Drones have many uses in surveillance and surveying, industrial inspection, aerial photography and filmmaking, package delivery and even dynamic light shows. Their fundamental ability to carry cameras, sensors and other equipment has opened up these diverse applications. Now, research published in the International Journal of Intelligent Machines and Robotics has explored another: the safe removal of moss and lichen from sloping roofs, an otherwise hazardous job for a worker.

A robot meant to inspect a coastline, a flooded street or a wetland can't count on one kind of ground. It might need to crawl over a hard surface, climb a bank, scale a curb or a step, then slide into open water, often within the same few feet.

For the modern LLM era, the standard recipe has rested on two pillars: a fixed tokenizer, and next-token prediction. Text becomes tokens, and the model learns by guessing the next one. Two lines of work are now challenging that recipe from opposite sides. Meta’s Byte Latent Transformer (BLT) asks whether we need tokens at all. […]