ChatGPT Advanced Voice Mode, a feature that first started rolling out in September, is now being added to the artificial intelligence (AI) chatbot’s desktop apps. Announced on Thursday, OpenAI’s native chatbot will now offer a human-like voice chat experience to Mac and Windows users. The feature was first unveiled at the OpenAI Spring Updates event in May and it can express emotions, modulate the voice, and react to what the user is saying. So far, only the paid subscribers of the platform have access to the feature.

ChatGPT Advanced Voice Mode Comes to Desktop Apps

In a post on X (formerly known as Twitter), the official handle of OpenAI announced that the Advanced Voice Mode has been rolled out to macOS and Windows desktop apps. The move is interesting as major AI firms have started focusing their attention towards the desktop to offer more powerful and comprehensive AI capabilities to users.

On the same day, Anthropic released its desktop apps for Mac and Windows, paving the way for the Computer Use tool. Google is also reportedly working on a new agentic AI browser tool that will be able to complete tasks such as booking movie tickets and purchasing a product. Now, with OpenAI’s Advanced Voice, users can finally utilise the full capability of voice-based AI in a desktop environment. Notably, so far the feature was only available to Android and iOS apps.

Some of the ways users can take advantage of the ChatGPT Advanced Voice Mode is by verbally prompting the AI to write a code, or having a back-and-forth while writing a research paper or college assignment. Users can also upload data files and then have a two-way conversation about its analysis and insights.

ChatGPT app users will find the option to turn on the Advanced Voice Mode by tapping the waveform icon placed next to the text field. Tapping on the icon activates the new voice mode. Users now have five new voices to choose from — Vale, Spruce, Arbor, Maple, and Sol. Each of these voices has a different pitch, tonality, and regional accent.

However, the feature is still only available to the ChatGPT Teams and Plus users. Additionally, those residing in the EU, the UK, Switzerland, Iceland, Norway, and Liechtenstein will not get the new feature.

ChatGPT is being rolled out as a Windows app for paid subscribers, OpenAI announced last week. The artificial intelligence (AI) chatbot could earlier only be used via the web client, but with this move, users can access a dedicated app to use all the generative AI functionalities. Currently, the app is available as an early beta version to ChatGPT Plus, Enterprise, Team, and Edu users. The AI firm stated that a public version of the app will be released later this year.

ChatGPT App for Windows

In a post on X (formerly known as Twitter), the official handle of OpenAI announced the roll out of the ChatGPT desktop app for Windows. The company said that the app is being rolled out to Windows operating system (OS) to enable a faster access to the chatbot for their daily use cases. Additionally, users will be able to quickly launch the app using the Alt + Space shortcut on the keyboard.

The app is currently available to only paid subscribers of ChatGPT, which includes the ChatGPT Plus, Enterprise, Team, and Edu users. It is available in an early beta version. OpenAI plans to introduce the public stable version sometime later this year. Notably, those using the free tier of the AI chatbot will not get access to the app at present.

However, unlike the macOS version of the app, the company did not announce any new features for the desktop app. As such, it is not clear whether users will get the quick screenshot feature which can capture either the entire screen or a part of it and automatically upload it to the app.

Users should be able to use all the standard AI features including generating text and images, asking queries that require web search, as well as having conversations with the chatbot. They will also be able to upload files and have it be processed by ChatGPT.

To access the app, one can head to the Microsoft app store and search for ChatGPT. Alternatively, they can also click here to directly go to the app listings page. To install the ChatGPT for Windows app, users will need to have 283MB of free storage. It supports Windows 10 version 17763.0 or higher.

ChatGPT is being rolled out as a Windows app for paid subscribers, OpenAI announced last week. The artificial intelligence (AI) chatbot could earlier only be used via the web client, but with this move, users can access a dedicated app to use all the generative AI functionalities. Currently, the app is available as an early beta version to ChatGPT Plus, Enterprise, Team, and Edu users. The AI firm stated that a public version of the app will be released later this year.

ChatGPT App for Windows

In a post on X (formerly known as Twitter), the official handle of OpenAI announced the roll out of the ChatGPT desktop app for Windows. The company said that the app is being rolled out to Windows operating system (OS) to enable a faster access to the chatbot for their daily use cases. Additionally, users will be able to quickly launch the app using the Alt + Space shortcut on the keyboard.

The app is currently available to only paid subscribers of ChatGPT, which includes the ChatGPT Plus, Enterprise, Team, and Edu users. It is available in an early beta version. OpenAI plans to introduce the public stable version sometime later this year. Notably, those using the free tier of the AI chatbot will not get access to the app at present.

However, unlike the macOS version of the app, the company did not announce any new features for the desktop app. As such, it is not clear whether users will get the quick screenshot feature which can capture either the entire screen or a part of it and automatically upload it to the app.

Users should be able to use all the standard AI features including generating text and images, asking queries that require web search, as well as having conversations with the chatbot. They will also be able to upload files and have it be processed by ChatGPT.

To access the app, one can head to the Microsoft app store and search for ChatGPT. Alternatively, they can also click here to directly go to the app listings page. To install the ChatGPT for Windows app, users will need to have 283MB of free storage. It supports Windows 10 version 17763.0 or higher.

Meta, the parent company of Facebook, unveiled a set of innovative AI models developed by its research division on Friday, reported Reuters.
Among the standout tools is the “Self-Taught Evaluator,” which may reduce the need for human involvement in the AI development process. This development is an important step towards creating AI systems capable of learning from their own mistakes, potentially paving the way for more autonomous and intelligent digital agents.
In addition to the Self-Taught Evaluator, Meta also released updates to its image-identification Segment Anything model, a tool for accelerating response generation times in large language models (LLMs), and datasets designed to support the discovery of new inorganic materials.
First introduced in an August research paper, the Self-Taught Evaluator uses the same “chain of thought” technique employed by OpenAI‘s latest models. This approach involves breaking complex tasks into smaller steps to increase accuracy in fields like science, coding, and mathematics.
Crucially, Meta’s researchers trained the evaluator entirely on AI-generated data, eliminating the need for human input during the training phase.
According to Meta researchers, the ability of AI to evaluate other AI models accurately opens new possibilities for autonomous AI systems that can self-improve. This could lead to the development of digital assistants capable of performing a wide range of tasks without human intervention.
Self-improving AI models may also reduce reliance on the costly and time-consuming process of Reinforcement Learning from Human Feedback (RLHF), which involves specialised human annotators verifying data and checking AI-generated answers for accuracy.
Jason Weston, one of Meta’s researchers, expressed hope that as AI becomes more advanced, it will become increasingly capable of verifying its own work, surpassing human accuracy.
Other companies, such as Google and Anthropic, have also been exploring the concept of Reinforcement Learning from AI Feedback (RLAIF).
However, unlike Meta, these companies have been more cautious in releasing their models to the public.

Facebook owner Meta said on Thursday it had teamed up with Blumhouse Productions, the Hollywood company behind popular horror films such as The Purge and Get Out, to put its new generative AI video model Movie Gen to the test.

The announcement comes after Meta unveiled Movie Gen, which it said can create realistic-seeming video and audio clips in response to user prompts, earlier this month. Meta claimed the tool could rival offerings from leading media generation startups such as OpenAI and ElevenLabs.

In a blog post, the social media company said Blumhouse had selected filmmakers Aneesh Chaganty, The Spurlock Sisters and Casey Affleck to try out Movie Gen and use clips generated by the tool in their short films.

Chaganty’s film would appear on Meta’s Movie Gen website, while films from Affleck and The Spurlock Sisters were forthcoming, Meta said.

Blumhouse CEO Jason Blum said in a statement that artists remain the lifeblood of the industry and innovative technology can aid in their storytelling. 

“We welcomed the chance for some of them to test this cutting-edge technology and give their notes on its pros and cons while it’s still in development,” Blum said. “These are going to be powerful tools for directors, and it’s important to engage the creative industry in their development to make sure they’re best suited for the job.”

With the partnership, Meta is signaling how it aims to collaborate with creative industries, whose members have largely recoiled in response to the arrival of generative AI technologies over concerns around copyright and consent.

Several groups of copyright owners have sued major tech companies, including Meta, over the unauthorised use of their works to train generative AI systems. Meta has argued its AI training is protected by the copyright doctrine of fair use.

However, Meta and other tech companies have shown they are willing to pay for certain types of content for AI. Aside from the Movie Gen partnership, Meta said last month it had struck deals with actors including Judi Dench, Kristen Bell and John Cena to give voice to its Meta AI chatbot.

Similarly, Microsoft-backed OpenAI has been meeting with Hollywood executives and agents this year to discuss possible partnerships involving its video generation tool Sora, which it first showed off in February.

No deals have been reported to have come out of those talks yet, though Lions Gate Entertainment said in September that it had struck a deal with another AI startup, Runway.

© Thomson Reuters 2024

(This story has not been edited by NDTV staff and is auto-generated from a syndicated feed.)

Alphabet’s Google is moving the team behind its Gemini AI assistant app to its DeepMind research lab, continuing a plan to consolidate the company’s various groups working on Artificial Intelligence (AI).

The search giant is simplifying its structure to “keep increasing the pace of progress” of AI development, Chief Executive Officer Sundar Pichai said in a blog post on Thursday. He also announced that Prabhakar Raghavan, the most senior leader of Google’s search and ads units, is leaving the role after four years helming the company’s flagship businesses. Raghavan will step into a new role as Google’s chief technologist, Pichai said.

Nick Fox, a veteran Google executive who acted as a search deputy under Raghavan, will step into the role leading the company’s search, ads, maps and shopping services.

Long the dominant search engine globally, Google has battled the perception for the last two years that it’s lagged behind the likes of Microsoft, OpenAI and other startups in rolling out new generative AI tools and services. Yet, as it works to stay competitive with new entrants upending search, it must also take care not to cannibalise its core profit machine. In his announcement, Pichai framed the reorganisation as a way to streamline the company’s progress in AI.

Over the last six months, Google has been consolidating its AI-focused teams, seeking to improve the Gemini models to challenge the likes of OpenAI and Anthropic. In April, the firm moved its models, research and responsible AI teams to the DeepMind division. Soon after, it merged DeepMind with Google Brain, a rival research unit within the company. 

DeepMind was founded in London in 2010 as an academic-style research lab, then acquired by Google in 2014. Executives say the lab has recently shifted away from its original mission and become more product-driven.

Both inside and outside Google, “so many of the leading research labs are actually product companies at this point,” said Eli Collins, DeepMind’s vice president of product, in an interview with Bloomberg News last month. He said DeepMind has had to “pick up the pace” to keep up with the rate of AI innovation.

Google has also been challenged by mounting antitrust scrutiny from federal officials. In August, Google lost a trial over claims by the US Justice Department that the company illegally monopolises online search and advertising markets.

In September, a trial covering the company’s alleged dominance over the technology used to buy and sell online ads concluded, with closing arguments scheduled for November. A decision in that case is expected by the end of the year.

© 2024 Bloomberg LP

(This story has not been edited by NDTV staff and is auto-generated from a syndicated feed.)

Microsoft on Monday said that its vice president of GenAI research, Sebastien Bubeck, is leaving the company to join ChatGPT-maker OpenAI.

It was unclear what role Bubeck would assume at the Microsoft-backed AI startup.

“Sebastian has decided to leave Microsoft to further his work toward developing AGI,” a spokesperson for Microsoft said, adding that the company looks forward to continuing their relationship through Bubeck’s work with OpenAI.

Bubeck did not immediately respond to a Reuters request for confirmation.

Most of Bubeck’s coauthors who worked on a research paper on Microsoft’s Phi LLMs, which are smaller than traditional large language models (LLMs), remain at Microsoft and plan to continue developing the models, according to the Information, which first reported his departure.

The development follows a series of departures from OpenAI, including that of longtime chief technology officer Mira Murati in September.

CEO Sam Altman has denied any link between the departures and a planned restructuring of the company.

© Thomson Reuters 2024

(This story has not been edited by NDTV staff and is auto-generated from a syndicated feed.)

ChatGPT, the artificial intelligence (AI) chatbot by OpenAI is reportedly about to get more expensive for paid subscribers. According to a new report, the AI firm is planning to increase the subscription price for ChatGPT Plus users by $2 (roughly Rs. 167) a month. The price hike is not expected to stop there either, as the company is said to push the monthly subscription cost to $44 (roughly Rs. 3,685) in the next five years. The reason behind pushing for a higher ticket price is said to be OpenAI’s revenue ambitions and expensive cost of running operations.

ChatGPT Subscriptions to Reportedly Get More Expensive

According to The New York Times, the AI firm is planning to increase the subscription price by $2 by the end of 2024. Citing financial documents viewed by the publication, the report further added that the final price of the ChatGPT Plus subscription might stand at $44 a month by the end of 2029, a steep climb from the current $20 a month in the US or Rs. 1,950 a month in India.

With the price hike, the company reportedly wants to secure a revenue of $100 billion (roughly Rs. 8.3 lakh crore) in 2029, a majority of which is expected to come from its subscription-based services. If the AI firm is able to achieve this target, it would be raking in annual revenue similar to Reliance Industries, Nestlé, or Comcast. OpenAI currently has approximately 10 million ChatGPT Plus users, according to the report.

The documents reviewed by the publication that was meant for the investors and OpenAI reportedly highlighted that it is currently making “billions” from ChatGPT, and expects to boost the numbers significantly in the coming years. Notably, the company is in the process of closing a funding round.

Despite the big numbers projected in its revenue estimation, the company is reportedly struggling with optimising its operational costs. OpenAI is said to lose approximately $5 billion (roughly Rs. 41.8 thousand crore) this year, most of which goes towards its AI-powered services. Other significant cost-centres include employee salaries and office rent.

Another major source of expense is reportedly cloud computing for which the company uses Microsoft’s services. Despite getting $13 billion (roughly Rs. 1.08 lakh crore) yearly due to the partnership, the AI firm spends much of that money on running cloud processing.

Sam Altman denied on Thursday that there was any link between the departure of three senior OpenAI executives and a planned restructuring of the company which he said the board had been considering for several months.

The world-leading AI firm’s longtime chief technology officer Mira Murati abruptly announced her departure on Wednesday. Within hours, two senior research executives Barret Zoph and Bob McGrew revealed they were also leaving the company.

On the same day, Reuters reported OpenAI was working on a plan to restructure its core business into a for-profit benefit corporation that will no longer be controlled by its non-profit board, in a move aimed at making the company more attractive to investors.

Speaking on stage at the Italian Tech Week conference in Turin, OpenAI CEO Altman said that “some stuff” reported around the executives’ departures was inaccurate, adding that the personnel changes were unrelated to the restructuring.

“That’s totally not true,” he told event organiser John Elkann, the chairman of car companies Ferrari and Stellantis.

“A lot of the stuff I saw was also just totally wrong, but we have been thinking about that (restructuring), our board has, for almost a year, independently, as we think about what it takes to get to our next stage.”

Altman praised the departing executives, adding he was excited to simplify the company’s structure and work more closely with technical staff.

“I have not been as involved in the tech recently as other things, because there’s been so much going on, I’m excited to do that,” he said.

“This will be, hopefully, a great transition for everyone involved, and OpenAI will be stronger for it, as we are for all our transitions.”

Details of the proposed restructure highlight significant changes behind the scenes at one of the world’s top AI companies.

Sources told Reuters the plans were still being negotiated between lawyers and shareholders at the time of writing, with the timeline for completion still uncertain.

© Thomson Reuters 2024

(This story has not been edited by NDTV staff and is auto-generated from a syndicated feed.)

A potential deal to buy Intel could accelerate Qualcomm’s diversification but will burden the smartphone chipmaker with a loss-making semiconductor manufacturing unit that it may struggle to turn around or sell, analysts said.

A buyout will also face tough antitrust scrutiny globally as it would unite two crucial chip firms in what would be the sector’s biggest ever deal, creating a behemoth with a strong share of the smartphone, personal computer and server markets.  

Shares of Intel rose nearly three percent on Monday, after media reports late on Friday about Qualcomm’s early-stage approach for the struggling chipmaker. Qualcomm’s shares were down 1.8 percent. 

“The rumored deal between Qualcomm and Intel is intriguing on many levels and, from a pure product perspective, makes a certain degree of sense as they have a number of complementary product lines,” said TECHnalysis Research founder Bob O’Donnell.

“The reality of it actually occurring, however, is very low. Plus, it is unlikely Qualcomm would want all of Intel and trying to break apart the product business from the foundry business right now just would not be possible,” he said.

Once the dominant force in the semiconductor industry, five-decade-old Intel is facing one of its worst periods as losses mount at the contract manufacturing unit it is building out in hopes of challenging TSMC. 

Intel’s market value has fallen below $100 billion (roughly Rs. 8,36,313 crore) for the first time in three decades as the company has missed out on the generative AI boom after passing on an OpenAI investment. 

As of last close, its market capitalisation was less than half that of potential suitor Qualcomm, which has a value of about $190 billion (roughly Rs. 15,88,934 crore).

Considering Qualcomm had around $7.77 billion (roughly Rs. 64,980 crore) in cash and cash equivalents as of June 23, analysts expect the deal will mostly be funded through stock and would be highly dilutive for Qualcomm’s investors, likely raising some apprehension.

Qualcomm, which also supplies to Apple, has quickened its efforts to expand beyond its mainstay smartphone business with chips for industries including automotive and PCs under CEO Cristiano Amon. But it still remains overly reliant on the mobile market, which has struggled in recent years due to the post-pandemic demand slump.

Amon is personally involved in the Intel negotiations and has been examining various options for a deal for the company, sources have told Reuters.     

This is not the first time Qualcomm is pursing a large acquisition. It had offered to buy rival NXP Semiconductors for $44 billion (roughly Rs. 3,67,973 crore) in 2016, but abandoned the bid two years later after failing to secure a nod from Chinese regulators.

Foundry Conundrum 

While Intel designs and manufactures its chips that power personal computers and data centers, Qualcomm has never operated a chip factory. It uses contract manufacturers such as TSMC and designs and other technology supplied by Arm Holdings. 

Qualcomm lacks the experience needed to ramp up Intel’s fledgling foundry business, which recently named Amazon.com as its first major customer, according to analysts. 

“We do not know why Qualcomm would be a better owner for those assets,” said Stacy Rasgon of Bernstein. 

“We do not really see a scenario without them either; we do not think anyone else would really want to run them and believe scrapping them is unlikely to be politically viable,” he added.

Intel’s foundry business is seen as crucial to Washington’s goal of growing domestic chip manufacturing. The company has secured about $19.5 billion (roughly Rs. 1,63,079 crore) in federal grants and loans under the CHIPS Act to build and expand factories across four US states. 

Some analysts said Intel would prefer outside investments instead of a sale, pointing to a recent move to make the foundry business more independent. 

Bloomberg News reported over the weekend that Apollo Global Management, already a partner in Intel’s Ireland facility, has offered an investment of as much as $5 billion (roughly Rs. 41,814 crore) in the company.

Qualcomm could also decide to buy parts of Intel’s business, instead of the entire company. Reuters had reported earlier this month that it had particular interest in Intel’s PC design unit.

© 2024 Bloomberg LP