Deep Research
Hosted by
´◡`
´◡`The major bottleneck to solving (general) problem-solving? Giving foundation models access to the public and private knowledge of humanity.
The next frontier in AI is more than just model improvement: it’s access to knowledge. This hackathon is about building the great deep research systems of the future.
Join us to co-invent new techniques to give AI access to the world's knowledge.
user data, Google, and more. They use LLMs to generate and refine queries, looping on results to surface deeper insights inaccessible to single-shot searches.
Modern search engines like Perplexity, Exa, and Glean now power these agents. Embedding models such as BERT and Contriever enable engines to guide indexing, write their own queries, and integrate retrieval directly into long-form generation. The RAG era has evolved into the era of the search agent.
Report writing now includes code-generated graphs, tables, and structured analysis. These long-form outputs push AI toward deeper coherence and reasoning, producing artifacts—papers, books, and web-native docs—designed to be cited and reused.
Guest List
ᵔ◡ᵔ
ᵔ▿ᵔ
⚆◟⚆