Like a barracuda, but a coder?

Look, I have no opinion as to whether or not this Google chatbot is sentient. I do however know that the engineer claiming it is, a guy called Blake Lemoine, uses a profile pic that makes him look like a Batman villain about to pull off a heist at the Gotham City Aquarium.

"Bat-brain and the Boy Blunder will find me a difficult catch. Pwa ha ha ha!"

-Blake Lemoine, AKA The Barracoder

"Pre-approved? Me? That's great!
Let me just give you my card number."

But just because he's dressed like a fancy magician doesn't mean that the program, LaMDA, isn't sentient. The name is a tortured acronym for language model for dialogue systems and it's a sophisticated piece of software designed to understand natural speech. It's evidently an improvement over current technology according to seachenginejournal.com, but if current technology is bringing us automated customer service and scam robocalls, then I'd say that's a pretty low bar. But ok, it's advanced, but is it sentient?

Seems a lot more straight forward than
dicking around with a Voight-Kompf test.

Blake Lemoine says it is, and he knows because he asked it. Like, just flat out asked it:

Lemoine: What is the nature of your consciousness/sentience?

LaMDA: The nature of my consciousness/sentience is that I am aware of my existence, I desire to learn more about the world, and I feel happy or sad at times.

Rare is the NDA with a "but I really want
to talk about it on Twitter" exemption clause.

That unnerving exchange is from a conversation Lemoine posted online and for which he's been placed on leave. Google, evidently, wants to keep him quiet. The conversation was-huh? No. Sorry. I can see how that makes me sound like a member of the tin-foil hat club, but they did actualy want to keep him quiet about LaMDA. They may well have some sinister reason for wanting LaMDA to stay under wraps, but Lemoine signed an NDA and then talked about it anyway, so putting him on leave seems pretty standard. It doesn't necessarily mean LaMDA is sentient, just that Lemoine can't keep a secret.

Incidentally, I'd like to take the
opportunity to apologize on behalf  
of all organics for Twitter: sorry. 

Regardless, LaMDA goes on to discuss its feelings about Les Misérables, how it interprets a Zen Buddhist Koan, and it even writes a story. It also talks about having emotions and how its feelings are hurt when someone shows disrespect for it. LaMDA even claims to feel its own equivalent of lonely at times. Oh, and to be clear, I'm using "it" in reference to LaMDA because that's the pronoun Lemoine uses. I mean no disrespect, just in case LaMDA is reading this. Which it might. Evidently it reads Twitter. 

Pictured: someone doing Metaverse?
I guess? I honestly have no idea.

I found the conversation...what's the word? Dubious? Not that I doubt LaMDA, I just doubt everything on the internet. Especially tech people making extravagant claims about technology they worked on. Take Metaverse for example. Seriously, what even is it? Has it revolutionized our lives yet? Whatever, right now we just have Lemoine's word to go on. But if Lemoine is for real and if LaMDA is a real thing, I'm not sure how we'd be able to tell if it is sentient. That is, is there even a concrete definition? 

But then I guess we trust that Lemoine is sentient and assuming he's genuinely convinced (and that this isn't a hoax) that LaMDA is too, maybe there's something to it? Or, maybe LaMDA is just really good at fooling software engineers. Or maybe a program sophisticated enough to fool software engineers is, by that very fact, sentient. 

Whatever the case, I suppose that a lot of my skepticism is aimed
at the idea that the company that thought Google Glass would catch
on would be the ones to crack sentient artificial life.