Meta’s top AI scientist said that large language models such as ChatGPT will never reach human intelligence as they have a limited understanding of logic. Yann LeCun told the Financial Times that these models “do not understand the physical world, do not have persistent memory, cannot reason in any reasonable definition of the term and cannot plan . . . hierarchically.” He also said that depending on LLMs to reach human-level intelligence is not right as these models need the right training data to answer prompts correctly, thus making them “intrinsically unsafe.”
Read more: OpenAI creators say this is the secret behind success of ChatGPT: ‘Humans feel…’
He said that he is working on a totally new cohort of AI systems which aim to power machines with human-level intelligence which could take 10 years to achieve. This comes as Meta saw its value shrink by almost $200 billion after CEO Mark Zuckerberg pledged to increase spending and turn the company into “the leading AI company in the world.”
Read more: OpenAI to pull Scarlett Johansson soundalike Sky’s voice from ChatGPT
Moreover, AI firm Scale raised $1 billion in a Series F funding round that valued the startup at close to $14 billion and the French startup called “H” revealed it had raised $220 million.
Read more: Sundar Pichai on OpenAI launching ChatGPT 4o one day prior to Google I/O 2024: ‘Does not matter’
Experts have questioned AI’s ability to “think” like humans as Akli Adjaoute, who has spent 30 years in the AI field said, “AI does not have the ability to understand the way that humans understand.It follows patterns. As humans, we look for patterns. For example, when I recognize the number 8, I don’t see two circles. I see one. I don’t need any extra power or cognition. That’s what AI is based on. It’s the recognition of algorithms and that’s why they’re designed for specific tasks.”
{{/usCountry}}Experts have questioned AI’s ability to “think” like humans as Akli Adjaoute, who has spent 30 years in the AI field said, “AI does not have the ability to understand the way that humans understand.It follows patterns. As humans, we look for patterns. For example, when I recognize the number 8, I don’t see two circles. I see one. I don’t need any extra power or cognition. That’s what AI is based on. It’s the recognition of algorithms and that’s why they’re designed for specific tasks.”
{{/usCountry}}