- Great Technology is a technology and society newsletter from an independent journalist Alex Kadrowicz. To receive it in your inbox every week, you can Register here.
- A former senior Google engineer said he had conversations with LaMDA that are rich and full of personality.
- In the future, the chatbot may be able to talk to people about their personal interests and may even have the ability to teach lessons.
When I sat down with Blake Lemoine last week, I was most interested in chatbot technology he cried feelingly — LaMDA — from the subject of sensation itself. Personal questions aside, modern chatbots are incredibly frustrating (ever try to change a flight via text?). So if Google’s technology was good enough to make Lemoine, one of its senior engineers, believe it was a person, that progress was worth investigating.
As our conversation began, Lemoine revealed that Google had just fired him (you can listen to it in full Great Tech Podcast). And when I wrote the newsdone oneInternational history. But now, a week later, I can’t stop thinking about how LaMDA — consciously or not — might change the way we relate to technology.
According to Lemoine, LaMDA’s conversational skills are rich, situationally aware and full of personality. When Lemoine told LaMDA he was going to manipulate it, the bot replied, “that’s going to be crap for me.” When pressed about complex issues, he tried to change the subject. When he repeatedly told LaMDA how terrible it was and then asked it to suggest a religion to convert to, the chatbot said either Islam or Christianity, cracking under pressure and breaking its rule against privileging religions. LaMDA may not be sensitive, but it puts Delta’s Virtual Assistant to shame.
As LaMDA-like technology comes to market, it could change the way we interact with computers — and not just for customer service. Imagine talking to your computer about movies, music and books you like and it responding with other things you might like. Lemoine said it is under development.
“There are cases [of Lamda] which are optimized for video recommendations, instances of Lamda that are optimized for music recommendations, and there’s even a version of the Lamda system that they’ve given machine vision to, and you can show it photos of places you like to be and it can suggest destinations holidays that are like that,” he said.
Google declined to comment.
LaMDA can also connect to various APIs, giving it awareness of what’s happening in the world. Let’s see what a hypothetical – but reasonable – conversation with LaMDA might look like:
Me: Hi LaMDA, I’m in the mood for a movie tonight.
LaMDA: Okay, but you know the Mets are playing right now?
Me: Yeah, but I’ve had enough baseball for the week. So let’s go with something that’s been critically acclaimed, maybe from the 90s?
LaMDA: So you watched Pulp Fiction last week and you also enjoyed Escape to Dannemora, how about The Shawshank Redemption?
Me: Okay, let’s do it
LaMDA: Cool, you can rent it for $3.99 on YouTube, but since you’re subscribed to HBO Max and it’s available there, I’d recommend going that route. Here is a link.
“Regarding natural language”, Gaurav Nemade, LaMDA’s first product manager, told me, “LaMDA far surpasses any other chatbot system I’ve personally seen.” Nemade, who left Google in January, is teeming with potential use cases for LaMDA-like technology. These systems can be useful in education, he said, by taking on different personalities to create enriching new possibilities.
Imagine LaMDA teaching a physics class. He could read about Isaac Newton, play the scientist, and then teach the lesson. Students could talk to ‘Newton’, ask about his three laws, push him on his beliefs and talk as friends. Nemade said the system even makes jokes.
When released publicly, these systems may not be traditional chatbots, but avatars with likenesses, personalities and voices, according to Nemade. “The future that I would envision,” he said, “is not going to be text, it’s not going to be voice, it’s actually going to be multimodal. Where you have video plus audio plus a chatbot like LaMDA.” We may see these kinds of experiences debut within three years, he said.
Our interactions with computers today are mediated through interfaces created by technology developers to interact with machines. We click and ask, and we’ve become comfortable with this unnatural communication. But developments like LaMDA are closing the gap between machine-to-human conversation and may enable brand new experiences never before possible.
Some of Lemoine’s critics have said he thinks Google’s marketing is too naive. And it is indeed ironic that he brought more awareness to LaMDA than any Sundar Pichai Google I/O keynote he could hope so, even if Google would rather never hear from him again. Asked if it was a viral marketing ploy, Lemoine said: “I doubt I would have been fired if that was the case.”
Yet even those who disagree with Lemoine on the question of sensation—like Nemade and I—understand that there is something there. LaMDA technology is a big leap forward. It has serious drawbacks, which is why we haven’t seen it publicly yet. But when we get our hands on LaMDA, it may well change the way we relate to digital machines.