How Veo 2 Can Eat Hollywood

Google has just introduced Veo 2, its latest version of a text-to-video AI generator.

Given the proliferation of AI-this and AI-that, it may seem that Veo 2 is probably no big deal.

But for those who are in the film industry—at virtually every level—calling it a “big deal” is to immeasurably understate the case.

And Google is driving this technology forward at a remarkable pace, and given the company has two additional tremendous assets, it could conceivably knock down the Hollywood sign and build its own version in Santa Clara County.

Veo 2 isn’t just another system that produces somewhat accurate but still slightly off renderings of scenery, objects and people (e.g., the hands with six fingers).

Not by a long shot.

What’s more, like its predecessor, Veo, launched in May 2024, it has video production as its reason for being.

According to Aäron van den Oord, research scientist, Google DeepMind, and Elias Roman, senior director, Product Management, Google Labs, with this new video generation tool the user can achieve resolutions of up to 4K and run times measured in minutes, not a series of seconds.

What’s more, Veo 2 is achieving verisimilitude with techniques that are analogous to what happens in the real world on soundstages everywhere.

As van der Oord and Roman describe it:

“Veo 2 understands the unique language of cinematography: ask it for a genre, specify a lens, suggest cinematic effects and Veo 2 will deliver. . . . Ask for a low-angle tracking shot that glides through the middle of a scene, or a close-up shot on the face of a scientist looking through her microscope, and Veo 2 creates it. Suggest ‘18mm lens’ in your prompt and Veo 2 knows to craft the wide-angle shot for which people use this lens, or blur out the background and focus on your subject by putting ‘shallow depth of field’ in your prompt.”

The director and cinematographer rolled into one system.

The output resembles actual scenes with actual people doing things in a natural way (or in geek-speak: Veo 2 has “an improved understanding of real-world physics and the nuances of human movement and expression.”)

As this is a text-to-video AI generator, it simply requires someone at the keyboard to come up with instructions, such as “Write a scene for a film noir movie shot in the style of John Huston with actors who resemble Brad Pitt/Ryan Gosling and Amanda Seyfried/Sydney Sweeney who are in a seedy bar with the appropriate background characters and alcohol accoutrements plotting to kill their spouses; include dialogue.”

And the Google tools do the rest.

This is not something that has the quasi-actual appearance of a high-end video game. This looks like the real thing.

“Veo 2 understands the unique language of cinematography” should be words that cause an array of members of a conventional film crew to reach for the Mylanta.

While Veo 2 is limited to minutes at present, how long before there is Veo 4, 5, and Infinity. . . that will even go to Marvel-movie length?

And while it might seem that this would be beneficial to screenplay writers, especially those who are comprehensive in their descriptions of the setting and action, isn’t it likely that a large language model like GPT-3 (or maybe 4, 5 or. . . .) simply does the job with a few prompts, thereby rendering the writer superfluous?

So Google has an incredible tool. Consider it one leg of the stool that is what it takes to make an industry. Veo 2 is the Means of Production.

But like any good stool, Veo 2 requires three legs.

In late October, a Russian court fined Google 2,000,000,000,000,000,000,000,000,000,000,000,000 rubles.

Or, translated into dollars, that’s $20 undecillion. To put that into perspective, that’s more than the GDP—of the entire world (a mere $110 trillion).

Now while the size of the fine may seem (and be) absurd, obviously the judges levying the fine realize that the current occupant of the Kremlin doesn’t want to be made to look silly.

So, in effect, that $20 undecillion fine shows that the Russian state is serious.

What, you may wonder, did Google do to get the fine?

It simply restricted access of 17 Russian state media outlets to YouTube. Google has owned YouTube since 2006, when it acquired it for $1.65 billion in an all-stock deal.

There are on the order of 4.3 billion videos on YouTube, growing at a rate of 9.7 million per day. This encompasses everything on how to fix a leaky faucet to how to survive a black hole, from the Coen Brothers’ No County for Old Men to Adam McKay’s The Big Short.

Another big number: YouTube has 4.95 billion monthly active users. To put that into context, Netflix has 221.44 million monthly active users, or 4.47 percent of the number of viewers YouTube has.

And the Russian state recognizes the global importance of YouTube, which is no minor observation.

So what do we have here?

•Money. Alphabet, Google’s parent company, is valued at $2.4 trillion.

•Distribution Channel. YouTube even has more monthly active users than TikTok, which is at some 1.8 billion.

•Means of Production. Veo 2 and its ancillary elements.

So if the world’s governments get out of the way, Google has the three legs necessary for the solid stool.

According to the Motion Picture Association, the film and television industry supports 2.74 million people—”from special effects technicians to makeup artists to writers to set builders to ticket takers and more”—and pays annual wages of more than $242 billion.

Big numbers. Not Google big, but certainly significant.

But one has to wonder: what will be the consequences should Google go all-in on the creation of entertainment products that they can produce with a fraction of the number of people that that these products currently require?

Maybe it won’t put Hollywood completely out of business—after all, there are still places that serve hamburgers, even though McDonalds has sold over 300 billion—but the consequences may not be pretty.