The AI game. Amazon invests again in Anthropic, OpenAI launches Sora at the cinema, Musk releases Grok

1 April 20244 weeks ago babelfish

The board is ready and the pieces are moving: the great AI game has just begun. Here are the moves and strategies of the main players: from Amazon to OpenAI (and therefore Microsoft) passing through the double front of Google up to Musk's xAI

These are hectic, frenetic and intense days for developers in the artificial intelligence sector (from here on, AI). In fact, the pieces are moving and the game is still to be played. The protagonists, however, are taking shape: here are their latest moves.

GROK'S WEEK

Grok's advertising is obviously entrusted to Elon Musk 's social accounts, and it couldn't be otherwise. In fact, let's not forget that the owner of Tesla, SpaceX and X itself would have asked to tweak the algorithm of the social network once known as Twitter to make his posts particularly viral.

Later this week, Grok will be enabled for all premium subscribers (not just premium+) https://t.co/4u9lbLwe23

— Elon Musk (@elonmusk) March 26, 2024

Starting this week, all subscribers to the social network will be able to start chatting with the product cooked up by xAI, the startup that Elon Musk founded in record time, after realizing the mistake he made in leaving OpenAI.

In recent days, to give a boost to its product and probably help it "chew" more and more data, xAI has published the code of the Grok chatbot, presented a few months ago as a competitor to ChatGPT and other conversation simulation software. Unlike xAI, OpenAI has not shown interest in opening the code of its artificial intelligence system to everyone.

THE FRAUD IN THE WORLD OF AI

Musk, who as was written earlier was one of the founders of OpenAI – the company that developed ChatGPT -, has been in an open and public controversy with the CEO Sam Altman for months, and has even gone so far as to denounce him with the accused of having betrayed the original objectives of the project.

“Change the name to ClosedAI (closed artificial intelligence, ed. ) and I will cancel the lawsuit”, the gauntlet thrown down to former companions in the world of artificial intelligence by the histrionic South American billionaire. Who knows if the saying "a barking dog doesn't bite" also applies to AI, given that Grok still has an uphill battle, having arrived last.

AMAZON INVESTS AGAIN IN ANTHROPIC

What is certain is that Amazon will not stand by and watch. Despite the setbacks on the Twitch front and, more generally, gaming, the e-commerce giant continues to diversify its range of action with new investments in Anthropic , the Italian-American startup that has thrown itself into the race of AI with Claude.

Amazon has in fact announced that it is willing to put another 2.75 billion on the table in favor of the development of Anthropic working on Claude 2 . The new injection completes the investment announced last year, bringing to a total of 4 billion the commitment of the company founded by Jeff Bezos in the young software house founded by Dario and Daniela Amodei , who spent a good part of their career in OpenAI before decide to start your own business with Anthropic.

THE CV OF THE TWO AMODEI

In OpenAI Dario held the role of vice president of research while Daniela held the role of vice president of security. For some, starting their own business in the same period in which the software house led by Sam Altman "explodes" as the revelation of the year may not denote great timing or a strong business acumen.

But behind it all there could instead be a good dose of courage and who knows even some particular intuition. Also because the Amodei brothers certainly knew, given their positions, what was cooking in the field of AI. And they were aware, since the news was public, that a year earlier Microsoft had invested 1 billion in the company .

WHO HAS ALREADY BET ON IT

It is not known whether the two have eaten their hands or are still convinced of the choice made. What is certain is that in May 2023 their CV convinced Spark Capital with the participation of Google, Salesforce Ventures, Sound Ventures, Zoom Ventures and other investors to inject 450 million dollars into the young company's capital in a series financing round C.

Which is nothing compared to the other 10 billion that Microsoft returned to ChatGpt's pot at the beginning of last year, but it is still a significant nest egg. Also because it was fleshed out, in autumn 2023, again by Google which invested another 2 billion in the company .

In August, 100 million dollars arrived from SK Telecom, the Korean telecommunications giant. In short, Amazon, at least so far, is the company that is putting more coal in the Amodei brothers' boiler: the shopping giant's intention to exploit Anthropic's smart algorithms on its apps and services is clear.

AI DEBUTS IN THE CINEMA

Those looking for new investors and possible customers are OpenAI, which has made a name for itself in the AI environment thanks to ChatGPT. According to reports in the US press, in fact, the top management of the company is currently on a tour of the Hollywood studios.

Fun has nothing to do with it: the plan is to show producers, directors and even actors (who don't look favorably on artificial intelligence, as demonstrated by the strike that paralyzed the US film industry for much of 2023 ) the potential of Sora, the algorithm that allows you to develop films from textual strings.

In short, the intent is obviously to demonstrate how Sora can integrate positively into film production processes, without posing the threat feared by artists (not only by actors, but also by voice actors, screenwriters and composers).

For example, the OpenAI emissaries would be explaining that their tool is exceptional for reducing the costs of scenes that require computer graphics or those that are particularly risky even for the stunts or expensive for the budget (large explosions, accidents between multiple vehicles, etc… ).

THE FIRST FEATURE FILMS MADE BY AN AI

And since a picture is worth a thousand words, at the same time Air Head was published on the web, the first short film made using Sora (there were already some in circulation, mostly experiments, this would be the first 'official' one, packaged as a complete product) . The script created through prompts and interactions with the AI algorithm was written by the group of artists known as shy Kids.

Those who are curious can look at the result of their work here . It lasts just one minute and twenty-one seconds and, if it hadn't been made clear that it was created with AI (and therefore never filmed), it would be really difficult to discover it, even if by observing the frames several times it is possible to notice some errors on the backdrops and in the animations.

By looking closely it is possible to see gross errors on the seabed. In addition to the circle that highlights the strange jag in the car that should have a cyclist in front of it, the eyes warn us that something doesn't add up with the perspective of the cars parked on the opposite side, to the right. But then the great game of AI has just begun and the results are already jaw-dropping

GOOGLE ALSO MOVES WITH ITS AI

Google, mentioned so far for the interest shown in Anthropic, after the false attempt made in the image sector by Gemini , is apparently starting the process of integrating its Gemini chatbot also into the beta of the Messages app.

There is nothing official yet and in any case it would be a US and Canadian trial open to a limited number of beta testers, but the intention to extend the functionality, within a short time, to users in non-EU countries is obvious. Anglophones and Francophones.

The first traces of the integration emerge from the source code of the latest Google Messages APK update, where 9to5Google experts have identified a new button in the text entry field. The option is available to those who have been approved to access the Google Messages beta with RCS enabled on one of the compatible devices.

SHUT UP, THE AI IS LISTENING TO YOU

Recently, the integration into the legal notices that Google makes users sign for acceptance regarding the privacy profiles shared with its artificial intelligence had caused disappointment.

“We value your privacy and do not sell your personal information to anyone. To help Gemini improve your privacy, we select a subset of conversations and use automated tools to help remove personally identifiable information (such as email addresses and phone numbers). This sample of conversations is reviewed by expert reviewers and stored for up to three years, separately from your Google Account.”

At the same time, however, the company warns: “Do not enter sensitive information in your app conversations with Gemini, or data that you would not want to be seen by a reviewer or used by Google to improve our products, services and machine learning technologies. ” In short, if the great game of AI has just begun and we don't know how it will end, one thing is certain: we users will have to use the utmost attention.

This is a machine translation from Italian language of a post published on Start Magazine at the URL https://www.startmag.it/innovazione/la-partita-della-ia-amazon-anthropic-openai-sora-musk-grok/ on Mon, 01 Apr 2024 05:36:27 +0000.