChatGPT — that awkward, monotonous, sycophantic voice everyone has gotten sick of. They have their finger on the pulse.

To me this ad has parallels to the 1984 Apple ad, regarded as one of the greatest commercials of all time. It aired during the Super Bowl and took a direct shot at IBM, suggesting they were Big Brother trying to control the world while Apple would liberate you. Was it all technically true? Not exactly. But it resonated. That’s exactly what Anthropic has done.

I said a few months ago that Anthropic’s going to be the heavyweight champion of AI in 2026. This is the moment everyone realized it’s actually happening.

This is the moment that Anthropic begins its valuation ascent past OpenAI.

In my course on brand strategy, I teach something called laddering. You say we are this and the competition is that. You find points that pass three hurdles: differentiation, relevance, and sustainability. 

Anthropic’s differentiation is clear: no ads. It’s relevant because the number one use case for AI is therapy. People are revealing their most intimate questions and concerns. The thought that OpenAI is going to take all your personal information and start saying, “You seem to be suffering from depression; have you thought about Lexapro?” It’s frightening.

I’m totally blown away by Anthropic’s execution. They registered 10x the ROI on this Super Bowl commercial before it even ran. 

Altman’s response was a mistake. When you’re the market leader, you don’t reference the competition. You never talk about them. Hertz never referenced Avis. Coke never referenced Pepsi. If he’d been asked about it, he should have said, “I’ve seen it. It’s a great ad. Good for them.” Anything beyond that looks defensive. 

This is a seminal moment. When we look back on this, this will be seen as a big moment in the AI wars.

Elon Musk has broken four wealth records in the past four months. Specifically, he is the first person to be worth more than $500 billion, $600 billion, $700 billion, and now $800 billion. 

Though his fortune is mostly illiquid, if he could convert it to cash, his current net worth would be enough to

Yet, if his tweets are to be believed, he’s miserable.

Is Elon right? Is it true that money can’t buy happiness? A much-cited study from Daniel Kahneman and Angus Deaton in 2010 found that happiness increased up to a point — $75,000 per year — and then leveled off. Then, a study in 2021 found that more money did increase happiness with no plateau. 

But the most recent research, a collaboration between both studies’ authors, found that the reality is more complicated. 

Money increases happiness for everyone up to $100,000 per year. 

But past the $100,000 threshold, the effect changes. For people who are persistently unhappy, earning more money doesn’t make them happier. In the words of the paper’s authors: “The suffering of the unhappy group diminishes as income increases up to 100k but very little beyond that.”

However, among people who already experience a lot of positive emotion, additional income continues to increase happiness, even past $100,000, and it does so faster than it did at lower income levels.

The chart below helps clarify.

The different lines represent people grouped by their average happiness level. The people in the 85th percentile are more consistently happy than those in the 15th percentile. 

The slope of the lines represents change in happiness. The yellow lines from 0 to $100,000 show that for all groups, happiness increases until $100,000. 

The green lines show the change in happiness for incomes higher than $100,000. For the folks in the least-happy group, the 15th percentile, the rate of increase in happiness (green line) is basically flat — meaning their increase in happiness is statistically insignificant past $100,000.

Small Influences and Big Influences 

Money impacts happiness and even increases it substantially for some people. However, as the authors of the study noted, “The relationship is weak, even if statistically robust.” Not even $850 billion will buy happiness for someone who is persistently unhappy. 

If money can’t reliably buy happiness, what can? The longest scientific study of happiness ever conducted tells us the answer: relationships. All relationships matter: friends, acquaintances, family, coworkers, and neighbors. The stronger our relationships are with others, the happier we are, and the longer we live.

At the End

Musk’s tweet about happiness was seen by over 100 million people and received over 100,000 comments in 48 hours. 

There's something ironic and deeply sad about the most well-resourced man in the world confessing his unhappiness to 100 million IP addresses, when we know that the key to happiness is real relationships. If only he would get off his phone.

In 12 months, Anthropic is going to be worth more than OpenAI.

Journalist Derek Thompson joins Scott to examine why Americans feel increasingly unhappy — even as many measures of health, safety, and quality of life improve.

Prof G Recommends: Stat Significant is a free weekly newsletter featuring data-centric essays about movies, music, TV, and more. When do we stop finding new music? Which TV shows got their finale right, and which didn’t? Which movies popularized (or tarnished) baby names? Subscribe to find out.