To AI or Not to AI? Industry Thought Leaders Make Their Cases in a Lively Debate at Streaming Media Connect 2024
On Thursday, August 22, 2024, at Streaming Media Connect, key industry thought leaders and stakeholders engaged in a lively debate on “The Real Value of Artificial Intelligence” in today’s streaming ecosystem. Panelists included media cartographer Evan Shapiro, Chris Giliberti of Avail, H Schuster of HUSSLUP, and Peter Csathy of Creative Media. Streaming Media Editor-in-Chief Steve Nathans-Kelly moderated the debate.
The argument for the lower risks of a short-term AI bubble vs. long-term
This panel expanded on the various arguments that began at the Streaming Media NYC debate “AI: Boon Or Bubble?” The first speaker, Peter Csathy, Founder & Chairman, Creative Media, suggested that while there may be a short-term AI bubble with VC-backed companies, long-term AI investments with big tech companies are not in a bubble.
“You have companies like Inflection who are trying to compete with the big tech companies and are realizing that they just don't have the resources necessary to be able to do that,” Csathy said. “And now you see a lot of the VC-backed companies being absorbed into larger ones. So, from a long-term perspective, we are not in a bubble. Short-term, it is much like the internet. When the internet came in and transformed our world, you had massive investments on the VC side, and about 80% of those, according to statistics from Harvard Business Review, failed over time. And similarly, we will see the same thing happening in the AI generative AI space.”
Csathy emphasized that ChatGPT is only less than two years old. “We have certain notions of where this is all going, but we have no idea ultimately what the big opportunities will be,” he said. “We do know that AI is already tackling some of the most existential issues on the planet, including climate change, healthcare, and pharmaceuticals. From a resource perspective, you see estimates that technology companies will invest over $1 trillion in AI by the year 2030. And when that much investment is made, markets are made.”
He further underscored the unpredictable nature of rapidly evolving technology and the potential demand for it, noting that when Apple first released the iPhone, most of the market thought it was a bad idea, while no one anticipated the coming scope and power of the app economy.
Why “all bubbles are bad,” short-term or not
Evan Shapiro challenged that view. He warned of the potential dangers for common investors and suggested that the bubble could affect major tech companies, potentially dragging the entire market down. His view is that, short-term or not, all bubbles are bad.
“The definition of a financial bubble includes the word irrational, and the reaction to the market or the participation in the market around bubble economics is always irrational,” Shapiro said. “It does not consider the actual facts or the actual values at the basis of an economically growing new ecosystem. You can look at the first internet bubble Peter [Csathy] referred to as a good example. That internet bubble bursting created a recession that cost people jobs and created a lot of damage in the marketplace. We can all remember the silly companies that were started during the first internet bubble. I worked for a lot of them. There is no question that on the other side of a bubble, the technology that is at the center of it may create tremendous progress and value, but the bubble itself is usually a very damaging process.”
Shapiro said that he does believe AI is a powerful tool with great potential. “We haven't begun to plumb with the depths of its possibilities, its opportunities, and its true value, very much like that first internet bubble,” he said. “We thought the internet was for one thing, and then suddenly broadband entered the picture, and we had a different product on the other side of it. And I think we are in the bubble before the change. [But] in the interim, we are in a dangerous era.
Shapiro provided several current examples to back up his argument, including the recent erraticism of AI giant NVIDIA’s valuation, how Microsoft’s stock value recently dropped 7% despite positive second-quarter earnings because they could not articulate to Wall Street why they are investing in so much AI and where the return on that investment will come. Google just experienced a very similar issue.
Shapiro noted other significant issues such as the “fake math” of AI startups, who often frame their revenue around annual recurring revenue (ARR), while their customers are actually paying month-to-month, which means those contracts are less stable than they say. “Additionally, hedge fund customers just wanted to test the startup software to recreate it for themselves in-house for less,” he said, quoting an article from The Information. “The whole market is predicated on the high-end demand for these chips, which is not being met, thus driving up the price of the chips. Well, the demanders, a lot of them, not all, have no fucking idea what they're talking about. They're creating a model because they see an opportunity in a bubble, and they're watching all this money go up. It's not like NFTs, but it sounds like NFTs.”
Ultimately, Shapiro said that what AI will truly accomplish is much more mundane than the current hype around it. “I think that the true value of AI is going to be in the really small boring shit behind the operating system, making the user experience better incrementally, minute by minute, second by second, inch by inch, millimeter by millimeter,” he said. “And so I think that the great use is going to be in the backend. In streaming, the ability to understand compression needs on the end user for the video product, the ability to better personalize each experience for the end user [and to] personalize [the advertising] for the brands that are out there as well.”
How AI might be the tool to burst the Hollywood bubble
H Schuster, Founder & CEO, HUSSLUP, suggested that AI could be the tool that bursts the Hollywood bubble, leading to a reimagining of the creative industries.
“I want to approach the question by looking at a specific segment of how AI is being used to solve problems in the entertainment sector because that's where I work, and it’s where my company is,” Schuster said. “And so is AI in a bubble? No. And, I think with the specificity of the creative economy, AI is going to pop a different bubble, which is the studio bubble, or the current traditional media Hollywood bubble.” She said that the period of “peak TV” was itself a bubble, and since 2022, studios have pulled back from production spending and used the industry strikes as a cover for massive downsizing. Yet, the demand for content has not gone away.
“You've got millennials watching four and a half, five hours of video content a day, and major studios do not distribute a lot of it,” she said. “So, creatives making content need tools to make better content more quickly and to monetize it with this audience outside of this traditional model.”
She showed a graphic that read: “Enter the AI revolution: like the printing press and the internet, AI will save trillions in costs and create hundreds of trillions of in value. Like most new technologies, AI endangers some jobs or at least changes the way we work. However, AI will democratize the creative process and level the playing field for a majority of creators, especially the creators who are excluded from the gated studio system.”
“So what does that new creative process look like?” she asked. “Instead of waiting on those four vertically integrated conglomerates to greenlight your content, you can use gen AI tools to create more IP faster, better, and more cheaply, and to help self-distribute your content.”
She noted that Hollywood’s creative class is justifiably scared. They don’t want generative AI to replace them. As outlined on another graphic she displayed, “These tools need input, guidance, and vision, but it is a different use of skills. They do not want their original IP to train LLMs for free. No one knows what a licensing deal is really worth. Will they be paid a pittance compared to the value created for the LLM over time?”
“Distribution is going to be a critical piece of this,” she said. “Gutenberg's biggest achievement was to print 200 Bibles in Latin on his printing press, but he had nowhere to sell them. And in fact, he died pretty destitute. eBooks similarly struggled. They didn't really have distribution until Amazon created Kindle. I think creators are using Gen AI tools, and they are starting to use them to make better content faster. They no longer need the studio to distribute it. They can do that on YouTube, social media, and platforms like Kinema and Mubi. But I believe AI will probably revolutionize the distribution piece of this pipeline, as well, through a lot of discovery tools. So, generative AI tools will be more productive and efficient than the traditional Hollywood models we're using right now, and I think they will burst the Hollywood bubble.”
How the overall accumulative and transformative potential of AI may negate any short-term losses
Chris Giliberti, Co-Founder & CEO, Avail, argued that “whatever small bubble-like aspects of what we're seeing today are going to be dwarfed by a new wave of innovation unlike anything we've ever seen before.”
Still, he acknowledged that the present situation is bubble-like and ultimately unsustainable. “The foundation model space is a great place to look if you want to make the case that there's a bubble,” he said. “If you look at Hugging Face, there are thousands of apps out there. So there's a lot of investment in a lot of competing technologies. And that sort of view of the landscape is certainly not sustainable. There's going to be a big shakeout. I think it's going to work the way it's worked across every other innovation cycle in tech, where there's a power law and the winners are concentrated. And I don't think we know who those winners are yet. The other thing that we see firsthand here at Avail is the market for training data, which is also really running up. Some might even say that irrational amounts of money are getting paid for training data by the big foundation models whereby one projection of market research, there's a 2.5 billion market here at this moment, and it's projected to be 10x over that in the next ten years."
Ultimately, he said, “I think you could look in the public markets and say there's some irrationality. You could look at the private markets and draw the same conclusion. But I think all of this is short-term noise. I think we are at the foundation of a massive sea change here and a massive amount of value creation.”
Giliberti compared the early internet bubble and its rapid rise in revenues to today’s increase in gen AI revenue. He argued that even with the initial 2000 crash, the overall accumulation of value that the internet brought was a net positive. “When we can unleash AI onto science and tech problems in earnest and [handle] mathematical problems in all their forms: coding, finance…I think we're going to see another huge wave of innovation that's even more difficult to predict at this moment. When the independent reasoning and critical thinking abilities of these models improve to the point that they're more similar to human intelligence [and we reach Artificial General Intelligence, AGI], all of this is going to get supercharged. And so it's just a tremendous amount of leverage in the economy. When I look at the valuations today, sure, a lot of investment is getting poured in. Some of it is misplaced. Investing is a power law business. I actually don't think bubbles are necessarily bad. If they're in service of something where we could fundamentally envision a future that's really exciting, at least we could argue the opposite of crypto. I think that's what we're seeing here, and that's why I'm ultimately sort of netting out in this place where we are not in a bubble.”
Click here to watch the entire "To AI or Not to AI?" Streaming Media Connect debate.
Streaming Media Connect will return November 12-14 for another round of deep dives and provocative conversations with the top thought leaders in the industry.
Want to join the conversation? Got a topic to propose?
Related Articles
Streaming Media presented its 14th Connect virtual conference August 20-22, 2024, with Media Industry Cartographer & CEO ESHAP Evan Shapiro and Streaming Media Editor-in-Chief Steve Nathans-Kelly as hosts and MCs, and featuring speakers from Google, Amazon, Roku, Bloomberg, DAZN, Fubo, IMAX, and more, and session topics ranging from streaming monetization to FAST strategy to OTT bundling to cloud streaming OpEx and live streaming at scale.
22 Aug 2024
On Tuesday, August 20, 2024, at Streaming Media Connect, Alan Wolk, one of the industry's most influential thought leaders, moderated a panel that confronted the current challenges (or, some would say, failures) to reliably deliver the right ads to viewers and measurable results to brands and discussed innovative technologies and strategies to make streaming ads perform. Panelists included Christina Chung of Estrella MediaCo, Julie Triolo of Vevo, C.J. Leonard of Mad Leo Consulting, and Erin Firneno of Advertiser Perceptions.
20 Aug 2024
AI enhancements and a Cloud platform build out are the twin developments promoted by Telestream heading into IBC. This includes the unveiling of a comprehensive Cloud platform which will eventually house all of its products, and a focus on AI to streamline user operations.
15 Aug 2024
In a recent interview, Simon Clarke, Chief Technology Officer of Telestream, discussed two key products, Wirecast and Vantage. Celebrating Wirecast's 20th anniversary, Clarke reflected on its evolution from a nascent streaming tool to a robust application used by a diverse range of streaming producers, including educators, business professionals, and houses of worship.
25 Jul 2024
On Monday, May 20, at Streaming Media NYC, two lively debates on the topic of AI in Media and Entertainment showcased leading industry figures. The first, "AI: Boon Or Bubble?" examined the ways that AI has entered every sector of the media, business, and technology worlds, and it explored how much of AI's insurgency is a fad, and how much of it is the actual future of streaming. The second, "The Real Value Of Artificial Intelligence," explored the fearmongering around the idea that "AI is coming for our jobs," and focused on how AI/ML might make workflows more efficient and effective, grow or refine reach, and help to boost bottom lines.
20 May 2024