Η Google DeepMind σχηματίζει μια νέα ομάδα για την ανάπτυξη μοντέλων τεχνητής νοημοσύνης ικανά να προσομοιώνουν τον φυσικό κόσμο. Διευθύνεται από τον πρώην συν-επικεφαλής της OpenAI (project Sora) by Tim Brooks and is based on Google's Gemini, Veo and Genie projects.
“DeepMind has ambitious plans to build massive models production that simulate the world", Brooks wrote in X. "I'm recruiting for a new team with this mission."
TechCrunch he says:
According to the job listings provided by Brooks in his post, the new team will work with Google's Gemini, Veo and Genie teams to tackle "critical new problems" and scale the models "to the highest levels of computation." .
Gemini is Google's flagship AI model suite for tasks such as analysis images and text creation, while Veo is Google's own video creation model.
Όσο για το Genie, είναι η άποψη της Google για ένα παγκόσμιο μοντέλο AI που μπορεί να προσομοιώνει παιχνίδια και τρισδιάστατα περιβάλλοντα σε πραγματικό χρόνο. Το πιο πρόσφατο μοντέλο Genie της Google, που παρουσιάστηκε σε προεπισκόπηση τον Δεκέμβριο, μπορεί να δημιουργήσει μια τεράστια ποικιλία τρισδιάστατων κόσμων με δυνατότητα αναπαρtreatments.
"We believe that training artificial intelligence on video and multimodal data is on the critical path to artificial general intelligence," reads one of the job descriptions.
Artificial general intelligence, or AGI (Artificial general intelligence), generally refers to an artificial intelligence that can complete any task a human can.
"World models will power many areas, such as visual reasoning and simulation, design for embedded agents, and real-time interactive entertainment."
According to the description, Brooks' new team will try to develop "real-time interactive production" tools as well as study how to integrate its new models with the company's existing multimodal models such as Gemini.
