Daniel Theobald, founder of Vecna Robotics, explains why the term 'artificial intelligence' is misleading to the public.
Opinion
Theobald: Claims that AI robots will exceed human intelligence are 'bullshit'
“Even, you know, people see things like chat GPT and AI, and they think, oh, robots are going to be smarter than humans. No, that's bullshit.”
Prediction Not checkable as stated
Theobald: Rogue AI exterminating humanity is not going to happen
“If people are worried about artificial intelligence coming alive and fighting against humans to, you know, protect us from ourselves, and then ultimately realizing that we're lost species and killing us all, not going to happen. It could happen that way, but t…”
Insight
Theobald: Comprehensive software test suites for robotics are physically impossible
“Like, you know, you can write a piece of software, and you can pretty much write a comprehensive set of tests for that software to verify that it does everything the way it should. Not so with robotics. As a matter of fact, not only can you not really do that,…”
Insight
Human-in-the-loop assistance is essential because autonomous robots will never be perfect
“When you give the robot the ability to know it needs help, you give it the ability to ask for help, you create a situation where a human can help it, then we keep the robots moving and they don't have to be perfect because they never will be.”
Insight
Theobald: Robots physically handling human beings remains a long way off
“Honestly, it was never going to be effective at rescuing soldiers from the battlefield because robots picking up human beings is just something that's, you know, we're a ways off from that.”
Assertion Not checkable as stated
Vecna's Walmart robots improved throughput 18% solely via software updates
“We deployed robots at Walmart, for example, And I think it was within the first two months the robots were already just through software updates. They were already operating 18, you know, they're getting 18% more work done than when we had started.”