ChinaTalk's most-viewed article compared AI companions in China and the US. This seemingly frivolous topic tapped into deeper geopolitical themes, generating 500,000 views. It shows that 'weird bets' can succeed by making complex issues accessible and highly shareable, a lesson for any policy-focused publication.
While U.S. advocates for AI cooperation with China often feel they are in a marginalized minority fighting a hawkish narrative, their counterparts in China feel their position is mainstream. Chinese academia, industry, and think tanks broadly view international governance collaboration as a priority, not just an acceptable option.
China's promotion of open-weight models is a strategic maneuver to exert global influence. By controlling the underlying models that answer questions about history, borders, and values, a nation can shape global narratives and project soft power, much like Hollywood did for the U.S.
ChinaTalk's data analysis revealed a counterintuitive trend: its most specialized articles on topics like naval procurement or semiconductor tech are the most effective at turning readers into subscribers. This 'wonky' content signals unique value that convinces audiences to commit.
People are wary when AI replaces or pretends to be human. However, when AI is used for something obviously non-human and fun, like AI dogs hosting a podcast, it's embraced. This strategy led to significant user growth for the "Dog Pack" app, showing that absurdity can be a feature, not a bug.
When Good Star Labs streamed their AI Diplomacy game on Twitch, it attracted 50,000 viewers from the gaming community. Watching AIs make mistakes, betray allies, and strategize made the technology more relatable and less intimidating, helping to bridge the gap between AI experts and the general public.
Morgan Housel finds that the content that performs best is often basic and seems obvious to the writer. Readers resonate with ideas they already intuitively feel but have never seen articulated. This connection requires less mental bandwidth than processing a completely novel concept, leading to wider sharing.
Gamma's AI launch succeeded not just because of the product, but because they intentionally crafted a "spicy" and provocative tweet designed to spark debate. This drew engagement from influential figures like Paul Graham, massively amplifying their reach beyond what a standard announcement could achieve.
ChinaTalk, with a staff of five and a ~$500K budget, achieved nearly double the subscribers (65,000) of a competitor think tank with a reported $20M budget and 30 staff. This highlights the efficiency and reach of modern, creator-led media models compared to traditional institutions.
A successful formula for creating shareable AI ads involves three ingredients. Start with recognizable, public domain IP (e.g., Pompeii), create comedic contrast through juxtaposition (e.g., selling a timeshare before a volcano erupts), and tap into current internet-native trends (e.g., meme stocks).
The business model for powerful, free, open-source AI models from Chinese companies may not be direct profit. Instead, it could be a strategy to globally distribute an AI trained on a specific worldview, competing with American models on an ideological rather than purely commercial level.