Dishing on generative AI with GamesBeat’s Dean Takahashi | The AI Beat

I’ve been on the AI beat for a year and a half, but my colleague Dean Takahashi, who leads our GamesBeat publication, has been covering the gaming industry for 27 years — 15 of those years right here at VentureBeat. And he doesn’t just cover games, mind you — he reports on all things chips, hardware and R&D as well.

I joke that Dean has probably seen a thousand tech trends rise and fall. But seriously, he has had the inside scoop on Nvidia’s rise to AI dominance for years (the company was founded in 1993 with a vision to bring 3D graphics to gaming and multimedia) with decades of coverage on nearly every major utterance of CEO Jensen Huang.

So I thought it would be fun to take a beat (pun intended) and dish a little bit with Dean about the current era of generative AI. We chatted on a video call from his digs in the Bay Area to mine in the New Jersey suburbs. 

The pace of generative AI may be surprising, but it’s no fad

For one thing, Takahashi emphasized that just because he has seen AI develop over the years, it doesn’t mean there were no surprises as generative AI began to boom last summer.  

“I had already been covering Nvidia for a long time — they had been talking about deep learning neural networks for about nine to 10 years,” he recalled, adding that Huang’s memorable GTC keynote moments around AI began around 2016. “Jensen said that neural networks were the invention to lead all inventions — that made me think we should expect cascading innovations to fall on top of each other — and they really have.” 

But the release of Stable Diffusion last August, and OpenAI’s ChatGPT in November, were still eye-openers: “It’s like I was awakened to AI many years ago, but I was happy to see so much more change happening,” he said.

That said, even though the generative AI space has the sense of explosion of other tech fads that took off in the gaming space, like virtual reality, eSports and blockchain games, Takahashi doesn’t see AI in the same way. 

“The level of AI is on a different scale — I think Nvidia tracks 3,500 AI startups or something like that,” he explained. “I think AI is accelerating now and it’s the kind of innovation that helps out every industry, and it is a continuous innovation that hasn’t gone away.”

The AI art of Dungeons & Dragons

Gaming AI news often neatly parallels what is going on in other creative industries, so it’s no surprise that I was interested in asking Dean for his thoughts on this week’s news that the Hasbro-owned role-playing game Dungeons & Dragons won’t allow its artists to use AI technology. This came after social media blowback around one of its illustrators using AI to create commissioned artwork for an upcoming book. There is growing pushback on generative AI from artists who believe that AI image generators were trained on their copyrighted data without consent, or that generated image output copies their work. 

Hasbro’s reaction, said Takahashi, has parallels to the reaction of gaming companies to blockchain games — where the large companies were excited at first but then pulled back when SEC regulators cracked down on crypto.

“I think you see something similar happening with AI here, where now there are legal and copyright concerns,” he said. “You’re not sure if you feed data from the public market into these large language models, whether it’s going to spit out something that infringes on somebody else’s copyright in the real world.” 

AI concerning for artists, writers and actors, but a boon for user-generated content

In Takahashi’s opinion, going slow and waiting for a legal framework to evolve makes sense. But just like with blockchain games, where many startups simply went forth to get ahead in the market, he said it’s possible that many smaller AI companies will also move quickly. 

That has led to concerns not just for artists, but for game writers and voice actors, he said, noting that many also work in Hollywood, where the media has been covering the SAG-AFTRA strike closely. 

“I definitely see a lot of writers who write for both games and films becoming worried that they can be replaced with these chat programs, and then the same for voice actors, when it’s possible to take a sample of somebody’s voice and put it into anything,” he said. “I think that the writers and actors on strike believe that they’re arguing their case for almost everybody in work — like someday, this is going to happen to you. And we’re the ones that have woken up to this.”

That said, Takahashi pointed out that the timing of the generative AI wave has been great for user-generated content. “Games like Minecraft or Fortnite or Roblox have lots of content that’s created not by the main companies involved, but those who are making their own contributions in the space,” he said. “And in some ways they have advantages because personalized games can be a lot more meaningful to people who are doing that kind of creation for themselves or their friends.” 

Is this a tech time like no other?

Finally, I had to ask Dean Takahashi how he thinks this generative AI era stacks up against other big tech moments over the past couple of decades. He chuckled and pointed out that “people are starting to say that this is a time like no other, there’s no better time to be in Silicon Valley.” While he added that this tends to happen at every big tech juncture, in the case of AI, “the gravity with which they believe it still requires you to pay attention.” 

I told Dean that I thought a lot of today’s AI leaders were likely inspired by games they grew up with — so perhaps there is a symbiotic relationship between the two areas. He agreed that AI, just like games, is about simulation, and offered a San Francisco-based company called Fable as an example. Its game, The Simulation, is a place where users create and nurture AI characters that dwell in different metaverse worlds.

“What they would like to see happen is the creation of general artificial intelligence inside [the game],” he said. “If that happens, then we get some idea of the consequences of that, and we either worry and we put a stop to it, or we see that it actually is not doing any harm.”

“So it’s like setting AI loose in The Sims to see what happens?” I replied.  

Takahashi nodded. “If the thing that happens in The Sims is World War Three and the extreme extinction of human life, then maybe this turns out to be a bad idea for the real world.”