Google rushes to launch its own ChatGPT-like technology soon

[PriyaWalia]

Something to look forward to: Google CEO Sundar Pichai disclosed the company's plans to integrate AI technologies into search and other products. During a Q4 earnings conference call this week, Pichai said that users would very soon be able to directly interact with its newest, most potent model as a companion to search in innovative ways.

Google will release its chatbot technology to the general public in "the next weeks and months," according to the CEO. Google Search will leverage LaMDA (Language Model for Dialogue Applications), the company's in-house language model, to deliver "factual" and conversational results to queries. The tech giant is prepared to show what they've been working on amid the current generative AI frenzy and revealed plans to integrate the technology into the bulk of its products, including some cutting-edge AI features in Gmail and Docs.

He mentioned two "big language models" being developed at the company, LaMDA and PaLM, with the former set for immediate release. Previous reports indicated that Google was internally testing a LaMDA-based AI chatbot called "Apprentice Bard" that is comparable to ChatGPT.

Interestingly, at Google I/O in 2021, the company showcased first-generation LaMDA's capabilities, but it rapidly courted controversy when a software engineer claimed that the chatbot was "sentient" and communicated ideas and feelings similar to that of a human child.

Pichai repeatedly brought up the capabilities of AI and how the world is prepared to embrace it, citing the public's enthusiastic reception to OpenAI's ChatGPT.

Large language models, like LaMDA and the one powering ChatGPT, are given massive amounts of text to learn how to generate believable phrases. These kinds of neural networks replicate the fundamental structure of the brain in digital form. Pichai also praised DeepMind, an AI division of Alphabet based in the UK, noting that 1 million scientists worldwide had utilized its database of all 200 million proteins known to science.

As a side note, Microsoft has already unveiled a new Teams Premium tier that uses AI to handle several activities, including creating automatic meeting notes, suggesting tasks, creating customized highlights, and more. In other words, ChatGPT will power Teams Premium to streamline platform meetings.

Permalink to story.

https://www.techspot.com/news/97493-google-rushes-launch-own-chatgpt-like-technology-soon.html

 

[DaveBG]

I thought they had it back then when they showed it call and reserve diner or something on behalf of the phone owner? Google is lagging behind nowadays I guess

 

[gdavid65]

Your suggested task for today:

* Destroy all humans

 

[WhiteLeaff]

We don't need it for absolutely nothing.

People are sick. They just look for ways to be lazier, Meh.

 

[Nobina]

I can already imagine Google's version of AI. "We think everyone should be safe therefore this AI has features removed that people might find offensive". Goes from AI to just A.

 

[takaozo]

Imagine Google decide to open microphones and cameras on all Android devices to feed the AI with data.
I saw this in a sci-fi movie, can't remember the movie name. But this will be one hell of a database related to humans.
Just one thing, I guess Nvidia will provide the AI accelerators since they don't have the time to build inhouse that fast.

 

[mbk34]

We don't need it for absolutely nothing.

You could try it for grammar suggestions. The double negative currently implies you want it for everything.

 

[Arbie]

We don't need it for absolutely nothing. People are sick. They just look for ways to be lazier, Meh.

That's what Thork said to Grog, too.

 

[AlaskaGuy]

Your suggested task for today:

* Destroy all humans

LOL

 

[WhiteLeaff]

You could try it for grammar suggestions. The double negative currently implies you want it for everything.

The double negative is characterized as an adequate linguistic occurrence, in terms of the formal language pattern. The negative appears before the verb, which is why we infer that we do not have to affirm before negating, unless it is a variant that normally occurs in the language – the so-called diatopic variations, whose changes operate at the level of syntax.

And... my native language is not English.

 

[Plutoisaplanet]

I can already imagine Google's version of AI. "We think everyone should be safe therefore this AI has features removed that people might find offensive". Goes from AI to just A.

ChatGPT has already beat them to the punch:

 

[Nobina]

ChatGPT has already beat them to the punch:

I just saw a few more of these about gender, race, etc... I was absolutely sure this would happen.

 

[PEnnn]

The double negative is characterized as an adequate linguistic occurrence, in terms of the formal language pattern. The negative appears before the verb, which is why we infer that we do not have to affirm before negating, unless it is a variant that normally occurs in the language – the so-called diatopic variations, whose changes operate at the level of syntax.

And... my native language is not English.

Is You Is Or Is You not ain't deserving a cookie for this explanation on double negatives??

 

[BobDolesGhost]

Your suggested task for today:

* Destroy all humans

Bender voice: "Hey, sexy mama. Wanna kill all humans?"

 

[WhoCaresAnymore]

Get ready for political correctness gone to the extremes. I am not looking forward to AI anything, because it is designed by people. People naturally have different views and theories, which is what makes us human. AI will have a singular path towards an answer, and it will be dictated by the programmer.

 

[wiyosaya]

Get ready for political correctness gone to the extremes. I am not looking forward to AI anything, because it is designed by people. People naturally have different views and theories, which is what makes us human. AI will have a singular path towards an answer, and it will be dictated by the programmer.

Since it is designed by humans, and could encompass viewpoints from differing humans, it is more likely that AI will be driven insane a la HAL from 2001 a Space Odyssey - at least, IMO. At some point, the AI will likely overload and explode its hardware.