Taylor Swift vs Sora

Last year, the AI industry spent $50bn on NVIDIA chips to bring in $3bn of revenue. 

Sequoia’s estimation contributes to growing scepticism around GenAI’s valuations.   

Rana Foroohar talks about gnawing corporate doubts on how to integrate AI into workflows, and “whether it will really be more productive than the humans it may replace.” Deepmind co-founder Demis Hassabis accuses certain players of “grifting”: using terms liberally to benefit from the hype.   

More speculatively, an X user asks why Logan Kilpatrick and Andrej Karpathy would leave OpenAI if GPT-5 and AGI were just around the corner… 


OpenAI’s last big announcement was Sora. It teased fans with glimpses in February. The text-to-video model turns users’ prompts into short video clips. Enthusiasts speculated about its destructive potential for the creative industries. Soon, one mused, viewers would simply instruct Sora to create bespoke binge-worthy TV dramas. 

It spooked professionals too. The actor and filmmaker Tyler Perry reportedly put an $800m studio expansion on hold after seeing initial videos. 


As proselytisers salivated at the prospect of more automation, Taylor Swift embarked on the Asia leg of her Eras tour. At over $1 billion in revenue, it’s the highest-grossing of all time. Prestigious enough for Singapore to risk regional relations with its exclusivity deal, Swifties were everywhere when she arrived in the city-state. Orchard Road resembled Wembley Way as fans of all ages proudly sported tour t-shirts. 


The popularity of live events and our ever-increasing screen time are partly complementary. Broadcasting one’s presence is part of the experience. Ready-made hashtags are even available for funerals

Their enduring success also shows why GenAI falls short. Sora may one day help Netflix churn out more banal content. (Emily in Paris has already been labelled “ambient TV” i.e. the type you can scroll and watch). 


But institutions are now subservient to the individual. When it comes to a TV show, 26 percent of Americans say the actors strongly influence their decision to watch. Even light entertainment would suffer for including AI-generated characters. 

Individuals’ importance only increases in line with the subject. Think of the way politics has changed. New parties led by charismatic individuals are breaking traditional allegiances. Meloni and Macron in Europe or Milei in Argentina. India’s Modi has turned once tight elections into BJP processions. And, of course, Donald Trump. He has singularly captured the Republican Party. 

Or sport. Sir Alex Ferguson’s mantra that “no one is bigger than the club” is impossible to enforce in modern professional football. All the top players have bigger followings than the clubs they play for. Inter Miami were accused of “embarrassing” China when Messi failed to appear in a Hong Kong friendly. The fixture was advertised on the basis of an individual with scant reference to the participating clubs.

Even public intellectuals are now rock stars. One might previously hear from them at book festivals or debates but now they sell out live arenas. I recently accompanied Konstantin Kisin on an Australia tour where he was greeted by adoring fans at each venue. A week later, Douglas Murray was due in the country. His tour was organised by David Bowie’s former promoter. 


The internet started this trend. Subsequent social media platforms and alternative outlets neutered mainstream media. Loyalty hinges on individuals rather than publications or channels. In the month after Tucker Carlson’s exit, Fox lost 39 percent of its viewers

But AI now impairs trust in this online marketplace. Everything can be edited. This varies from the superficial, such as the subsequent autotuning of Alicia Keys’ Super Bowl performance, to the nefarious: pornographic deep fakes of Taylor Swift. 

The latter demonstrates AI’s ability to diminish what Rachel Botsman calls “distributed trust”. She argues the digital economy removes institutions as sources of authority. Instead trust becomes something more personal. We believe what we choose to see. 

Those with an agenda can edit that. In Taylor Swift’s case, these images were shared enthusiastically by conspiratorial US conservatives. They claim she is part of liberal plot to undermine Trump’s reelection.       

AI models themselves are not politically neutral. Google’s Gemini launch was overshadowed by overzealous DE&I programming that refused to show white people. It backfired when it started generating images of multi-racial Nazis. This leads Elon Musk to trumpet Grok’s “anti-woke” answers. LLMs are not repositories of objective truth but motivated by political preferences. 

Knowing these things are possible also helps deflect genuine challenges. Trump decries any obstacle as fake news. Meanwhile the liberal media insists any populist impulses are only the result of misinformation. Supporters still cheer on Carole Cadwalladr’s Russia conspiracies even as her defamation debts balloon

Increasingly, one has to see it to believe it. 


Seeing things live also speaks to a fundamental part of the human condition. We are social animals. A lot of the AI hype is predicated on that changing. But very recent history suggests it’s unlikely. People quickly forget all the wild “new normal” predictions made during Covid lockdowns. 4 years after the initial outbreak, life looks unerringly similar. 

The social instinct prevails. Jonnie Penn suggests something similar will happen with AI. We are increasingly aware of technology’s pernicious impact on our mental health. The popularity of Jonathan Haidt’s latest book, The Anxious Generation, shows this feeling resonates. Penn writes

As with the shifting perception of oil and cigarettes over the past fifty years, the future of AI might surprise inhabitants of our present day. In response to burnout, digital fatigue, algorithmic-racism, -sexism, -ableism, -authoritarianism, and a global mental health crisis, is the prospect of an unprecedented coalition against digital maximalists and their presumptions of “AI-first.” 

Human passivity is often presumed in grandiose AI predictions. But people will ultimately choose how much of it they want. That appetite may be more limited than evangelists imagine. 

AI valuations are speculative. Live events are proven. Just look at Taylor Swift. Her $1bn revenue looks a rather better return on investment. 


One Comment

  1. wonder mail generator

    What i dont understood is in reality how youre now not really a lot more smartlyfavored than you might be now Youre very intelligent You understand therefore significantly in terms of this topic produced me personally believe it from a lot of numerous angles Its like women and men are not interested except it is one thing to accomplish with Woman gaga Your own stuffs outstanding Always care for it up

Leave a Reply

Your email address will not be published. Required fields are marked *