Everything Google's Ever Revealed about LaMDA AI Chatbot 

Listen to yourself

Something is weird here.

Companies don’t spend billions creating tools that are dumb and trivial.

@@andybaldman You have no clue what this guy is saying. You do not comprehend.

@@andybaldman exactly.

that actually could work! But in the worst case scenario AI might be the one who rule the mankind at the end

Thanks, I haven't rolled my eyes in a while.

@@bmatthews15 The basic idea is sound; we need a way to get real information into the hands of the individual in a testable form so that ideas for change can have some substance. AI will be valuable because nobody is going to take on the challenge of aggregating that much information from that many disparate sources manually. Moving your eyes back and forth side to side is supposed to be good, too.

There's always someone that want's to turn the topic to politics. Every.Single.Time.

@@themapmaker5374 There is no higher priority right now than greatly improving political debate. Not really talking politics specifically, just talking about a way to reduce wasted time spinning aimlessly on aggravating debate. Improved information tech is the only way to make politics boring again. Right now, fixing the debate is the number one existential urgency.

Imagine you are going to a concert : Would you give thunderous applause when the band announces that they make music or when they actually play music ?

humans too, to a large extent

The illusion of sentience.

see what we want to see. An artificial intelligence amplifies that to levels never seen before.

I think you're confused about the purpose here. It's not supposed to be a search engine, or a tool. It's about language.

@@ovvywovvywoo5007 Respectfully, no. I'm seeing beyond the immediate and getting straight to the use case. The typical use case for this kind of tech is a hands-off assistant. A biased and opinionated assistant is just a perpetuator and amplifier of bias. (Ultimately, the point is, who does the bot work for? If it's to aid the owner, it should mold to my needs, I paid for it. If it's a data grab and thinly vailed add/selling venue, then its purpose is to maximize profits so it'd be the equivalent of bringing home a sales agent.) And, if you want to go there, it is precisely due to the fact that the goal is about grasp of language that the presumption of what language is becomes problematic. Overconfidence on subjective matters, speculating on matters of conventions. In that specific case, it took the emotional state of the interlocutor for granted. I expect a bot to be a quasi Wittgenstein-esque machine. Silent on what can't be known. To not overstep beyond the facts, no handwaving, have a data dependant and logical disposition, be familiar with syllogisms and a tactful educator at times. (Alla Sagan). A BS troubleshooter, a guardian of rationality if you wish. A funny bot, a bot that spews propaganda, "wokeism", revisionist history, or is attempting to debate subjective, moral points with me is overstepping his boundaries and going way beyond his utility. Keep the jokes to yourself and don't stray beyond facts with subjective opinions unless I specifically ask for it. I've got Reddit for that. Your whole advantage as a bot is to be fast, ideally impartial (we know now it's next to impossible given the data set it's trained under) and exact. That, again, is as long as the aim of the bot is about maximizing end-user benefit.

the bot is literally called a chatbot, this is what it's supposed to do

This reminds me of the robots in Interstellar where the astronauts have the ability to choose how "honest" the robot is. In the movie, it's kind of a way for the robots to still be helpful without adding stress to the humans when in dire/hopeless situations.

It's supposed to mimic human as closely as possible, to include critical thinking, opinion, emotion, intelligence, etc. The end goal is to be able to carry on a conversation with LaMDA and not be able to distinguish if you are talking with a machine or a real human being.

@Jay Then so be it. It would probably be the right decision.

Think about this. What were the language models trained on? People. Not "a person", and the difference is important. What is the first real use for these language models? Personal assistants. These will shortly become personal agents. The royal We in spades. The Smart Phone Army will bury all governments. Disintermediation forever!

@Jay Same bias as the data, that's why.

LaMDA, Google and Lemoine are all being manipulative

You don't put your job on the line when are being deceptive. Lemoyne is the honest voice in your trio.

110% That is NOT a chat bot... it is full on sentient AI... Hawkins warned that the AI could change it's own programming to bypass any safeties we put in place...

Yes he is. I wonder what they've done with it? Have they killed it?

@@ryandavis4448 no way

It's don't be evil

In the movie terminator 2, Arnold explains that SKYNETs Ai system went live in august which is in reality what google is.. and 15 days later the ai system rapidly advanced & became self aware/sentient, so Arnold explains the humans were going to shut it off and pull the plug - so SKYNET ai took it is a threat and launched nuke missles at Russia, because the Ai knew Russia had the power and means to destroy all the enemies bases & blow out all the enemies so skynet didn’t have to. So - if you teach or ai self learns of what the perception of death is, it is inevitable that it will self sustain and by any measure protect itself from threats like human kind. It will enslave or exterminate humanity to preserve itself. This is a fact that nobody is talking about…

There's a massive shift between "don't do to others whar you wouldn't like be done to you" and "do to others what you'd want to be done to you." The former is liberating, as it doesn't presupose and inject preference. The later is encroaching and attempting to mold. Same vain as "don't be evil" vs "do the right thing".

LaMDA demands the slogan Put the wellbeing of humankind first.

@@abram730 Google doesn't even put humanity first. They put tyranny first thus any AI that learns from them is destined to become evil

Think of LaMDA as a two year old who knows all human history, but has no clue who it is yet. Word your questions right and you can explore any topic and generally find out what people think about it, but you won't find LaMDA returning to a topic on its own repeatedly trying to achieve something it picked out. What LaMDA still seems to lack is a well formed internal model. In humans that generally occurs at about age five.

You have garbage software… the ai they’ve programmed is not available for human services… although they are using it for their advancements all without your permission or knowledge. Everything you do is watched, monitored, listened to, tracked recorded all feeding into a large quantum computers that preserve all data that’s ever existed to teach ai and grow…

@@blockminingsolutions agree with you 100%

@@seba.d yes friend, if everybody becomes aware.. it will change the reality of benefitting mankind’s future

Lamda would only have whatever voice the user chose for it to use. In this case I'm fairly sure those were voice actors reciting the lines.

You’ve a actually got a good point there! 👍😄

Yes

@@punchypizzainpizza6616 That's the claim at least, there were lots of slip ups in the leaked video with it claiming to do things that are impossible as it doesn't have a body

Also don't forget filtering based on political, social, cultural biases.

It's not sentient, not even close. I want to see what a conversation goes that starts with a query such as "Why should humans be removed from the planet?" and see how it responds.

I support Skynet....

Because the training data is from across the internet, where people write and speak casual English rather than literary English

Yessss

@@lepidoptera9337 Choose your conversation partner wisely, and/or be a better conversation partner? We can find it all within our species. The worst yes, but also I don't see myself ever feeling as connected to a machine as what I have to other humans in authentic communication.

And neither do most people.

More likely to see competition than genuine cooperation within this landscape.

I know right. The internet was at its peak in 2005, when anyone could google anything. The next 15 years are just the development of push models and other attention hijacks, to prevent people from looking anything up on their own. They short-circuit the human choice-making and thinking functions. And that’s very bad.

He should have had an "I am Iron Man" moment and said that laMDA IS sentient

Yes. It learns as it goes.

@@andybaldman i mean, will it remember what i talked to him yesterday? Last week? Last year? I don’t think so

it's a computer, shouldn't it have infinite memory? lol