Three Generational Shifts Shaping Technology & Culture

This article was originally published in Chinese by Runze Dai of Leiphone, translated and republished with the author’s permission. You can also refer to Momentum Works’ “Who is Temu” and “Who is SHEIN” reports for structured analysis and insights about these two companies. You can also read our book “Seeing the unseen: behind Chinese tech […]

+How to win Product Hunt 👏 every 👏 time

The green energy supplier is answering a third of all customer enquiries with AI

Italy’s GDPR restrictions were just a taste of what lies ahead for ChatGPT.

Election administrators across the United States say they’re concerned their offices will be targeted for fake Twitter accounts that'll confuse or mislead voters after the social media platform changed its long-standing verification service. Some election officials are trying to take steps to ensure that voters can tell the difference between the official election office account and any impostors that might pop up in elections this year or during the 2024 cycle. Twitter’s uneven response is causing concern about the potential for a flood of social media misinformation around voting and elections.

BookTok is the only profitable publishing trend of the year. How much of the profit goes to BookTokers?

Bailey Richardson (an early Instagram hire and current partner at People & Company) has studied ordinary people who started extraordinary communities. Here, she shares the nuts and bolts of how startups should approach community building, highlighting eight valuable lessons on how the best communities start, gain steam, and evolve as scale takes hold.

Greater access to the code behind generative models is fueling innovation. But if top companies get spooked, they could close up shop.

Recent work claims that large language models display emergent abilities, abilities not present in smaller-scale models that are present in larger-scale models. What makes emergent abilities intriguing is two-fold: their sharpness, transitioning seemingly instantaneously from not present to present, and their unpredictability, appearing at seemingly unforeseeable model scales. Here, we present an alternative explanation for emergent abilities: that for a particular task and model family, when analyzing fixed model outputs, one can choose a metric which leads to the inference of an emergent ability or another metric which does not. Thus, our alternative suggests that existing claims of emergent abilities are creations of the researcher's analyses, not fundamental changes in model behavior on specific tasks with scale. We present our explanation in a simple mathematical model, then test it in three complementary ways: we (1) make, test and confirm three predictions on the effect of metric choice using the InstructGPT/GPT-3 family on tasks with claimed emergent abilities, (2) make, test and confirm two predictions about metric choices in a meta-analysis of emergent abilities on BIG-Bench; and (3) show how similar metric decisions suggest apparent emergent abilities on vision tasks in diverse deep network architectures (convolutional, autoencoder, transformers). In all three analyses, we find strong supporting evidence that emergent abilities may not be a fundamental property of scaling AI models.