How analysis dies
https://benn.substack.com/p/how-analysis-dies When we read some argument or piece of analytical work—be it a proposal for a business strategy, a pitch for an investment, or an op-ed making a political point—we typically fancy ourselves as rational thinkers and impartial jurors. But we aren’t good at being either. As Randy Au called...
Discovering latent knowledge in language models without supervision
https://arxiv.org/abs/2212.03827 https://github.com/collin-burns/discovering_latent_knowledge How can we figure out if what a language model says is true, even when human evaluators can’t easily tell? We show that we can identify whether text is true or false directly from a model’s unlabeled activations. Existing techniques for training language models can be misaligned with...
A deep-learning model of prescient ideas demonstrates that they emerge from the periphery
https://doi.org/10.1093/pnasnexus/pgac275 https://osf.io/preprints/socarxiv/j24pw/ Where do prescient ideas—those that initially challenge conventional assumptions but later achieve widespread acceptance—come from? Although their outcomes in the form of technical innovation are readily observed, the underlying ideas that eventually change the world are often obscured. Here we develop a novel method that uses deep learning...
Individuals prefer to harm their own group rather than help an opposing group
https://www.pnas.org/doi/10.1073/pnas.2215633119 Group-based conflict enacts a severe toll on society, yet the psychological factors governing behavior in group conflicts remain unclear. Past work finds that group members seek to maximize relative differences between their in-group and out-group (“in-group favoritism”) and are driven by a desire to benefit in-groups rather than harm...
What does Meta AI's Diplomacy-winning Cicero mean for AI?
https://garymarcus.substack.com/p/what-does-meta-ais-diplomacy-winning Human-level play in the game of Diplomacy by combining language models with strategic reasoning (FAIR et al., 2022) https://www.science.org/doi/10.1126/science.ade9097 Diplomacy, a complex game that requires extensive communication, has been recognized as a challenge for AI for at least fifty years. To win, a player must not only play strategically,...
Acquisition of chess knowledge in AlphaZero
https://www.pnas.org/doi/10.1073/pnas.2206625119 We analyze the knowledge acquired by AlphaZero, a neural network engine that learns chess solely by playing against itself yet becomes capable of outperforming human chess players. Although the system trains without access to human games or guidance, it appears to learn concepts analogous to those used by human...
Data's day of reckoning
https://benn.substack.com/p/day-of-reckoning Facebook and Google, for example, can clearly make a lot of money by using their data to serve better ads. Netflix can clearly improve its business by using viewership logs to better promote the content that they offer. Uber can clearly make their service better by shortening wait times....
Traits you can change, and traits you cannot
https://staysaasy.com/leadership/2022/11/06/traits-you-can-change-and-traits-you-cant.html In my experience some personal traits can be changed, and some cannot. Mistaking one case for the other causes two of the most common and costly talent retention mistakes: Failing to believe in a rising star, because you thought that their weaknesses couldn’t be addressed. What it looks like:...
What to watch in AI
https://www.generalist.com/briefing/what-to-watch-in-ai Copilot for everything. AI is already streamlining illustration, writing, and coding. It may soon become an assistant for all knowledge workers. In the future, we may have versions of GitHub’s “Copilot” feature for lawyers, financial analysts, architects, and beyond. Tracking value accrual. As AI startups often rely on publicly...
What hunter-gatherers can teach us about the frustrations of modern work
https://www.newyorker.com/culture/office-space/lessons-from-the-deep-history-of-work Drawing from this field work, Woodburn argued that hunter-gatherer communities like the Hadza often relied on what he called an “immediate-return” economy. As Woodburn elaborates, in such a system, “People obtain a direct and immediate return from their labour. They go out hunting or gathering and eat the food...
About
a collection of excerpts from interesting research, blogs, events, etc...