Jiajun Wu, an AI researcher, has been tweeting about various topics lately. In one of his tweets, he expresses excitement about winning the Best Systems Paper Award at #CoRL2023 for their work on RoboCook. He gives credit to the incredible team behind the project and shares a link to learn more about it.

In another tweet, Jiajun Wu introduces HEIM, a benchmark for holistic evaluation of text-to-image models like DALL-E. He emphasizes the need for transparent evaluation of these models' capabilities and risks.

Jiajun Wu also shares their latest work exploring whether robots can design tools to overcome limitations imposed by their morphology. They provide a link to their research page where interested readers can find more information.

Furthermore, Jiajun Wu introduces several other projects and works they have been involved in recently. These include introducing a benchmark for real-world object inverse rendering called 𝗦𝘁𝗮𝗻𝗳𝗼𝗿𝗱-𝗢𝗥𝗕 #NeurIPS2023 and presenting NOIR (Neural Signal Operated Intelligent Robots), which is a novel brain-robot interface system.

From analyzing Jiajun Wu's tweets, it is clear that there are trends related to AI that he frequently discusses. Some of these trends include grounded language learning with universal grammar rules, geometric concepts understanding in AI systems like GPT-4V, evaluation benchmarks for text-to-image models like DALL-E, and exploring robot tool design capabilities.

Overall sentiment analysis of Jiajun Wu's recent tweets suggests a positive outlook towards the direction AI is heading. His excitement and enthusiasm when sharing his own work or congratulating others indicate a positive attitude towards advancements in AI and its applications. However, it's important to note that sentiment analysis is subjective and based on the interpretation of the text.