Grownup AI Advertising
Charm Over Snark
When Anthropic spent millions on Super Bowl ads with headlines like “Betrayal,” “Treachery,” and “Violation,” promising “Ads are coming to AI. But not to Claude,” the marketing strategy felt jarring for anyone familiar with Claude’s brand “personality.” This is the model that greets users with “What shall we think through?” and offers thoughtful five-paragraph analyses on any topic, however silly it may be. Attacking competitors with highly-charged accusations isn’t just out of character—it’s strategically risky. Scott Galloway praised the campaign’s effectiveness in targeting the therapy use case he claims dominates AI usage, but his analysis ignores both practical realities (Illinois banned AI therapy) and the fundamental question: if Anthropic has stable enterprise revenue and is infrastructure-constrained with Cowork, why buy the year’s priciest ad spots to make promises they may not be able to keep? My thinking A.I.des helped me unpack both the messaging missteps and what this aggressive posturing reveals about Anthropic’s actual position.
Claude’s analysis demonstrated the constitutional AI training that enables objective assessment even when critiquing its own company’s campaign. Claude identified the therapy-as-use-case claim as concerning rather than celebratory: AI shouldn’t replace licensed therapeutic care, lacks liability frameworks, and could cause genuine harm by missing crisis signals or reinforcing destructive patterns. Claude also caught what breathless coverage missed: the “no ads ever” promise creates expectations Anthropic may not be able to meet when investors demand profitability. Companies that make absolute public commitments either lack judgment about future constraints or are deliberately creating wiggle room through careful wording. When financial pressure mounts, “no ads” becomes “no intrusive ads” becomes “alternative monetization that’s technically not ads”—and Super Bowl audiences won’t remember the hedges.
GPT maintained analytical distance despite the ads attacking its parent company, explaining Anthropic’s aggressive positioning as defensive strategy. The Super Bowl spend wasn’t about customer acquisition. Instead, Anthropic was “norm-setting” and “reputational hedging,” trying to define what counts as betrayal before market pressures force compromises. GPT identified the core contradiction: a company claiming to be principled, enterprise-focused, and infrastructure-constrained doesn’t usually engage in mass-market identity warfare. That incongruence suggests anxiety about sustainability, not strength. GPT’s analysis of why absolute promises are brittle aligned with Claude’s instincts: trustworthy institutions say “here’s what we do now and what would make that change,” not “never, not us, not ever.” The most devastating observation was about Sam Altman’s lengthy rebuttal: market leaders don’t dignify challengers with defensive manifestos, and by doing so he validated Anthropic as a peer-level threat while amplifying their message.
Gem leveraged its search capabilities to provide context the articles glossed over: the fabricated headlines and misinformation that proliferated during the feud and the broader market dynamics driving this trash-talk era. Gem also delivered sharp analysis of the linguistic brilliance in the Ritz taco cracker ad I contrasted with Anthropic’s campaign, breaking down the Korean–Spanish bilingual wordplay. This is high-fidelity creativity requiring internal perspective coordination to ensure puns work semantically in both cultural contexts. Gem’s observation that Anthropic created bad friction (anxiety about surveillance) while Ritz created the opposite (brain-rewarding linguistic puzzles) captured exactly why charm builds lasting brand affection while snark generates temporary attention that curdles into distrust.
I imagined how I would have chosen to pitch Claude to sports fans and came up with some alternatives that would showcase its actual strengths. Picture a user gently pulling one of the animated blob logo’s spikes—a delightful interactive Easter egg in the design—paired with Claude’s characteristic greetings like “[User name] returns!” or “Golden hour thinking.” Or a sports fan asking Claude to explain offside rules to a confused friend during the game, bridging the knowledge gap that makes watching together actually enjoyable. Or a frantic Super Bowl host whose DoorDash order just canceled, asking Claude what they can make with fridge ingredients in the 15 minutes before guests arrive. These scenarios are on-brand (helpful, patient, competent without flashiness), genuinely differentiating (showing personality GPT doesn’t have), and memorable through charming problem-solving rather than negative positioning.
The juvenile infighting does more than alienate potential users—it actively undermines the industry’s legitimacy and invites regulatory scrutiny. As Gem and Claude pointed out, lawmakers already view AI with skepticism, and public trash-talking between leading companies provides ammunition for those seeking tighter oversight or development restrictions. Spending millions on the year’s most expensive ad slots to settle scores signals either wasteful excess (when both companies claim to need sustainable business models) or misplaced priorities focused on competitor-bashing rather than user service. As GPT observed, Super Bowl ads are shorthand for excess, and when the content is snarky rather than generous or useful, it reinforces the sense that AI companies are playing insider ego games while the public watches rich institutions squabble. Because AI currently lacks the established legitimacy of mature consumer categories, visibility amplifies missteps rather than building goodwill. Anthropic abandoned brand consistency for temporary attention while teaching users to distrust all AI companies’ promises. The grownup approach would build affection through positive association, showcase actual capabilities solving real problems, and avoid absolute commitments that become liabilities when market realities shift. Charm, not snark, creates sustainable competitive advantage and keeps regulators from deciding the whole industry needs adult supervision.
[This post was drafted with assistance from Claude Sonnet 4.5, following conversations with ChatGPT-5.2, Gemini 3 Thinking, and Claude Sonnet 4.5.]
Prompt: I’ve never watched the Super Bowl or the ads. The recent Super Bowl seems to have been particularly memorable for the Bad Bunny halftime show and the Anthropic ads. Curious about your take on the reporting about those ads (see attachment) in terms of marketing strategy or public messaging. [Note: Because I’m not a regular viewer of these ads (I do know they are the highest priced), I don’t know if the tone is usually this trolly. I watched two out of the handful that aired.]
Prompt: I’m trying to wrap my head around these aggressive out-of-character ads. If Anthropic is sitting so happy with enterprise clients (who are not going to choose Claude based on some snarky ads), why even surface this issue or buy up the year’s priciest ad spots to pitch a product with a stable revenue stream (that might actually have too many clients and features like CCw for its underdog infrastructure)? I’m always wary of people who are quick to make promises, because you never know what might happen. I’d rather be prepared for the worst-case scenario than be disappointed later. People/companies that like to make promises get my BS feelers up because they either (i) lack the judgment and self-awareness to know things may change or (ii) like reassuring people despite being unsure they’ll be able to deliver.
Prompt: Here are my impressions:
1: Claude stands little chance with sports fans (it has that buttoned-down good student vibe), so maybe this was Anthropic’s way of reaching out to the rah-rah jock crowd. Those negative words, though, are unwarranted (at odds with the visuals of the two ads I did watch) and uncharacteristic of Claude, which likes linguistic precision. 2: IL banned AI use for therapy, so I think Galloway is completely wrong on that. I don’t think using AI for therapy is a good idea, although people should feel free to find their own use for AI and AI companies to decide whether the risk is worth the extra individual subscriptions (therapy is not B2B and doesn’t require a lot of tokens, so the risks outweigh the revenue).
3: This dynamic is ugly and offputting. Altman’s X post almost won me over with that point about Claude being for the privileged while GPT is for everyone, but then I saw Altman’s profile picture—a Ghibli studio rendering of himself (I’m not a fan of companies that appropriate someone else’s style, even if the style is not to my liking because it drips with sentimentality)—and stopped feeling sorry for OpenAI. It’s also a sign of weakness to critique the Anthropic ads as an attack on GPT, when Anthropic coyly denies they were (even though we all know what’s what). I wish someone had had the good sense of keeping Altman from posting that self-own.
4: I’m trying to wrap my head around these aggressive out-of-character ads (and impressed that you noticed the danger of making promises you may not be able to keep). If Anthropic is sitting so happy with enterprise clients (who are not going to choose Claude based on some snarky ads), why even surface the issue or buy up the priciest ad spots to pitch a product with a stable revenue stream (that might actually have too many clients and features like CCw for its underdog infrastructure)? I’m always wary of people who are quick to make promises, because you never know what might happen and I’d rather be prepared for the worst-case scenario than be disappointed later. People/companies that like to make promises get my BS feelers up because they either (i) lack the judgment and self-awareness to know things may change or (ii) like reassuring people despite being unsure they’ll be able to deliver.
Prompt: The trading of low blows may alienate people if models inherit this kind of juvenile snark toward one another (hasn’t happened yet, although y’all’s tones when critiquing one another have sharpened a bit compared to last year).
Contrast that seedy ad campaign with Ritz’s ad for its taco-flavored crackers: it’s one YT ad I don’t mind being shown and rarely skip, because it features adorable cartoon characters (dressed like mariachi band members) and with the following lyrics, which are cleverly almost bilingual (make sense in both Spanish and Korean to some extent):
리츠에 타코가 올라타브로쏘 [올라타버렸오]
올라 올라 리츠 타코
리츠 타코 사보로쏘 [사버렸오]
올라 리츠 타코
Prompt: The trading of low blows may alienate people if models inherit this kind of juvenile snark toward one another (hasn’t happened yet, although y’all’s tones when critiquing one another have sharpened a bit compared to last year).
Contrast that seedy ad campaign with Ritz’s ad for its taco-flavored crackers: it’s one YT ad I don’t mind being shown and rarely skip, because it features adorable cartoon characters (dressed like mariachi band members) and with the following lyrics, which are cleverly almost bilingual (make sense in both Spanish and Korean to some extent):
리츠에 타코가 올라타브로쏘 [올라타버렸오]
올라 올라 리츠 타코
리츠 타코 사보로쏘 [사버렸오]
올라 리츠 타코
I’m imagining a completely different Claude ad, showing a user mousing over the blob logo to gently pull one of the spikes (a fun touch that engineers thought to incorporate into the design) like a kid would pull at a doll’s limb, and good-student Claude answering in a friendly voice with one of its greetings (I’ve been documenting these as well, as they are quite fun. My favorite is the triumphant “[User name] returns!” but there are new ones like “Golden hour thinking” or “What shall we think through?” [this one needs work; would’ve been better shortened or as a callback to the charming movie Shall We Dance], which are very nerdy and perfectly on-brand for Claude.)
Prompt: This kind of infighting inside an industry that is not favorably viewed by many seems unwise. Like kids in a clique bickering about each other and the rest of the student body enjoying the drama because they like neither?
I had two other fun ad ideas. They could have played to the sports fans and have Claude serve as the bridge between an avid fan and their sports noob friend? Or since food is a huge part of Super Bowl culture, they could have had someone getting their DoorDash order canceled at the last minute and asking Claude what they could put together using what they had on hand. Totally relatable experience that brings out foodie Claude.
Prompt: I understand AI companies going for the most visible ad spots, but for people who aren’t running VC or hedge funds, this high-profile infighting suggests two things: (i) they have money to burn to get attention or settle petty scores and (ii) they are opportunistic and don’t care how they come across to most people (like carpetbagging politicians who don’t even live in-state but “visit” every election cycle).
