Twitch CEO: AI rush missing the point about human intelligence – connection
The debate about whether the emerging generation of digital entrepreneurs is building a superintelligence misses a key point, according to the CEO of Amazon-owned streaming service Twitch. He told attendees at SXSW Sydney yesterday that human intelligence is fundamentally different to machine intelligence because it is based on a desire to connect with other humans and belong. It's not about analytical reasoning. And while not a doomsayer on AI – he believes it will yield both good and bad results – he doesn't think we have the capacity to avoid bad results any more than business leaders during the industrial revolution could have avoided climate change. "We can't fix it now, why would we think they could fix it then."
What you need to know:
- We misunderstand human intelligence when we focus too much on its capacity for logical sequential analysis.
- That's just a side effect of our desire to collaborate, per Twitch CEO Dan Clancy.
- Community is the foundation of what it means to be human, he told attendees to his keynote speech at SXSW Sydney yesterday.
- But how long will it be until we can authentically fake sincerity? We're not there yet say brands, but maybe one day we'll get there says Linktree's CEO.
- Meanwhile, AI's aristocracy is divided into two camps, those who believe we will create a superintelligence (who coincidentally are trying to raise money off the promise) and those like Meta chief scientist Yann Lecun who says, "The term AGI (artificial general intelligence) to designate human-level intelligence is nonsense."
I think we undervalue the complexity and sophistication of human intelligence and what we can do in this world when we say it's going to be as smart as us.
Too often when we think about AI, we misunderstand human intelligence, according to the CEO of Amazon-owned streaming platform Twitch.
"All of us think that what is so unique about being human is our intelligence, our intellect and how smart we are... how we can play chess, all that sort of cool stuff," per Dan Clancy.
To counter that he referenced research suggesting that our ability to think deep thoughts is more of a consequence of developing the big brains needed to facilitate our desire to collaborate with other human beings.
Clancy runs a business built on the concept of communities and connection – a lot of its success, for instance, has come from aggregating users of the gaming community who he says are better thought of as “people who game” rather than gamers. Twitch facilitates that sense of community in gaming, which is one of the strongest aspects of the gaming culture.
The idea of collaboration and belonging is especially important to remember as we develop artificial intelligence. In Clancy's telling, it is important not to lose focus on the "artificial" nature of machine intellect.
“We forget that [community] is the foundation of what it is to be human. It’s that connection with other human beings. We get caught up in our reasoning and our thinking and our brains because it is quite impressive. But really, the foundation of being human is that connection to other people and that sense of belonging.”
Almost as an afterthought, he noted it is also the reason brands are attracted to the platform.
“Brands are trying to create an emotional connection with their audience, and what they're trying to do is reach someone at a time that they are emotionally engaged and positive, so they get a positive brain association. And that's what's happening when somebody's on Twitch – they're open to that connection.”
'Not yet'
It's already a live issue that brands need to consider.
Speaking on a separate panel about using AI-created influencers, Danielle Pearce, the founder and CEO of shoe retailer Merry People said, "As a brand, we've spoken about it internally, and I don't think AI is there yet in terms of content creators. And for us, authenticity and connection are just huge. And we wouldn't do it as yet, because I just don't think it's anywhere near a level of working with a content creator who has a genuine connection with an audience."
Her fellow panelist, Linktree co-founder and CEO Alex Zaccaria, said, "Maybe one day we'll get there. I think that authenticity just matters so much. It's the connection that you have with the creator. That's the reason you get influenced, or your content is influenced by creators because you have a connection with them. And generally, that's a human connection or some kind of connection about their style ... AI may eventually get there, but I think there is a human connection right now it's incredibly beneficial."
He noted however that, "More and more people can use AI to create content, to make the way they manage their business easier, the way they manage their finances easier, the way they create content is a lot easier. And so we're going to see a lot more people entering, hopefully lowering that barrier to entry."
A brain by any other name
Had we called AI "cool algorithms" early on, as opposed to artificial intelligence, Twitch's Clancy believes that much of the confusion and fear that exists about it could have been avoided
“I think we undervalue the complexity and sophistication of human intelligence and what we can do in this world when we say it's going to be as smart as us.”
He is not alone in his thinking. A clear division is emerging in the AI world between those who believe we are on the cusp of releasing a superintelligence into the world and those who believe nothing of the sort.
In the first camp are people like OpenAI's and Anthropic's CEO's Sam Altman and Dario Amodei, while the latter camp includes researchers like Yann Lecun, Meta’s chief scientist, and an ACM Turing Award Laureate who says, “Human intelligence is highly specialised. The term AGI to designate human-level intelligence is nonsense.” (A cynic might also point out that another big difference between the two camps is that the first is trying to raise money from investors off the back of its promises of superintelligence, and the second doesn’t need to.)
Clancy gives the impression he is more comfortable in the second camp.
“When I was growing up, I had a calculator that was a hell of a lot smarter than me at math, I don't know about you guys, but this idea that AI can do stuff that I can't do, computers have been doing stuff that I can't do [for as long as] I can remember.”
More importantly, he says, human intelligence is not just about logical reasoning, it builds upon human relationships, and the bonds and the connections we make, he says.
For good and ill
Clancy also believes AI will be both good and bad, but even if we can perceive the bad aspects we likely lack the ability to do much about it just like the people confronted with earlier historical technology disruptions. That’s a familiar pattern he suggests.
“Whenever there is a big change in the world – like the industrial revolution, there are obviously negative things that can come from any big technological change. During the industrial revolution, they didn't realise that it was going to lead to global warming.”
“With social networks, they didn't realise what could come from that.”
Even as we anticipate the bad things, we lack the wherewithal to change, he suggested.
“Sometimes we like to think that we can anticipate and prevent bad things from happening. Suppose you told someone during the industrial revolution that years from now this is going to cause global warming. Do you think they could have solved that problem back then? We can't solve it now, there is no way they could have solved the global warming problem at the beginning of the industrial revolution.”
Likewise social networks. "If someone told everybody in social networks early about the problems it would cause when it first emerged could they have solved the problem, we can't solve the problem now.”
It’s the same with AI he said. “Are there going to be challenges? Certainly, that's the very nature of technology. The challenge is not with the technology, it's actually with the people and use of the technologies and the connections and the relationships that they have.”
The degree to which algorithms divide us into discrete camps is a big issue with AI today, he suggested, echoing the sentiment first expressed by comedian Aziz Ansari that we are all trapped in our own algorithms.
Per Clancy, “A lot of people familiar with system one and system two thinking … system one is that intuitive, unconscious and emotional [part of our brain] and 95 per cent of what we do is driven by that. What usually happens is our rational brain finds a justification for what we want to believe.”
“And where do so many of our beliefs come from? Our communities. We believe things because our friends believe things. And because we seek that belonging with other human beings, that causes us to believe and share those beliefs with other human beings we care about.”
Clancy offered no solutions as to how we solve the unintended negative consequences of this.
But at least at Twitch, he’s found a way to monetise it.