Perplexity Frontier Research Details Six Directions from Multi-Agent to Frontier Evals
andrewgwils · x · 2026-11-16
Parts 3-4/5 of the Perplexity Frontier Research thread lay out the philosophy—going beyond black-box predictions to systems that produce fundamentally new knowledge—and six research directions: multi-agent orchestration (automated research scientist, co-training, agentic search), pre-training (open weights, new architectures, synthetic data), continual learning, reinforcement learning (verifiable rewards, environments from real usage), systems, and frontier evals. The team says it's a place where "you could propose the next transformer."
Related event: Perplexity Launches Frontier Research to Build Autonomous Research Systems(6 posts)→
More from Research
- Andrew Wilson: AI research beyond black-box prediction could yield the next Transformer — andrewgwils · 2026-11-16
- Perplexity Launches Frontier Research to Openly Advance Models, Data and Algorithms — andrewgwils · 2026-11-16
- Jon Barron's CVPR23 Talk Revisited: A 1950s Speed Forecast That Overestimated the Future — jon_barron · 2026-11-16
- Google opens SatWeave and PARDIG, the foundational datasets behind WeatherNext 3 — ymatias · 2026-11-16
- Anthropic: agent memory works best by forgetting selectively—notes beat 400+ archives 4x in Pokémon Red — The Decoder · 2026-11-16
- Plasma proteomics outpredicts HbA1c for 10-year cardiometabolic disease risk in Nature Metabolism study — EricTopol · 2026-11-16