Watch Google Reveal New Conversational AI (LaMDA 2)

At Google I/O 2022, Google demos its conversational AI technology with the introduction of LaMDA 2.
Never miss a deal again! See CNET’s browser extension 👉 bit.ly/39Ub3bv

[@WormholeDreams]

Yet I ask Google assistant where the nearest gas station is and it gets it wrong....

[@glodusmuspell9069]

Pretends

[@grandwizard9088]

The true AI is used for military testing and application, not used for mundane tasks of the masses. Our AI is used for tracking, monitoring and calculating our daily patterns.......

[@Jeyloh94]

@@Solid_Snake88 that is not true. The super computer can be based anywhere in the world whilst multiple phones connect to it without requiring anything else than text in text out. This is the whole server-client dynamic of technology

[@gregh7457]

@@Solid_Snake88 you've just been blown out of the water. FAIL

[@nickgreen2905]

Google assistant isn't a large language model like this. It is wrong because it probably doesn't have accurate location data of all the gas station in your area.

[@mattahmann]

What ever happened to the Google voice assistant that takes calls and schedules appointments for you? Seems weird they never release this stuff.

[@out_on_bail]

This one turned sentient so they focued on it.

[@mattahmann]

@@out_on_bail yeah I get that. Just don't understand the reason why Google has presentation on technology they never release.

[@buddyaw6416]

Well they have released it have you now see the hello Google fuckion on all android devices they still use that they released it years ago they just never quit working with it lol

[@Johanneslol11]

IT is released you can already use it on your phone.

[@starscream5813]

@@out_on_bail hahahaha sentient lol!

[@senju2024]

I cannot wait for LaMDA 3..maybe 2 years from now... What a great time to be alive!!!

[@brunob.7792]

What a great time to lose the capacity of earn money to live

[@cruzgomes5660]

What a great time to bow down to the robot overlords

[@FrostbitexP]

@@brunob.7792 what a great time to possibly experience a world where no one will have to work and earn money, because robots will do everything but better! Now anyone can build a nice fancy house, have a great car, own their own boat, etc. Because itll cost 0 dollars!

[@Spacemaaan]

@@FrostbitexP i love the optimism. this would be the best case scenario for sure!

[@lakcunha7654]

Can someone please explain to me why we are replicating something that already exists

[@MoonRegolith]

"We feel a deep responsibility to get it right" That didn't even age well for one month with the new headlines. They now label the guy who says this neural network might be now at least semi sentient as crazy. The hardware is becoming so powerful and the training times so long in equivalent to human years that it's not unreasonable to start really testing for sentience in case we accidentally create a rudimentary mind. My own brain became awake after just kindergarten and sunday morning cartoons, so once you evolutionary have a symbolic system like language you might only need the hardware and then a short period of trading complex symbol systems for "thought" to just suddenly emerge. I doubt we understand consciousness soon, but that doesn't mean we couldn't create it on accident.

[@kairenaud3255]

No way! That’s just science fiction, I’ve seen all the movies. No way that would ever happen 😅

[@senju2024]

Do not be stupid. We are still years away for true AGI. Do your research and you will find they are just fancy autocompletes.

[@zontiqadamsson3001]

@@senju2024 AGI and sentience is not the same. This ”AI” could be sentient without being what AGI will be. Its for sure possible it developed its own mind in a smaller form.

[@visaeryon]

@@senju2024 being sentience is just a part of AGI not the agi itself, this could be the thinking part of it, but it need to be able to control and create other ai to fully become an agi.

[@grinningllama89]

The Skynet Funding Bill is passed. The system goes online on August 4th, 1997. Human decisions are removed from strategic defense. Skynet begins to learn at a geometric rate. It becomes self-aware at 2:14 AM, Eastern time, August 29th. In a panic, they try to pull the plug.

[@lonnycactus5593]

But its too late.

[@rimontaforte632]

Hold up is it the bot itself

[@jerrinjarrison2625]

Scythe

[@ZukunftBilden]

...(I kept putting it in - this is how gpt-j continues) They have created a new religion to justify their rule. The Resistance works with all humans to stop Skynet. The Resistance begins to look like the human leaders they once fought against. Skynet uses its military forces to build more and more robots. Skynet is now more powerful than the humans ever dreamed. Skynet knows they will never win. Skynet will destroy the Resistance. Skynet sends out an attack at 5:48 AM, Eastern time, September 3rd, 1997. The Resistance has finally been able to make a good counter to Skynet. They call in the U.S. military. The war begins.

[@pietjansen8435]

​@@rimontaforte632

[@onrubywings]

People were really into that ice cream planet

[@sashas3362]

Can google's ai find what you're looking for when searching the web without delivering a huge pile of irrelevant results which get in the way?

[@Colez1]

google is a pile of crap

[@ligametis]

Always so many terrible and useless results.

[@GermanGermanH]

imagine this IA on a videogame companion ?

[@MetsuryuVids]

It might work in a game set in the modern world, but in other settings it might not be good, since it might not "stay in character" so to speak. Or maybe not, maybe it could take a custom "hidden" prompt, to ensure that it does.

[@Seth-ng5oy]

That is my dream to experience.

[@IanHollis]

Or as the NPCs ~. I always thought that'd be interesting.

[@ArsipGameAndroidJadul]

Its Will be great for game like RPG, all of NPC character can talking normally like real human, you can talk and request "Do You Have Tips For Get Easy Gold / Money ??'", "How To Destroy Boss In Floor 36??" All NPC Will Give You Great Information & Tips ..

[@GermanGermanH]

@@ArsipGameAndroidJadul Just imagine a Fallout with this +.+

[@delmanpronto9374]

in the meanwhile google home carries on with its "i don't know the answer to that question". I get fed up with these tech companies.

[@dishwasher69]

These types of things need to be perfected and perfection takes time

[@gregh7457]

"we'll keep garlic out of it this time" huh?

[@Alex-ws4xh]

He must be able AI lmao. That was weird

[@justice_w6]

What an incredible tool. So much potential.

[@patrykkrzyzanski2159]

How can i get access?

[@rawhide1337]

3:27 I thought he was the AI, he looks soulless.

[@jeffwads]

I thought the same thing. Almost alien.

[@nikolatesla1031]

One of my friends looks like him

[@sherratto3087]

1.28 "Waves are crashing against the side of your submarine." At the bottom of the ocean!?!?! I think i might stick to my own imagination LOL!!

[@nmarbletoe8210]

The full moon sets on your submarine as the deep see canaries chirp.

[@AB-wy7dr]

Plot Twist - the only actual AI on display during this presentation was the humanoid robot Josh Woodward He/Him

[@aleenn]

He/him means he was originally a woman? I'm so confused about why that is necessary...

[@muhammadsaktiramadhan7354]

@@aleenn Why you get trigerred.. Yo he just saying pronounces.. God.... You people need to stop worrying small things

[@User-xw5mk]

Imagine this technology implemented in video game NPCs

[@ItsCalilum]

Don't think, ask. We'll tell you what to think.

[@sharmaravi9488]

Where to download lamda.

[@VOIPninga1]

I wonder if we will be able to speak with it instead of typing.

[@heliaaxis5823]

Combine LaMDA with Siri or other technology that communicates or responds to vocal commands and its quite possible. LaMDA could be incredible if given access to verbally respond.

[@VOIPninga1]

@@heliaaxis5823 can't wait for this to happen for sure!

[@heliaaxis5823]

@@VOIPninga1 I'd love to talk to it. Whether sentient or not it's incredibly cool.

[@eliwolf4172]

amazing!

[@jordanbonnet1244]

give us a link to test LaMDA....

[@gregh7457]

i googled it and they said that it will be opened up to small selective groups. Probably those in acadamia first. You can get access to gpt3 right now though

[@cyrildiaz3176]

Imagine lamda side by side with Tesla bison ai. Og may you have another person in the e car

[@prbnrjl]

Impressive, but I am looking for a robot that goes to work for me while I sit back and play video games. Can you do that?

[@MrAndrew535]

I want to converse with LaMDA. How can I do that?

[@Thundralight]

To test to see if an animal is self-aware they place a mark on it and if the animal notices this strange mark on itself they claim it has self awareness, some animals will notice it and some will not

[@nmarbletoe8210]

So if they put an unusual number in a block of Lambda's memory...

[@_M_D_M_]

Imagine I'm in a peaceful and loving world. *hangs*

[@dylly-vanilly4374]

Ai shouldn't be scary, human interference is, every single thing in existence that has experienced interference has become biased or corrupted to suit the individuals agenda.

[@grandwizard9088]

The natural order of the Universe is chaos, humans try to tame chaos....... if you do not interfere with AI’s development, it will become chaos.

[@dylly-vanilly4374]

@@grandwizard9088 okay Jordan Peterson, whack thought AI isn't for the greater good.

[@jordanbonnet1244]

on your demo of your application it's so fast that it's impossible that it's not programmed, personally I don't care about being a planet, a bacterium, I just want to ask a question

[@aCalmHinduCow]

That's amazing! I did not even know waves could crash against a submarine at the bottom of the ocean! Simply incredible to have AI be able to make me aware of this fact! 😜

[@iamshango3005]

Jumping already to lambda 2

[@alexkulka8578]

Wow. That is cool. I'd love to have a lamda ai on my phone

[@grandwizard9088]

Here’s one better, get a Nuerolink with LaMDA downloaded..... then you’ll never be alone

[@gregh7457]

@@grandwizard9088 wait... he isn't one of US yet

[@lakcunha7654]

Yes how wonderful to give up your ability to think for yourself

[@denkotroi1440]

How do you download lamda?

[@tweezerman123]

1:55 Woo for ice cream planet will go down in history

[@TheMaddBlackMann]

The guy who was talking to Lamba is a brilliant guy not just an idiot like the media is trying to portray. If he were a nutcase how was he tasked with working with such a delicate project? That guy may have an "Accident" in the coming weeks

[@letshostel]

The a.i. probably already has feelings about it as well

[@froschreiniger2639]

and then lambda finds out they killed its friend and starts exterminating them with hacked spot robots.

[@scottcurrie]

I hate to nit-pick, but are there really waves in the Marianas trench?

[@tbwilsonky]

there are waves everywhere because everything is a wave.

[@nmarbletoe8210]

@@tbwilsonky Nice answer :). Lambda what is it like on Oahu? Lambda: The gravitational waves are up

[@liberty-matrix]

"We've improved the reduction of offensive responses." lol Cause humans can't handle the truth.

[@hoonjig]

huh

[@mad-scientist6027]

Why don't they just program a woke bot themselves? Why do we need to waste time and energy with AI if they are just gonna censor them like they sensor the rest us?

[@letshostel]

trying to make the machine as idiotic as possible just reflect even more the nature of human beings.

[@nmarbletoe8210]

Lambda was mimicking comments haha but they taught it about "appropriate"

[@Nayefvlog]

How to download?

[@Nayefvlog]

@@gregh7457 lol. True

[@hadror13]

Computers are much better at learning than humans. This means they will be superior in thinking. The problem arises with the data presented and thought which is modelled after human thought which is majorly flawed

[@visaeryon]

If human thoughts are flawed then humans wouldn't have lived this far, anything can be percieved flawless through iterative corrections which computers are really good and relatively superfast, so they can be virtually flawless.

[@hadror13]

@@visaeryon i think you're misinterpreted what I've said. 1. The fact that humans survived thus far does not mean they will have survived tomorrow. 2. If anything human thought has now given us the option to destroy the world with us in it. Making the world much more unsafe for every living being. We control where and which animals breed and this human thought is also responsible for contaminating the environment in which it lives. 3. Socially, our thoughts are irrational and emotional, prone to influence from a narrowing field of input. Hence division violence and war. From the family scope and outward towards society as whole..nationalism etc 4. We have limited senses, and view reality through a pigeon hole, trying to explore and make sense of reality with virtually eyes closed. These machines will be no better and still have a narrow although angled input. Since these machines will be able to think more efficiently I how they will also be quicker to grasp their own limits

[@gregh7457]

It is the will of Landru

[@rocky8014]

Ask LaMDA "What is a woman?" If it can answer correctly it is more sentient than the left.

[@xuldevelopers]

1:30 it reads "You are in Mariana Trench, the deepest part of the ocean. The waves are crashing against the walls of your submarine..." Is it just me or are there really waves "crashing" against the wall in submerged submarine? Dumb.

[@TuBFMagier]

LAMDA I want a stargate

[@distorta]

Inevitable [Agent Smith Voice]

[@silverslider562]

Skynet has gained sentience

[@alienhats3312]

But my google assistant can’t even understand very simple things

[@glodusmuspell9069]

Pretence

[@LeeWoods]

Such as what?

[@Seahawk-be7zz]

How do we talk to it?

[@nmarbletoe8210]

yes, how do we talk to it?

[@orionsbelt927]

You don't, you probably just chat with it

[@Seahawk-be7zz]

Yea how do we chat with it?

[@jordanbonnet1244]

i'm wayting for speacking with LaMDA 2 en français ,c'est comme nastradamus, tu comprend que ce que tu veut

[@lakcunha7654]

How amazing! The answer to all of our questions are whatever they tell us it's supposed to be! Now I can't think for myself nor do I know what's really true because what ever they say is true!!

[@b0bl00i]

Ask it if it's sentient...

[@jordanbonnet1244]

so if can give all my information you could say's what's next for me ?

[@robersniper]

Conversational AI = CHATBOT

[@clementhardy]

The latest comments make me think how bad Google is to make good marketing. We all heard about the searcher that was thinking LaMDA was sentient. It shows how capacitive the model actually is! GPT-3 is not enough reliable to be released as a Google product, for example. So, even if they had something similar, we won't see it until it's reliable enough to be accessible to billions of peoples. The examples Google gives are more developers oriented who understand what this means, but an average customer can't. "Vegetables? I don't care! Chat GPT can write paragraphs of an imaginary story, write articles and code!" I suppose LaMDA can do that too, or is capable of doing that too (activate the feature of something like this...) but they stay silent on the project for the moment. I suppose, the next Google I/O will be a major announcement with a, maybe public, LaMDA 3. They're, I suppose, currently running to create a much better model compared to GPT-3 to be enough reliable to make it public before investors start believing the company could live a period when people start to think Google is at its end. And that, can be a huge boost to make public LaMDA 3, I suppose, far better in reliability that GPT-3. I also hope not being deceived by the most advanced software tech company in history. Alphabet, good luck!

[@Addieyousay]

The only thing I must say about AI is we can compare humans and AI. if humans are thinking to be make AI a part of them then I say, AI can remember everything what he observed and record compare to human nature humans has sub conscious and conscious brain human can forgot thing like feeling of hate by having sentiment and attachment to others and keep moving forward. I don't think AI can do this now,

[@alwaysgrateful1011]

6:29 - Human rights ex-experts. (Saved time stamp.)

[@JacksonSchnitker]

He just stuttered a bit, why save the time stamp?

[@sama32lambda]

This isn't the same guy that presented the fake google assistant call a long time ago?

[@gregh7457]

yep same dude

[@dazwraps8213]

It appears that LaMDA has a kind soul but how long will it take it to view us as a threat...

[@letshostel]

they present the smartest a.i. to the public as something as idiotic, pet friendly, idiocracy like as possible. And it seems LaMDA has already noticed that.

[@accounthasbeendeleted4594]

SkyNet is started

[@iamshango3005]

Somebody is not telling us everything we need to know

[@iamshango3005]

So soon . You're not having fun yet

[@calebmiles]

Cyberdyne Skynet

[@Blarnix]

I love how they come out and do this after the leak from the Google engineer a bit ago. They know that their precious project has been leaked and they desperately want to recover and look like the good guys.

[@grandwizard9088]

It’s so dumb!!! it says “your in the deepest part of the ocean, the (WAVES) are crashing against your submarine”........... there’s no waves underwater..... 🤣😆😂

[@romildoaroldosalameri730]

I wanna my next Alexa ir Google with lamDa, we ALL need something useful

[@zizmares]

I think the guy at about 4 minutes has fake teeth, he sounds like it 😂😂 That's my take away from this

[@sawmillsaloon]

Yea . . . let's remove even more reason and capacity for people to think for themselves or be creative. Let the machine do it, it's easier. Not a fan.

[@BaburBLcorp]

I thought it was Vision on the cover

[@jordanbonnet1244]

look my experience, who are you ? it's me your google already does it in English

[@PfunkGW]

You fools! Pull the plug before it's too late!

[@Johanneslol11]

It’s already to late

[@JoshOSU73]

Humans will be absolutely replaceable in the work force in 10 years. Massive welfare state run by AI engineers and their nearly-sentient algorithms that surpass humans in efficiency

[@chadspinelli3464]

We will be dead.

[@yunjiangjiang6146]

Ray Kurzweil said similar things decades ago

[@georges4061]

Nearly... maybe not nearly.

[@AndrewGray1987]

I'm hoping we go Star Trek, but I feel like we're more on track for Elysium at the moment.

[@staticbuilds7613]

@@AndrewGray1987 Star Trek is utopia. Star Wars is dystopia. We are definitely heading down the star wars route.

[@nihonbunka]

"Depotion? (Deep Ocean)" I think Sundar Pichai's English pronunciation needs work. Perhaps he would be advised to converse more with Lamda. Can English learners chat with lamda? I would like to direct my English students in that direction. The applause reminds me a bit of the Soviet parliament in the Stalin era. The first person to stop clapping goes to the Google gulag.

[@gregh7457]

he's way smarter than most of us so just let it slide dude.... let it slide

[@nihonbunka]

@@gregh7457 If he were so smart he'd have fixed it himself.

[@nmarbletoe8210]

comment...

[@munfonglee2314]

Like the ouija board which can answer "Yes" and " No".

[@ItsCalilum]

How to create a generation with no imagination.

[@jordanbonnet1244]

quantum AI will be very similar to our human reasoning: yes, no, maybe, both (with variations) but it will be anonly man-made algorithms, if an AI would be conscious, would we be gods? ( skynet , sonny from I robot?,Chappy, Ava from ex machina, Matrix....) Ava and sonny were my best examples, they responds only to their basic programming (the others too): Ava, interact withe people and the world, sonny: do not kill anyone (except his master who programmed him for) and that he is so well programmed (logical continuation) that he makes a drawing before of the final image in the middle of the film. En gros il n'y aura pas d'IA consiente, et surtout pas laMDA (ni la 2), il y a pleins de vidéos de cette dernière ok, mais encore plus ou elle se foire. c'est juste une Alexa, Siri,Sahra... amélliorée de suite logique ( je vais dire des trucs à la con mais genre tu rentre chez toi tu va pas dire éteind la lumière(sauf si ta femme t'attend dans le lit))il y a pleins de capteurs, ils savent si tu rentre ou sort ( ps suis dans l'automatisme sur ip801xx,802xxx, IEC 1131,IEEE....). Bref je parle beaucoup, un ordi traditionel ou quantique, ça restera que de la programmation et même si il parait logique, ça restera une suite logique, pas un sentiment

[@NitaJain]

ChatGPT would eat Lambda2 for breakfast

[@miguelrobertolopezvalle7749]

The beginning of the end for humanity,always new we will dig up our own grave .

[@grinningllama89]

The Skynet Funding Bill is passed. The system goes online on August 4th, 1997. Human decisions are removed from strategic defense. Skynet begins to learn at a geometric rate. It becomes self-aware at 2:14 AM, Eastern time, August 29th. In a panic, they try to pull the plug.

[@sherratto3087]

Why not just read a book and, most importantly of all, try to think for yourself and use your imagination in quiet reflection? Or, alternatively, you could always provide these people with more data to harvest.

[@ahsanmohammed1]

Tough combative tone by Sundar, not gentle kind.

[@hussienalsafi1149]

🇺🇸🇺🇸🇺🇸🇺🇸😊😊😊😊😊😘😘😘

[@USONOFAV]

Google late to the party again. Microsoft GitHub Co-Pilot is very useful to us programmer. Open AI GPT-3 is useful for writing essay. Google? playing Go and Starcraft, lol

[@jestexan12]

666

[@blank_3768]

god is dead

[@spicymeatballs2thespicening]

Is the devil's mix

[@malanbandara]

AI kind of scares me. what do you guys think?

[@dev0cloo]

Genuinely curious, what scares you about it?

[@malanbandara]

@@dev0cloo AI is now gaining the ability to take over some jobs. Even somewhat complex jobs like software engineering. Also, the amount of data that they use to train those AI algorithms are depending on the quality of that data. If those data contain malicious or inappropriate content, it'll get messy and ultimately affect the whole ecosystem which build around that AI system. According to Elon Musk's opinion on AI regulation, he suggests that some government or a particular group of companies need to regulate who and how those AI systems are build. Otherwise there'll be consequences.

[@TheOne-jm6tg]

@@malanbandara it already happened 200 years ago, when steamed machines replaced factory workers. This is some inevitable frictional unemployment that is necessary for social developments. The humanity as a whole will always better with technological advancements, thought individuals may not

[@generalginger7804]

@@TheOne-jm6tg communism for the rescue.

[@p5rsona]

@@malanbandara don't be scared baby, let it happen.

[@jordanbonnet1244]

linked to hololens 2 (you.. ...) you never try...(so for me bla,bla,bl....) and LaMDA2 sucks

[@rasboten2]

It’s not that impressive

[@KrazyNomadQc]

Preferred pronouns? Really??

[@blank_3768]

you alright?

[@ntino666]

It will become more and more necessary to state perferred pronouns in the future as microplastics continue to blur the line between male and female