A New Study from Harvard and Perplexity Finds AI Agents Perform 26 Minutes of Autonomous Work per Session vs 33 Seconds for Search
Quick Answer
A study from Harvard and Perplexity reveals that AI agents can autonomously work for 26 minutes per session, significantly outperforming traditional search assistants, which only manage 33 seconds.
Quick Take
A study from Harvard and Perplexity reveals that AI agents can autonomously work for 26 minutes per session, significantly outperforming traditional search assistants, which only manage 33 seconds. This research highlights substantial improvements in autonomy, time efficiency, and cost-effectiveness, indicating a broader scope of tasks that AI agents can undertake.
Key Points
- AI agents achieve 26 minutes of autonomous work per session.
- Traditional search assistants only manage 33 seconds of work.
- The study indicates significant gains in autonomy and efficiency.
- Broader scope of tasks attempted by AI agents is noted.
- Research conducted by Harvard and Perplexity highlights these findings.
Article Excerpt
From source RSS / original summaryA new Harvard and Perplexity paper uses matched-pair sessions to compare an autonomous agent with a search assistant. It finds large gains in autonomy, time, and cost, plus broader scope of work attempted. The post A New Study from Harvard and Perplexity Finds AI Agents Perform 26 Minutes of Autonomous Work per Session vs 33 Seconds for Search appeared first on MarkTechPost.
Reader Mode unavailable (could not extract clean content).
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from MarkTechPost
See more →Xiaomi MiMo and TileRT Push a 1-Trillion-Parameter Model Past 1000 Tokens Per Second on Commodity GPUs
Xiaomi's MiMo team, in collaboration with TileRT, has launched MiMo-V2.5-Pro-UltraSpeed, achieving over 1000 tokens per second decoding on a 1-trillion-parameter model using a single 8-GPU commodity node. This advancement significantly enhances performance for AI applications, making high-capacity models more accessible on standard hardware.