Marissa Mayer recently drew fire for suggesting an AI tool should help her write a school assignment for her daughter.
Sam Altman has proposed that parents record personalised podcasts for their children’s commute. He has also stated he cannot imagine parenting without ChatGPT. Mark Zuckerberg has suggested that meeting social and romantic needs through AI will become increasingly acceptable.
TechCrunch’s Amanda Silberling asked whether executives are too disconnected to build products for normal people. The question matters as Anthropic prepares for an IPO while Wall Street loses appetite for heavy AI spending.
Kirsten Korosec, Sean O’Kane, and I spoke with Silberling on the Equity podcast about why leaders sometimes seem tone deaf. She wondered if this is a marketing failure or a product failure.
Silberling said it is a little bit of both. She argued the industry is pushing a future where basic human interaction gets optimised out of actually occurring. She added that humans being able to talk to each other is probably a net positive for society.
The disconnect
Anthony Ha noted there is widespread societal scepticism about AI. People feel the people running these companies do not have their best interests at heart. He suggested executives should avoid making statements like Mayer’s because people view them with suspicion.
He thinks CEOs are trying to walk a delicate balance. There is a financial incentive to argue the technology should be used in every aspect of life. But making that case paints an unappealing picture. In some cases, it is downright dystopian.
Korosec said if they took Ha’s advice, there would be nothing to talk about on the show. She noted that AI making life easier in personal relationships often overcomplicates simple tasks like observing your child. This disconnect seems incredible when reading their quotes.
She asked who they are speaking to. Are there people out there who can relate?
Silberling pointed to other use cases. Bumble CEO Whitney Wolf Heard suggested AI concierges where an AI dates another person’s AI before introducing the humans. Mayer also mentioned her AI coordinating with her husband’s AI.
Silberling feels people are trying to over-optimize out of just having conversations with people. She has seen this happen in her own generation. Ordering takeout feels weird now if it involves calling a place. She is used to doing it online without talking to anyone.
She believes there will be more situations where basic human interaction gets more and more optimized out of actually occurring. She thinks humans being able to talk to each other is probably a net positive for society.
The business case
O’Kane asked about the return on investment. How much revenue can a company generate off of a conversation somebody is having in meatspace?
Silberling said dating is really inefficient. It sucks that you have to go on a bunch of dates with different people. You might waste an evening and be like, “Well, that was a bust.” That is why people wonder if their AI could go on dates for them. But there is just something not human there.
She also noted you are outsourcing the decision of who you are going to date to a computer you are trusting knows you better than yourself. She does not think that is true.
Who is the audience?
Korosec asked if this is a product problem or a messaging problem. Are they talking about it in the wrong way? Or is it actually a product problem where they are focusing on optimizing the wrong things with AI tools?
Silberling said it is a little bit of both. The marketing problem comes from high-up people at massively valuable companies talking about pain points in their lives. They focus on making dinner reservations or booking flights. Everybody does those things, but she is not booking flights once a week. She is not making dinner reservations all the time. When she does, those are not things she finds to be a particular challenge in her life.
She acknowledged people find value from things like ChatGPT and Claude. But the value the average person is finding is not necessarily like booking a flight for her.
Ha mentioned companies like Uber. When it first launched, it targeted a tech early adopter set in San Francisco. He remembered being confident this would never catch on more broadly. This was just for a narrow audience of young urban millennials who have money to spare on cars, and it works uniquely well in San Francisco.
For a lot of things, the tech industry has succeeded by betting people have not caught up with them. They believe they are the leading edge and they reflect what the consumer of the future looks like. The family of the future looks like what they envision.
In this case, the bet is not borne out yet. There are a variety of reasons. Partly because the society the technology is being promoted into is very different. People are starting to say they do not love the influence the technology is having in their life. They do not love the pitch that every experience in their life should be optimized out of. Actually, they want to live their own life. They actually want to parent their own children.
What it means
Silberling noted it is not a coincidence that over-optimization is happening among people in tech at the same time there is a movement reverting back to wanting to use flip phones. She has talked about this on Equity before. People are buying CDs because they are tired of an algorithm telling them what they want to listen to.
It is an interesting cultural tension that is happening. They have not found the balance of what optimization actually helps us and what optimization just makes us feel like we are not as human as we want to be.
