You can ask it questions (“What’s the difference between DPO and PPO?”), look up topics (“Graph Convolutional Neural Networks”), learn about papers (“KAN: Kolmogorov-Arnold Networks”), and research authors (“Yann LeCun”). Behind the scenes, it will try to find the most relevant computer science papers on arXiv, then synthesize their findings to generate a detailed, research-backed answer for you, all in a matter of seconds.

Unlike tools like ChatGPT, Emergent Mind is hyper-focused on computer science. It factors in citations and social media metrics to help rank papers (including Hacker News upvotes). It provides references so you know what papers the answers came from, is always up-to-date with arXiv (about 670k comp sci papers and counting), and encourages exploration using automatically-generated follow-up questions and topic links (similar to Wikipedia).

The tool is still fairly new and there’s endless room for improvement, but we wanted to share it with you all to get any feedback to make sure our product roadmap is aligned with what folks would find most useful.