AI is bigger than the internet. The industry must make it ours, or be replaced. Treat this as year one.

Pereira was the kid always in the corner, drawing, writing or programming, and he started his career as a programmer. The thread running through everything is that he has always been an outsider combining images, words and code, and for a while nobody understood him. At his first advertising job he was asked whether he was a copywriter or an art director; he said he was a programmer, and was told programming was not creative, it was tech people's work. It took the industry thirty years, he says, to realise technology can be creative too. Now he can call himself all three at once.

Everyone claims to be curious, Pereira says, but the moment they decide something might be a waste of time, they stop. Real curiosity requires being ready to waste time, and he enjoys it, because it is investment in the future.

The second half of his method is to pursue ideas that are not doable. The business trains people to only propose what can certainly be done, because an undoable idea makes the agency suffer. But if an idea turns out impossible, for that moment you are ahead of the technology, and when it catches up you are already half built, because the idea is ready.

What excites Pereira most is the ability to think, see and do all at once. He has won pitches by having a thought, going to his backyard with his phone, and making the spot in a day rather than writing a script. Whatever is in his head, he is now responsible for making, rather than praying someone else understands what he means. It is not very scalable, he admits, but it is fun as hell.

Great work has to pay back for both the consumer and the brand, and that demands discipline. Walking Cannes, Pereira sees an industry obsessed with breaking the rules to feel fresh while disregarding the discipline of the business. A campaign that needs a case study to be understood, he argues, is unlikely to work in the real world. He says this after winning Grand Prix with web series and feature films, so format is not the issue.

The Intel series ran four years on one relentlessly repeated idea, that what's inside counts, building on the Intel Inside sticker. The story was just a way to get people to talk about that idea; you didn't need to watch it all. People saw big brands and formats beyond thirty seconds and missed the point. If the story had fit in thirty seconds, he says, he would have made it in thirty seconds. It is a very Bruce Lee thing: they pointed at the moon and looked at their finger.

The idea that brands are either more ambitious or more cautious is, for Pereira, a fake dichotomy. Clients and agencies are both getting more ambitious. The real divide is between two festivals at Cannes, the Palais and the beaches, where the conversations differ. A generation of judges trained to craft case studies keeps saying, I can see the case here already, and that is a dysfunction. When creatives complain clients aren't ambitious, their ambition is simply different from the client's. The cool a CMO wants is not the cool the industry is chasing. They are pointing at different moons, and there is only one.

The company began by accident. Flying to Cannes ten years ago as jury president, Pereira wrote a script about kung fu versus artificial intelligence, turned it into a novel, and spent eight years on it, interviewing ethicists, engineers and brain scientists. As the book neared launch, the rumblings of generative AI began around GPT-2, and he started experimenting. The output was poor but fascinating, and because he had always worked with technology he could see it would get there.

Having run digital agencies in the first wave, Pereira and his partner remembered how traditional agencies could not adjust fast enough, bound by their structure. So they built a lab, not an agency, unbound by legacy, and named it Silverside after the small Bay Area fish that sits in the middle of the food chain, helping the ecosystem thrive. It grew from a lab doing things they didn't understand into a company doing strategy, creative, production and technology, and helping clients transform. It could have been a waste of time, he says, but not this time.

People talk about AI as if it were one thing, Pereira says, but it is a generation of technologies underneath every task in every profession: in your headphones, in a song's production, in the algorithm that served it. To stay in front, you have to realise the task at hand is only thirty percent of what you're doing. Every step is a glimpse of what's next. This is a very experiential technology that you cannot understand from a report.

That is why he distrusts executives who do a research deal and think they are ahead for two months, until someone outdoes them. The gap between interns and CEOs has never been smaller, because the youngest people living inside the technology often know more than the leaders. For Pereira this is a genuine economic and structural change, echoing the early digital days when he, as a junior, was the one person the agency CEO could not out-know.

Pereira wrote a book where AI was the villain, so he painted the worst picture and does not discount the dangers: a society that can't believe what it sees, a handful of men with more power than countries, content used as fodder with no respect for its makers. The law is too slow. But from the inside he sees people crafting their work carefully out of respect for their peers, and twenty-three-year-olds directing films on national TV instead of doing coffee runs.

His warning is blunt. Zuckerberg has said advertisers won't need agencies because the platforms' tools will do the work, and the engineers are genuinely trying to build technology that replaces the whole industry. If the industry does not steal that technology and make it its own, they will succeed, and it will be the worst time in advertising. Eventually the platforms would relearn creativity and a new industry would grow up that is not us. Take it and make it ours, he says, and they have no chance.

Three big changes are hitting at once. The economics of creativity have inverted: what was expensive is now cheap, what was cheap is more expensive, fast is slow and slow is fast. Everything built to be efficient with ideas is up for grabs, and the industry is not ready to rethink it. The creator economy is remaking how information and entertainment exist, and the business is unprepared for that too.

The soul of a brand, the intuitive sense of how it should speak, used to belong to the creative director. That is passing to the CMO, through creative directors' failings or CMOs' egos; blaming won't help. If the industry fights it, it makes enemies of allies. Put it all together and the business has to reinvent itself. This is Pereira's twenty-seventh Cannes, but we should all look at it as year one, because nothing we know really applies for next year.