Jua's model is said to be superior to others due to its ability to ingest more information, making it larger by a multiple of 20x over Google's GraphCast. The company is also working on making the model more efficient to reduce operational costs. The startup's founders, Andreas Brenner and Marvin Gabler, have experience in the energy sector, fleet management software, weather forecasting, and deep learning technology. The company is also considering safety and reliability issues as it develops its model.
Key takeaways:
- A Swiss startup called Jua has raised $16 million to build a large "physics" model for the natural world, with its first application being in modelling and predicting weather and climate patterns.
- The company plans to target industries such as agriculture, insurance, transportation, and government with its model.
- Jua's model is said to be better than others because it ingests more information and is larger, and the company believes that understanding the physical world is key to achieving artificial general intelligence.
- The company is also aiming to build a more efficient system to bring down operational costs, claiming that their system uses 10,000 times less compute than legacy systems.