One of the biggest problems with frontier AI models right now is that they still can’t properly hear or watch the world around them.
Models like Fable 5.2 and GPT-6 Astra can solve incredibly complex problems, write code and reason at a crazy level.
But they still lack proper audio and real-time video support.
Even computer use often relies on taking screenshots, instead of continuously watching what’s happening. This is inefficient and slow.
Imagine an AI that can continuously hear sounds, watch things move and understand what’s happening around it, just like a human.
How can we expect AI to power advanced robots if it can’t properly see and hear the world in real time?
@OMondieuu Il y a surtout pas mal de postes au sein de la fonction publique qui ne servent strictement à rien et qui coutent une blinde tout les mois
ça m'étonnerait qu'on atteignent les 800 000, mais facilement les 50 000 en tout cas