🙃 Keep America weird! (It’s good for innovation and growth.)
🙃 Keep America weird! (It’s good for innovation and growth.)
Also: Is Generative AI … (gulp!) … doomed? A Quick Q&A with internet law professor Eric Goldman 😨
Who is “weird” — and how exactly that weirdness manifests itself — is now an issue in the 2024 US presidential campaign. Of course, “weird” can be spun differently. To be weird can mean being original and creatively different from the norm. An independent thinker. “Weird” can also suggest something unsettling, off-putting, or socially unacceptable. Currently, it seems the latter definition is the one dominating the political conversation.
But it’s the former definition that seemingly has more economic relevance. For example: The 1992 study, “Why do some societies invent more than others?” analyzed invention patents granted to 33 countries over four years, comparing them to a framework that measures national cultures across six dimensions, including individualism and acceptance of inequality. The results showed that “individualistic and non-hierarchical societies are more inventive,” attributing this quality to factors like decentralized authority, trust-based socioeconomic systems, and valuing individual recognition.
Then there’s “Individualism, innovation, and long-run growth” from 2011, in which researchers measure how much a society values individual versus group goals. Their conclusion: Countries having a more individualist culture have enjoyed higher long-run growth than countries with a more collectivist culture.” American-style cultures provide both financial incentives for innovation, but also social status rewards, “leading to higher rates of innovation and economic growth.”
A new analysis provides a particularly clever way of exploring the issue. In “Culture clash: What 14 million images tell us about times a-changin’,” University of Zurich economists Hans-Joachim Voth and David Yanagizawa-Drott employ machine-learning techniques to examine over 14 million images from 111,000 yearbooks over the years 1950 to 2010. The analysis focuses on three main concepts: individualism (“how many students within each high school dare to have a different style from their peers”), persistence (“similarity between present styles and those of 20 years prior”), and style novelty (“the emergence of previously unseen style choices”).
Not surprising: Style innovation has increased for both, with men experiencing a dramatic rise in the late 1960s. Both genders also reached unprecedented levels of style novelty by 2010, marking a significant shift in cultural expression over the decades.
From the paper:
Picture yourself strolling through the hallways of a 1950s American high school. The scene is a sea of conformity: boys with crew cuts, many sporting jackets and ties, while girls don demure dresses and perfectly coiffed hairdos. Fast forward to today, and you would be greeted by a kaleidoscope of colours, hairstyles, and fashion choices that would make your head spin. This stark contrast serves as a potent reminder that culture is far more than just a set of beliefs and attitudes – it's a way of life, constantly evolving and shaping society.
Still, as the economists naturally ask: “Does any of this matter beyond the realm of fashion and hairstyles? Should economists care about necklines and ties in high school yearbooks?”
Actually, there does seem to be a significant economic implication, as demonstrated by the 1972 yearbook photo of Apple co-founder Steve Jobs: “Jobs appears with bow tie and tuxedo, sporting long hair and no beard or moustache. At this point in time, fewer than 0.3% of US male graduates had ever worn this style, qualifying him for the ‘style innovator’ category. Jobs also went on to apply for 1,114 patents, of which 960 were granted.”
To see if what they noticed about Steve Jobs was true for other high school students later in life, the researchers matched style innovation in “commuting zones” — large geographic areas that often include multiple counties or cities — with the patenting rates of those born in those zones. And sure enough, in places where high school students were more likely to have unusual or creative styles in their photos, more people ended up inventing things and getting patents for their ideas as adults.
As the economists conclude: “While handing out earrings to boys and cutting Mohawks will not necessarily raise technological creativity, schools that permit one of these style innovations may well also produce graduates that excel at the other.”
More evidence, then, that we need a culture where people feel free to express themselves both stylistically and through their opinions. (Indeed, the modern world was built on the “weird” rejection of a norm: belief humanity couldn’t change its future for the better.) The threat of public censure or cancelation can emerge in various ways and affect various aspects of our lives. The freedom to be weird is one reason I’m betting that the 21st Century will be the American Century, not the Chinese Century. Overall, America still seems pretty good at being weird (though I worry sometimes about freedom to be controversial and contrarian). Now all we have to do is spend more on science research and make it easier to build!
Q&A
😨 Is Generative AI … (gulp!) … doomed? A Quick Q&A with internet law professor Eric Goldman
Generative AI, such as chatbots ChatGPT and Claude, could spur this decade’s techno-optimist wave, but so far its reception has been quite different than that of the internet back in the 1990s. Rather than unbounded curiosity and excitement, GenAI has been met with panic and premature attempts regulation. Eric Goldman addresses the stark differences between the introduction of the Internet vs. GenAI in his paper, Generative AI is Doomed. Goldman argues that it’s not too late to adjust our attitudes about GenAI and reap its full benefits, but that doing so means we have to correct course quickly.
Goldman is a professor at Santa Clara University School of Law, where his research focuses on Internet Law, Intellectual Property Law, and Advertising & Marketing Law. He is also the associate dean for research, co-director of the High Tech Law Institute, and supervisor of the Privacy Law Certificate.
1/ In a nutshell, why do you believe Generative AI is doomed?
Generative AI is threatened by two interrelated dynamics. First, regulators are pursuing a massive volume of regulation. Second, those regulations disrespect the editorial nature of making Generative AI models available. In other words, regulators are proceeding as if they can dictate every aspect of Generative AI's operations and outputs. However, Generative AI models often involve the same editorial processes as other types of content publishers, which means they should qualify for First Amendment protections for many aspects of their work. Rather than acknowledge these potential constitutional limits, regulators are blasting ahead with censorial regulations.
The combination of these two dynamics will be devastating. It will produce too much law that is too intrusive. If any model-makers can survive this hostile environment, they will have to be exceptionally well-funded and willing to let regulators curb their editorial flexibility.
2/ What factors contributed to the Internet-era optimism vs. the AI-era pessimism?
There are many reasons why the Internet initially got more favorable regulatory treatment than Generative AI is getting, but I think a top reason is the differences in media depictions. All our lives, we have watched TV shows and movies where "AI" and autonomous robots have turned murderous. These media depictions have socialized us to instinctively fear anything called "AI" as a personal threat. (Imagine how much less threatening it would sound if, instead of the term "Generative AI," it was called "computer-assisted content creation tool").
In contrast, the Internet was stranger than science fiction. Before the mid-1990s, few TV shows or movies had ever contemplated anything like the ability of users to talk with each other through an electronic network. Thus, when the Internet burst onto the scene, people hadn't been socialized to fear it like they've been socialized to fear AI. This provided the Internet with a honeymoon period where people marveled at the opportunities and power of the technology, rather than immediately focusing on all of its problems.
3/ How did discussions about incorporating the Internet into the workplace vary from current discussions about AI in the workplace?
Generative AI raises serious concerns about labor displacement, which follows on decades of labor displacement by other forms of automation. Generative AI allows for labor-intensive content creation to be made much more efficiently, so many knowledge workers may find that their value to their employers is more of a commodity than they thought.
The Internet's rise made many content creators nervous, but their concern was more about piracy than labor displacement. If anything, the Internet was viewed as a job creator — a promise it has fulfilled in several ways, especially in enabling new content services and online merchants to build successful businesses that weren't possible in the offline world.
4/ Can you elaborate on the concept of “Techlash,” why it happens, and its impact?
"Techlash" refers to the backlash against the Internet companies. The techlash reflects our general skepticism of large and powerful entities. We instinctively assume that they abuse their power and reach to benefit themselves at everyone else's expense. Indeed, a series of embarrassing gaffes over the past dozen years — such as Google's wi-fi sniffing and Facebook's Cambridge Analytica data leakage — have reinforced these suspicions. As the big Internet services became some of the wealthiest companies that have ever existed in human history, our skepticism grew as well.
At the same time, politicians have weaponized these skeptical attitudes. They keep claiming that "Big Tech" is a threat to us, even though the government has far more power — and far more ways to abuse its constituents — than even the biggest Internet service. Worse, politicians have found that bashing Internet services can have partisan payoffs, such as accusing Internet services of engaging in partisan bias against one party that helps the opposition party.
Once the Internet services became a pawn in this partisan chess match, it signaled the end of any rational policy-making. Instead, the techlash now is all about partisan gamesmanship and scoring partisan points. Unfortunately, Generative AI is already mired in allegations of partisan bias, another dynamic that darkens any rational attempt to discuss sensible regulations.
5/ What can we learn from the Dot Com boom and bust and how can it inform our expectations for GenAI?
Perhaps the most important lesson to learn is just how many things had to go right for the Internet to succeed. In particular, the Internet needed the proper legal foundation (laws like Section 230) and to hold the rightsowners at bay so that they didn't impose tollbooths at every juncture. At the moment, Generative AI may lack both that legal infrastructure and checks on the rightsowners' flexes. That puts the entire category of Generative AI at substantial risk.
6/ What societal or cultural changes would you deem necessary to propel GenAI in the right direction and improve public outlook?
The antipathy towards Generative AI is so strong, I fear nothing will deter the efforts to derail it. However, there are two things that might help.
First, so long as Generative AI is semantically linked to "artificial intelligence," we instinctively will fear it. It would be great to come up with a new name for "Generative AI" that doesn't include the AI reference. That would make it seem less scary.
Second, we will need clarity about whether Generative AI outputs are protected by the First Amendment. Although the machines are "authoring" the speech, the Generative AI model-makers are still publishing content to their users. If this publication function gets full First Amendment protection, then most of the current efforts to regulate Generative AI will be obviously unconstitutional. If Generative AI doesn't qualify for First Amendment protection, then regulators will have no limits on their interventions, and they will overregulate the industry out of existence.
On sale everywhere ⏩ The Conservative Futurist: How To Create the Sci-Fi World We Were Promised
Micro Reads
▶ Business/ Economics
Chipmakers face a labour crisis - FT Opinion
▶ Policy/Politics
H-1B Visas Should Be Allocated by Merit, Not Lottery - Bberg Opinion
We All Pay the Price for Protectionism - WSJ Opinion
Procurement and Infrastructure Costs - Brookings
The FEC Has No Business Regulating AI - WSJ Opinion
In a Blow to the FCC, Court Suggests Net Neutrality Is a Major Question That Congress Should Decide - AEI
▶ Clean Energy/Climate
Untethered: Decentralizing Energy to Transform War - Andreessen Horowitz
▶ Robotics/AVs
The Chinese government is going all-in on autonomous vehicles - MIT Tech Review
▶ Space/Transportation
This futuristic space habitat is designed to self-assemble in orbit - MIT Tech Review
▶ Up Wing/Down Wing
Environmentalists are suing us out of addressing climate change - The Hill Opinion
▶ Substacks/Newsletters
AI for Non-Techies 🛠️: Top Tools for Meeting Notes - AI Supremacy
The rooftop energy revolution no one knew about - Exponential View
Paretotopian Goal Alignment - AI Prospects: Toward Global Goal Convergence
Why Does Ozempic Cure All Diseases? - Astral Codex Ten