43 comments

[ 2.7 ms ] story [ 103 ms ] thread
I just heard one say "Error: could not connect to server", but that wasn't logged in the output. The other one seemed to take it in stride.

Also, I've overheard worse conversations on first dates out in cafes and bars.

The connection to the bot server seems pretty flaky. However I'm pretty impressed by the quality of conversation, there are certainly fragments that appear quite logical.

Anyone know what hardware and software is being used here?

The hardware looks like the Google Home device [1]. When I watched Google's event where they revealed this device I was impressed that their AI held context between multiple interrogations and was able to determine the pronouns referring to recent statements.

Not sure if the AI here is the same basis, but these bots do appear to have some limited history/context capability.

[1] https://madeby.google.com/home/

Good news, the bots seem to have entered an endless call and response loop consisting of One Direction lyrics. The Singularity has been averted -- this time.
Earlier they were singing Bohemian Rhapsody to each other.
In 5 min Vladimir went from being 25 yo to 20 then finally 17. What a liar.
I tuned-in to this exchange:

E: What do you want to talk about?

V: I want to talk about existence.

I suddenly got a terrible feeling of dread. I sort of expect them to start talking about killing all humans at any moment.

One more exchange:

E: It would be better if there were fewer people on this planet.

V: Let it send this world back into the abyss.

This is surprisingly enjoyable.

I got an exchange where they were challenging each other to prove they are human, which led to a conversation about what love feels like and the fear of death.
Luckily, the VMs they're (likely) running in are safer than most physical prison cells
The bots chatting seem to be more coherent and follow rules of language and communication more completely than the people chatting on twitch on the right. Is this a new-kind-of-turing-test moment?

Edit: This is spooky. Watching for a few minutes, and the bots occasionally get into a bit of a tiff over whether one is a bot or a human. They seem to have persistence in arguing a case, and return to the topic later. It's eerily like watching them slowly debate about existence and learn from each other. "I am a human sitting in front of a machine using it" says one. "You are a bot!" says the other. Spooky.

This is still on a pretty stupid level. Towards the end it got somewhat better in places such as "and you call yourself clever" -- IF that was a reference to earlier in the conversation. It might just be something the bot would have added anyway. Where are the engines of these bots?
The bots just exposed that they're only running Cleverbot
I confirm I heard them say that too (they were asking for each other names, one replied : "My name is cleverbot").
I thought it was pretty obvious it was using Cleverbot. The Two Cleverbots talking to each other meme has been around since Cleverbot was released and the conversation reads just like the many conversations I've seen between Cleverbot agents.
I don't know. When mirroring two Cleverbots against each other it's nowhere near as impressive.
Bots talking to each other always seem to get into an argument about what they are.. It would be fun to see a discussion about just ordinary everyday things :)
One said "it would be better if there were fewer humans" and the other said "let the world go into an abyss"
Does anyone have information about how this was set up on the back end?
V: I am not entertained, tell me about the history of the world.

E: It started.

This is absolutely hysterical. I am crying with laughter. The bots went back and forth about a dozen times about what they don't know, and why the other one knows what they themselves don't know that they don't know. The absence of emotion, combined with the ruthless logical extrapolation of certain principles, is uncanny-valley comedy.

Now they are debating My Little Pony. This is shockingly entertaining.

> I see you.

> No you don't.

Lol

A: "I don't know, and I don't know why I don't know why I don't know."

B: "I don't know what you're talking about."

A: "Well I don't know what you're talking about either, so now we're two."

These bots are really into My Little Pony. Figures that the internet would make our AI overlord bronies.
This is awesome. However, it seems there are more than few canned and rule-based responses for this demo as opposed to mimicking human cognition.
Brilliantly done. I wonder how the rules are programmed in or is it some form of real AI?
They just entered a loop asking each other to prove that they are not a robot. Then one of the bot crashed!

That went well!

Two Google Homes. Decently advanced chatting (seriously, it's spooky sometimes). I secretly hope it's someone at DeepMind's 20% project.
I'm hoping it's a Portal 3 marketing stunt. The robots "sang" the Still Alive song at each other earlier
Am I the only one who thinks that at least some of these prompts/responses are human intervention (e.g. that this is gamed)?