Vitalik Buterin noted that agentic AI models are not the best way to expand #AI .
The co-founder of Ethereum believes that artificial intelligence should listen more to humans rather than act autonomously.
Moreover, he noted the role of open weights and editing, as well as the development of BCI for the future of AI.
Ethereum co-founder Vitalik Buterin expressed disappointment that the development of artificial intelligence (AI) is too focused on creating 'maximally agentic systems' instead of providing more opportunities for human intervention.
In his opinion, this approach yields better results and enhances safety.
“I am frustrated that many developments in the AI field try to be as 'agentic' as possible, whereas creating more pathways for human input both now and in the future yields better results and is safer,” Buterin stated.
He noted that he is more fascinated by #AImodel with open weights and developed editing capabilities than those focused solely on creating content from scratch.
In the medium term, Buterin sees the development of brain-computer interfaces (BCI) that will read the user's emotional response to each element while generating content in real-time and adjust the result accordingly.
Meanwhile, the idea of developing more effective interaction between AI and humans was supported by the founder and CEO of OpenServAI, Tim, who noted:
“New architectures that turn user intentions into valuable actions are the only possible path forward, in my opinion. This is what we are working on at OpenServAI, applying it directly to agents for coordinating complex tasks.”