AMI Labs CEO says terms like AGI and superintelligence are not useful

In short: AMI Labs CEO Alexandre LeBrun says his company does not use labels like “AGI” or “superintelligence” because he thinks the terms are unclear.

What happened

Alexandre LeBrun, the CEO of AMI Labs, told TechCrunch that his startup avoids calling its work “AGI” or “superintelligence.” AGI is short for artificial general intelligence, a term people use to mean AI that can do many different kinds of tasks like a person. LeBrun said these labels are not well defined, and he does not find them helpful.

LeBrun runs AMI Labs, a startup co-founded by Yann LeCun, a well-known AI researcher. The company is working on “world models,” which are AI systems designed to predict what will happen next in the real world, not just in text. LeBrun compared it to everyday intuition, like knowing a glass will tip and spill if you push it off a table.

LeBrun spoke while visiting Seoul, where he said AMI Labs is looking for partners in robotics, manufacturing, and electronics. He argued that today’s robots often follow fixed routines and can struggle in open settings like homes or streets. He said safer robots will need better “context,” meaning a clearer understanding of what is happening around them.

AMI Labs raised $1.03 billion in March at a $3.5 billion pre-money valuation, according to TechCrunch. The company does not have a product to sell yet, and LeBrun did not share a timeline.

Why it matters

Many companies use big labels to describe their AI, but LeBrun’s comments highlight a simpler question: can AI work reliably in the physical world, where mistakes can hurt people. If world models help robots better understand their surroundings, they could affect jobs and safety in places like factories, hospitals, and homes.

Source: TechCrunch AI