Sam Altman got exceptionally angry about Claude’s Super Bowl ads

Anthropic’s Super Bowl commercial, one of four the AI ​​lab dropped on Wednesday, begins with the word “betrayal” splashed in bold across the screen. The camera pans to a man seriously asking a chatbot (clearly intended to portray ChatGPT) for advice on how to talk to his mother.

The robot, portrayed by a blonde woman, offers some classic advice. Start listening. Try a nature walk! It then turns into an ad for a (hopefully!) fake cougar dating site called Golden Encounters. Anthropic concluded the point by saying that although ads will reach the AI, they will not reach its chatbot, Claude.

Another commercial shows a young man looking for advice on building abs. After displaying his height, age and weight, the robot presents him with an advertisement for height-enhancing insoles.

Human commercials are cleverly targeting OpenAI users, with that company recently announcing that ads are coming to ChatGPT’s free tier. It caused an instant buzz, spawning headlines of “mocking”, “skewering” and “dipping” on OpenAI.

They’re funny enough that even Sam Altman admitted on X that he laughed at them. But apparently he didn’t really find them funny. They inspired him to write a novel-sized rant that turned into calling his rival “dishonest” and “tyrannical.”

In this post, Altman explains that the ad-supported tier is intended to shoulder the burden of offering ChatGPT for free to many of its millions of users. ChatGPT is still the most popular chatbot by a wide margin.

But OpenAI’s CEO insisted the ads were “insincere” in implying that ChatGPT would skew the conversation to insert an ad (and perhaps for an off-color product, to boot). “It is clear that we will never run ads in the way Anthropic envisioned,” Altman wrote in a social media post. “We’re not stupid and we know our users will say no to this.”

In fact, OpenAI has promised that ads will be discrete, labeled, and will never impact chat. But the company also said it plans to make it conversational, the central claim of Anthropic’s ads. As OpenAI explained on its blog. “We plan to test ads below answers in ChatGPT when there is a related sponsored product or service based on your current conversation.”

Altman then went on to make some equally questionable assertions at his rival. “Anthropic offers an expensive product to the rich,” he wrote. “We also feel strongly that we need to bring AI to the billions of people who can’t afford subscriptions.”

But Cloud also has a free chat tier, with subscriptions of $0, $17, $100, and $200. ChatGPT levels are $0, $8, $20, and $200. One could argue that the subscription levels are fairly equivalent.

Altman also claimed in his post that “Anthropic wants to control what people do with AI.” He argues that he blocks the use of Claude Code from “companies they don’t like” like OpenAI, and said Anthropic tells people what they can and cannot use AI for.

It’s true that Anthropic’s entire marketing deal from day one has been “responsible AI.” The company was founded by two former OpenAI graduates, who claimed they were concerned about the safety of AI when they worked there.

However, both companies have usage policies, AI guardrails, and talk about AI safety. While OpenAI allows the use of ChatGPT for sexual arousal while Anthropic does not, OpenAI, like Anthropic, has decided that some content should be banned, especially related to mental health.

However, Altman took the anthropological “what should you do” argument to an extreme level when he accused anthropology of being “authoritarian.”

“One authoritarian company will not get there alone, not to mention the other obvious risks,” he wrote. “It is a dark road.”

The use of the word “tyrannical” in the rant about a Super Bowl ad is misplaced, at best. It’s especially tactless when considering the current geopolitical environment in which protesters around the world have been killed by agents of their own government. While commercial competitors have been competing in advertising since the beginning of time, Anthropy clearly struck a chord.