Meta Announces Inference Chip Project Developing In-house

By

-

May 19, 2023

Meta announces inference chip project — Image Credits: Meta

On Thursday, Meta Platforms revealed fresh information about the initiatives it was pursuing to improve the suitability of its data centers for supporting artificial intelligence work, including a proprietary chip family that it claimed to be developing in-house.

In a series of blog posts, the owner of Facebook and Instagram claimed that as part of the Meta Training and Inference Accelerator (MTIA) programme it designed a first-generation chip in 2020. The programme aimed to increase the effectiveness of the recommendations models it uses to serve ads and other content in news feeds.

In a prior story from Reuters, it was stated that the corporation was already developing a replacement and had no plans to extensively deploy its first in-house AI processor. The initial MTIA chip was presented in the blog articles as a teaching tool.

“From this initial programme, we have learned invaluable lessons that we are incorporating into our roadmap,” the statement read.

According to the posts, the initial MTIA chip was only focused on the AI technique known as inference, in which computers trained on vast quantities of data make decisions about whether to display, for example, a dancing video or a cat meme as the next item in a user’s feed.

A Meta representative declined to comment on deployment schedules or go into further detail about the company’s aspirations to create chips that could also train the models.

Telecom Giant BT To Cut 55,000 Jobs With Up To A Fifth Replaced By AI

By

Sahil Pawar

-

May 18, 2023

Telecom giant BT to cut 55,000 jobs — Image Credits: BT

Due to cost-cutting measures, telecom giant BT plans to eliminate up to 55,000 positions by the end of the decade, primarily in the UK.

As labor is replaced by technology like artificial intelligence, up to a fifth of those cuts will be made in customer support. Staff and contractors are also included in the headcount reduction from the existing workforce of 130,000.

“Whenever you get new technologies, you can get big changes,” stated CEO Philip Jansen. According to him, “generative AI” tools like ChatGPT, which can compose screenplays, poems, essays, and computer code in a manner akin to a human, “gives us confidence we can go even further.”

Even while BT has stated that by the end of the 2020s, it will have a “much smaller workforce,” Mr. Jansen continued, “new technologies drive new jobs.”

The leading internet and mobile provider in the UK, BT, is actively expanding its fiber network as it transitions away from copper. The business said that once the work was over, fewer employees would be required to create and maintain its networks.

In addition, it claimed that fewer workers will be required to serve clients in the future due to newer, more effective technology, such as artificial intelligence.

Meta to Face Privacy Fine Over Facebook Data Transfer from EU to US

By

Sahil Pawar

-

May 18, 2023

Meta privacy fine Facebook data transfer EU US — Image Credits: Meta

According to two persons familiar with the situation, Meta Platforms is likely to receive a record-breaking European Union privacy penalties for failing to heed a top EU court’s warning about the transfer of Facebook user data from EU to US servers.

According to the sources, the fine will be larger than the previous record 746 million euros ($821.20 million) charged to Amazon.

Meta declined to comment. Also, the requests for comment from the Irish Data Protection Commission (DPC) and the European Commission were not immediately answered.

Because they were concerned that American spy agencies would gain access to the data, EU regulators under the direction of Ireland’s Data Protection Commissioner Helen Dixon have been working to finalize a prohibition on the legal mechanism used by Facebook to transfer European user data.

They had stated in April that the Irish DPC had a month to issue a ruling banning Facebook’s transatlantic data flows. Mid-May may see the implementation of the ban.

A data transfer agreement between the EU and the US was declared unconstitutional in 2020 by the highest court in Europe, citing privacy concerns. A prohibition on the method Meta employs to send data from Europe to the US could compel it to halt Facebook services in Europe, Meta said last year.

AI Startup Sanctuary AI Designs Humanoid General-Purpose Robot Phoenix

By

Sahil Pawar

-

May 18, 2023

Phoenix is a brand-new, highly developed general-purpose robot that was revealed yesterday by Vancouver-based Sanctuary AI, an artificial intelligence and robotics startup. According to the business, it has created the first humanoid general-purpose robot in history, powered by Carbon, a unique AI control system.

To deal with the employment issues afflicting many firms, the SOTA AI system gives robots intelligence like to that of a person and enables them to perform a variety of functions. An important step in the company’s development towards full commercialisation, the company announced its first commercial deployment in March of this year. The business has now unveiled the sixth generation of technology in less than two months.

Sanctuary AI, a company founded in 2018, is on a goal to develop the first general-purpose robots with human intelligence that will aid in safe, effective, and sustainable human labour. In addition to being part of the founding teams of Sanctuary AI, the members at the startup are also a part of D-Wave, Kindred, Creative Destructive Lab, and other companies.

The startup has collaborated with organizations like Apptronik, Bell, Common Sense Machines, Contoro, Cycorp, Exonetik, HaptX, Magna, Tangible Research, Verizon Ventures, Workday Ventures, and others in an effort to achieve the ambitious goal of giving general-purpose robots human-like intelligence.

The startup obtained Series AI investment in March of last year. The company got a C$30 million funding from the Canadian government’s Strategic Innovation Fund (SIF) in November 2022, increasing its total funding to over C$100 million.

According to Sanctuary AI, their technology has already been used to do hundreds of jobs that customers from more than a dozen different industries have identified. Phoenix was created to be the most sensor-rich and physically capable humanoid robot ever developed, according to Sanctuary AI’s chief and cofounder Geordie Rose, and to enable Carbon’s rapidly developing intelligence to carry out the widest range of labor conceivable.

Stability AI Announces Release of Open-source StableStudio

By

Sahil Pawar

-

May 18, 2023

Stability AI release StableStudio — Image Credits: Stability AI

Stability AI today announced the release of StableStudio, the company’s premier interface for its Stable Diffusion model of art-generating artificial intelligence. A free and open-source alternative to DreamStudio is StableStudio.

After its release, DreamStudio quickly became known as a one-stop shop where users could access Stability AI’s text-to-image creating functions without having to know any programming. Users might easily produce artwork to test out new features as they become available by filling in instructions. In April, the system received the Stable Diffusion XL beta model, which can deliver even more vibrant and photorealistic pictures than before.

With the open-source release of StableStudio, developers will have total control over the interface for Stable Diffusion, enabling them to build and use custom plugins that can alter the underlying AI models made available by Stability AI.

As a result, StableStudio and its ecosystem will keep advancing in response to developer feedback. By integrating the project’s ongoing development into its public GitHub repository and creating incentives for improvements and new features, Stable Diffusion hopes to achieve this goal.

An employee of Google LLC recently complained in writing that leading providers of generative AI, such as Google’s Bard and OpenAI LP’s ChatGPT, were being rapidly encroached upon by the open-source AI community. The already-existing conflict in the industry between closed source and open source reached a peak as a result.

DRDO Introduces Online Certifications in Cyber Security and Artificial Intelligence

By

Sahil Pawar

-

May 18, 2023

DRDO online certifications cyber security artificial intelligence — Image Credits: DRDO

The Defence Research and Development Organisation (DRDO) has introduced online certification programmes in cyber security and artificial intelligence.

Applications are being accepted for a 16-week and 12-week online course on cyber security and artificial intelligence (AI), respectively, through DRDO’s institute, the Defence Institute of Advanced Technology (DIAT).

Participants will be given an overview of the 200 hours of course material for the cyber security course, which covers advanced subjects including malware analysis, drone and anti-drone technologies, and 120 hours of course material for artificial intelligence and machine learning.

Both programmes are open to undergraduates, recent graduates, IT professionals, tri-service officers, and anyone working in research and development from all academic disciplines.

The participants must register here and fill out their information for both online courses. Within two working days of submission, they will get an email with information on the next steps. The start dates for both online courses are set for June 12, 2023. Registration for these courses must be completed by May 25, 2023, and payment must be made by June 5, 2023.

Participants will pay Rs 29,500 for the online AI training and Rs 35,000 for the cyber security course. Five days a week, two hours will be dedicated to each of the two online courses. Participants will receive certificates if they have finished the course successfully.

OpenAI Planning to Release a New Open-source Language Model

By

Sahil Pawar

-

May 17, 2023

OpenAI planning new open-source language model — Image Credits: OpenAI

ChatGPT creator OpenAI is planning to release a new open-source language model, thus preparing to make a huge move in the world of artificial intelligence development. This comes as open-source AI models gain traction, posing a threat to industry titans such as Google and OpenAI.

This move comes only weeks after a leaked letter from a Google top developer acknowledged the growth of open-source AI technologies and their direct danger to Google and OpenAI.

The specifications of OpenAI’s future open-source model have yet to be revealed. It is not intended to compete directly with the flagship ChatGPT. Instead, it is expected to contribute to the developing open-source AI community.

When OpenAI released ChatGPT to the public, its generative AI gained great traction. This sparked a scramble among internet companies, with Google launching its own Bard chatbot shortly after. Meta Platforms, not wanting to fall behind, provided access to powerful machine-learning models capable of interpreting conversational language.

These models were subsequently converted into open-source software, paving the way for the creation of free alternatives to ChatGPT and other proprietary AI tools.

EU Regulators Approve Microsoft’s Acquisition of Activision Blizzard

By

Sahil Pawar

-

May 17, 2023

EU regulators approve Microsoft’s acquisition Activision Blizzard — Image Credits: Activision

Microsoft’s proposed $69 billion acquisition of gaming developer Activision Blizzard was approved by European Union regulators on Monday, subject to Microsoft’s suggested remedies.

The European Commission, the EU’s executive arm, stated that Microsoft provided antitrust remedies in the new field of cloud gaming. These solutions focused on allowing customers to broadcast Activision games purchased on any cloud streaming platform.

After the transaction was blocked by the United Kingdom’s top competition body last month, Europe’s approval represents a big gain for Microsoft. Global regulators have been investigating whether Microsoft’s acquisition of Activision will hinder competition in the console and cloud gaming markets. Regulators questioned whether Microsoft would take Activision games and keep them solely on its own platforms.

Earlier, the Federal Trade Commission (FTC) had filed a legal case against Microsoft to try and block the company’s plan to buy Activision Blizzard for about $68.7 billion. On the other hand, Chile’s market regulatory institution, the Fiscalia Nacional Economica (FNE), published its approval of the contentious Microsoft acquisition of Activision Blizzard in December last year.

Activision is responsible for some of the world’s most popular console and PC games, including the Call of Duty franchise and World of Warcraft. The EU decision comes after the UK Competition and Markets Authority blocked the transaction earlier this month, citing fears that it would decrease competition in the embryonic cloud gaming market.

According to the CMA, Microsoft would benefit financially from making Activision’s important games, such as Call of Duty, exclusive to its own cloud gaming platforms. Nonetheless, the CMA stated that the acquisition would not reduce competition in the console market. Regulators and some of Microsoft’s competitors, including PlayStation gaming system maker Sony, have expressed resistance to the purchase.

Adani Foundation Becomes World’s First Skill Centre In Metaverse

By

Sahil Pawar

-

May 17, 2023

Adani Foundation World's First Skill Centre Metaverse — Image Credits: Adani Foundation

By becoming the world’s first skills center to access the Metaverse, the Adani Foundation skill development vertical has achieved a significant milestone. Adani Skill Development Centre (ASDC), or Adani Saksham, has begun its Metaverse journey with two courses, seven years after its opening on May 16, 2016.

The non-profit organization, which aspires to skill India’s youth and contribute to nation building, stated that this achievement advances its purpose of leveraging cutting-edge technology to train future-ready professionals.

“With Metaverse, ASDC is stepping into an exciting phase where knowledge, skills, and technology converge to create an immersive educational experience through virtual classrooms,” the firm said in a release.

The ASDC has prioritized courses such as General Duty Assistant to support the healthcare and hospital industries, as well as Fire Safety in the Metaverse, and stated that additional courses will be offered in the future.

“This is a game changer, providing trainees with a deep understanding of their chosen field that extends far beyond the confines of a physical space.” Learners at 40 Adani Skill Development Centres in 13 Indian states would be able to enroll in these courses in metaverse, according to ASDC. The company will also make the courses available globally so that anyone with a computer can enroll

The Adani Foundation is the Adani Group’s community support and engagement arm, dedicated to making targeted social investments to generate long-term results across India.

OpenAI CEO Sam Altman Testifies at Senate Hearing

By

Sahil Pawar

-

May 17, 2023

OpenAI CEO Sam Altman testifies Senate hearing — Image Credits: CNN

Sam Altman, the CEO of OpenAI, emphasized his concern on the potential of artificial intelligence (AI) to tamper with election integrity in testimony before a Senate panel on Tuesday, highlighting the need for regulation in this field.

Altman emphasized that he is quite concerned about the possible risks connected to AI and elections and that these issues require the establishment of clear laws and norms.

The threat of false information was emphasized by Senator Mazie Hirono as the 2024 election neared. She brought up an incident where fake photos of former President Trump being detained by the NYPD went viral, and she questioned Altman about the possible harm brought on by such fake content. In response, Altman advised creators to make it very evident how an image is produced.

Altman suggested that the United States should take into account enacting licensing and testing standards for the creation of AI models as a standard procedure during his first appearance before Congress. Altman cited models’ capacity to influence or convince a person’s opinions as a crucial threshold when asked to describe which categories of AI should be subject to licensing.

Altman also backed the notion being discussed on Capitol Hill, which would allow businesses to choose whether or not to allow the use of their data for AI training. He emphasized, however, that content accessible to the general public on the internet should be regarded as fair game.

Additionally, Altman stated his desire for a subscription-based approach while stressing his openness to the possibility of advertising. The CEO emphasized the value of international cooperation and provided incentives for safety compliance when talking about the future of AI.