Vision Language Models (VLMs) have recently been adopted in robotics for their capability in common sense reasoning and generalizability. Existing work has applied VLMs to generate task and motion planning from natural language instructions and simulate training data for robot learning. In this work, we explore using VLM to interpret human demonstration videos and generate robot task planning. Our method integrates keyframe selection, visual perception, and VLM reasoning into a pipeline. We named it SeeDo because it enables the VLM to ''see'' human demonstrations and explain the corresponding plans to the robot for it to ''do''. To validate our approach, we collected a set of long-horizon human videos demonstrating pick-and-place tasks in three diverse categories and designed a set of metrics to comprehensively benchmark SeeDo against several baselines, including state-of-the-art video-input VLMs. The experiments demonstrate SeeDo's superior performance. We further deployed the generated task plans in both a simulation environment and on a real robot arm.

From small-scale home systems that take the guesswork out of growing for beginners to commercial operations and predictive software, the technology has the ability to change how the world grows, sees and uses cannabis, from hemp and CBD to marijuana, along with the other crops we're more accustomed to like cucumbers and tomatoes. "Right now, it's kind of an obscure topic," said Nathaniel Morris, founder of William Bond Ai in Ontario, which uses AI to train machines to grow cannabis. "It's not going to be obscure for long. Morris said AI imaging can find impurities in the plants, like mold, or a male plant that can impact the entire crop, better than the human eye, once trained to do so by an expert. The technology isn't really new -- but adapting it to fit the needs of cannabis growers is.

Powered by AI and Machine Learning technology, Seedo enables anyone to grow anything with no experience and the same amount of space you would need for a mini-fridge. Founded in 2015, the Israeli AgriTech firm's self-contained device generates "high yields of lab-grade, pesticide-free herbs, and vegetables," states Seedo's website. But the company is well aware that the herb that little Seedo will most be responsible for growing, is cannabis. In fact, the device's impressive growing abilities have been translated from the knowledge of the company's founder, retired expert cannabis grower Yaakov Hai. Seedo's biggest market is in the United States where growing and using cannabis recreationally is now legal in 11 states in the USA and in 22 states medical cannabis has been recognised as an effective treatment for numerous health conditions including PTSD, depression, chronic pain and for those undergoing cancer treatment.

An Israeli agtech company called Seedo might have the solution for the challenges of urban agriculture in vulnerable areas such as the Caribbean, that struggle with environmental and climate factors that lead to crop loss. Latin and America and the Caribbean is the most urbanised region in the world with up to 80% of the region's population residing in cities (UN-Habitat 2012). While urbanization is an important element of economic growth and modernization, the diminishing ratio of food producers to food consumers in urban settings negatively impacts local food systems, causing populations to be more susceptible to non-communicable diseases, obesity and undernourishment. Urban farming practices such as rooftop gardens, community greenhouses and vertical farms have provided an alternative to rural agriculture, but given the high cost of urban land, space and size limitations, non-conducive environmental conditions and limited human resources, these methods have not been without their challenges. Vertical farming's "closed and controlled" approach has been successful in eliminating the risk of insects, pests and diseases that are prevalent in traditional agricultural systems but the infrastructure required has typically been cost-prohibitive and highly reliant on fossil fuels (solar power is typically not enough).