AMD Announces MI300X, its Most Advanced GPU for AI

By

-

June 14, 2023

The MI300X, AMD‘s most advanced GPU for AI, was announced on Tuesday. According to analysts, AMD’s launch poses the most significant threat to Nvidia, which now holds a market share of more than 80% for AI chips.

According to AMD, its new MI300X CPU and CDNA architecture were created for advanced AI models and large language models. The MI300X can fit even larger AI models than other chips because it can take up to 192GB of memory. For instance, the H100, a rival to Nvidia, only supports 120GB of RAM.

Because they perform an increasing number of calculations, large language models for generative AI applications consume a lot of memory. The Falcon model, which has 40 billion parameters, was demonstrated by AMD on the MI300x.

Additionally, AMD said that it would provide an Infinity Architecture that integrates eight of its M1300X accelerators into a single unit. For AI applications, Nvidia and Google have created comparable systems that pack eight or more GPUs into a single box.

AI programmers have historically favored Nvidia chips because they have access to the chip’s essential hardware characteristics through a well-developed software package called CUDA. On Tuesday, AMD claimed to have its own software, dubbed ROCm, for its AI chips.

AMD did not provide a pricing, but the move may put pressure on the cost of Nvidia GPUs like the H100, which may cost up to $30,000. The MI300X will be ready for sample this fall, but larger shipments won’t begin until next year, AMD CEO Lisa Su stated on an earnings call last month.

Microsoft AI Introduces A 13-Billion Parameter Model Orca

By

Sahil Pawar

-

June 14, 2023

Orca, a 13 billion parameter model that learns step-by-step thought processes and complex explanation traces from GPT-4, has been introduced by a team of Microsoft researchers. The performance of current state-of-the-art instruction-tuned models is greatly enhanced by this novel method, which also addresses issues with task diversity, query complexity, and data scaling.

The researchers agree that the GPT-4 query and response pairs can offer helpful guidance for student models. Therefore, researchers enhance these pairs by adding detailed responses that offer a better understanding of the reasoning process employed by the teachers when generating their responses. Orca bridges this gap by adding the explanation traces and giving student models better reasoning and understanding abilities.

The Flan 2022 Collection is used by the research team to further improve Orca’s learning. The team chooses tasks at random from this large library to ensure a variety of challenges. These activities are subsequently subsampled to provide intricate prompts that act as LFM questions. With the help of this method, the Orca develops a diversified and extensive training set that allows strong learning and equips it to do a variety of tasks with ease.

To evaluate Orca’s capabilities, the researchers perform thorough tests with a focus on its generative, reasoning, and comprehension skills. They assess Orca’s performance in comparison to reliable benchmarks like Text-Davinci-003, ChatGPT, GPT-4, and Vicuna.

The findings, which show an improvement of over 100% on BigBench Hard (BBH), highlight Orca’s supremacy over cutting-edge instruction-tuned models like Vicuna-13B. Additionally, in zero-shot environments, Orca displays competitive performance on academic exams, demonstrating its potential for real-world applications.

The study’s findings support the idea that the tremendous potential of learning from step-by-step explanations in enhancing model performance. Orca makes substantial progress in instruction-tuned models by including thorough explanation traces and scaling challenges with complex prompts. This strategy not only helps student models to outperform current benchmarks but also empowers them to improve their reasoning and comprehension capabilities.

Accenture to Invest $3 Billion in AI to Accelerate Client Growth

By

Sahil Pawar

-

June 14, 2023

Accenture to Invest $3 Billion in AI — Image Credits: Accenture

In order to assist clients across all industries progress and apply AI ethically to achieve more growth, efficiency, and resilience, Accenture has announced a $3 billion investment over three years in its Data & AI practice.

Accenture will make investments in assets, business solutions, ventures, talent acquisitions, ecosystem partnerships, and human resources in order to deepen and expand its skills and capabilities in the areas of diagnostic, predictive, and generative AI. 80,000 AI professionals will be added to the Data & AI practice through a combination of hiring, acquisitions, and training.

A generative AI-based platform called the new AI Navigator for Enterprise will assist clients in defining business cases, making decisions, navigating AI journeys, selecting architectures, and comprehending algorithms and models in order to responsibly deliver value. The platform will feature resources intended to speed up responsible AI practices and compliance programmes, building on Accenture’s own work.

Accenture will develop pre-built industry and functional models that make use of new generative AI capabilities, as well as accelerators for data and AI readiness across 19 different industries.

Its Centre for Advanced AI is committed to maximizing the value of this technology across clients and inside Accenture in order to develop the uses of generative AI. This entails significant R&D efforts and financial commitments to rethink service delivery utilizing generative and other cutting-edge AI capabilities.

To redefine the kind of work delivered to clients, Accenture will invest in both new and current relationships across its leading cloud, data, and AI ecosystems. For instance, designers might improve prototyping by leveraging pre-built models, or creators can build dynamic virtual worlds that can change in response to changes in the outside world.

Tech Mahindra CEO Accepts Sam Altman’s Challenge To Build Indian ChatGPT

By

Sahil Pawar

-

June 13, 2023

Tech Mahindra CEO Accepts Sam Altman Challenge Indian ChatGPT — Image Credits; NDTV

Recently, OpenAI CEO Sam Altman spoke at an event in Delhi, where Rajan Anandan, the previous CEO of Google India inquired about his advice on how Indians could develop something akin to ChatGPT in India. Sam responded by adding that while one could attempt to construct something similar in India, it would be “pretty hopeless.”

When Tech Mahindra CEO CP Gurnani interpreted Altman’s statement as a “challenge,” the conversation attracted even more attention. “OpenAI founder Sam Altman said it’s pretty hopeless for Indian companies to try and compete with them,” he commented on Twitter after sharing the video. “CHALLENGE ACCEPTED,” he added.

Further confirming his commitment to the challenge, CP Gurnani today tweeted, “To all those who are asking… Yes, the CHALLENGE IS ON! The purpose of this challenge is not one-upmanship or competition or headline hunting. It is a challenge to ourselves — to push the frontiers of technology for nation building. I’ve always believed that AI is not “artificial’ intelligence –- it is ‘augmented’ intelligence.”

He further added, “Tech Mahindra’s AI think tank is in action already. We will work with academia and the government to figure out how we can “apply” human and machine intelligence to boost innovation and transform India, one industry at a time.”

The CEO of OpenAI had earlier responded to Gurnani’s tweet on Saturday, claiming that the question he was given was the incorrect one and that his response had been taken out of context. He tweeted, “This is really taken out of context! The question was about competing with us with $10 million, which I really do think is not going to work. But I still said try! However, I think it’s the wrong question.”

Center Denies Direct Data Breach of CoWIN User Data

By

Sahil Pawar

-

June 13, 2023

Centre denies direct data breach CoWIN user data — Image Credits; Hindustan Times

The Centre has taken note of an automated account on the messaging app Telegram that was reportedly passing sensitive personal data of Indian individuals, including their Aadhaar and passport numbers, who registered for the CoWIN portal for their Covid-19 vaccine.

With ref to some Alleged Cowin data breaches reported on social media, @IndianCERT has immdtly responded n reviewed this

✅A Telegram Bot was throwing up Cowin app details upon entry of phone numbers

✅The data being accessed by bot from a threat actor database, which seems to…
— Rajeev Chandrasekhar 🇮🇳 (@Rajeev_GoI) June 12, 2023

The administration said on Monday that all such reports are without any basis and mischievous in nature, hours after rumors surfaced that CoWIN data had been accessed by a Telegram bot.

The Indian Computer Emergency Response Team (CERT-In), the main cyber security organization, assessed the purported breach, according to Minister of State for Electronics and IT Rajeev Chandrasekhar. He said, “It does not appear that the CoWIN app or database has been directly breached.” According to him, the information the Telegram bot was accessing came from a database of “threat actors” that “seem to have been populated with previously stolen data.”

The explanation from the health ministry did not, however, clarify how the Telegram bot was able to provide user information associated with a phone number. Since the government has never officially recognised that Aadhaar data has been compromised, there are no specifics on previous data breaches that have been brought up by the government.

When the phone number that was used to register for the CoWIN site was messaged to the bot, the Telegram account, which has not been inactive since Monday morning, displayed personal information about a person. The Telegram bot displayed the person’s name, the government identification they used to receive their vaccination, and the location where they got it, all while claiming to get their information through the CoWIN portal.

The health ministry stated in its statement that “CERT-In, in its initial report, pointed out that the backend database for Telegram bot was not directly accessing the APIs of the CoWIN database.” According to sources, the APIs are being used to access CoWIN data by over 110 institutions, including 78 government entities.

The ministry claimed that the platform’s development team had confirmed that no public APIs could access data without an OTP, but one API did allow users to share data with external organizations like the Indian Council of Medical Research (ICMR) by calling the phone number linked with their Aadhaar number. This API, they clarified, only allows queries from trusted APIs that CoWIN has whitelisted, according to the statement.

Salesforce Expands its Generative AI Fund from $250 Million to $500 Million

By

Sahil Pawar

-

June 13, 2023

Salesforce expands generative AI fund $250 million $500 million — Image Credits: Salesforce

Salesforce today announced that it’s growing its Generative AI Fund from $250 million in size to $500 million. The fund is the part of the company’s Salesforce Ventures VC off-shoot which backs startups developing responsible generative AI.

Paul Drews, managing partner of Salesforce Ventures said, “The expansion will allow Salesforce to work with even more entrepreneurs, accelerating the development of transformative AI solutions for the enterprise. We’re already witnessing how AI is altering how society functions, and we’re eager to expand on the success of our Generative AI fund. We’re thrilled to support the next wave of creative entrepreneurs.”

The Generative AI Fund, which was established in March, has already made investments in a number of businesses at the forefront of generative AI technology, including Cohere, Anthropic, You.com, Hearth.AI, and more recent additions Humane and Tribble. Salesforce hopes to set itself apart with their tranche by giving priority to what it refers to as “ethical” AI solutions, despite not being the only fund investing solely in generative AI.

The wide variety of technologies that the Generative AI Fund is funding is another way that it stands out. For instance, Humane is developing a wearable AI assistant without a display, while You.com offers an AI-driven search engine with text and generative art capabilities.

The Salesforce AI for Impact Accelerator, the second accelerator programme ever offered by the firm, launches at the same time as the Generative AI Fund’s expansion. A group of education, workforce, and environmental nonprofits will each get $2 million from AI for Impact to advance the equitable and ethical use of responsible AI.

Mangalore University Introduces BCA in AI and ML, B.Com in Business Data Analytics

By

Sahil Pawar

-

June 13, 2023

This year, colleges connected with Mangalore University will be offering new degrees including B.Com in Business Data Analytics and BCA programmes in Artificial Intelligence and Machine Learning in the upcoming sessions. The program structures are still under development.

The Beary’s First Grade College, located in Kodi, Kundapura, in the Udupi district, will offer a BCA programme in artificial intelligence and machine learning. The academic council of the university approved the admission of 40 students to this college at its meeting on Monday.

St. Raymond College in Vamanjoor, Mangaluru, and the Kanachur Institute of Management and Science, located in Deralakatte, both will offer the course. The Academic Council approved the admission of 60 and 30 students to the two colleges, respectively.

Beary’s Institute of Emerging Learning will offer a four-year B.Com in Business Data Analytics programme, with a cap of 90 students. On its campus in Adyar, Minerva College, formerly known as the Mangalore Institute of Fire and Safety Engineering, will also offer this program. 80 students may enroll, according to the Academic Council.

These programs will provide an overview of artificial intelligence and ethics and foundation of data analysis. The foundations of AI & ML based languages, tools and techniques are covered in the BCA with a specialization in artificial intelligence and machine learning to develop a strong base in programming. B.Com data analytics course entails analytics and applications in order to develop tangible business solutions.

Goa Board Introduces AI as Subject From Class 9 Onwards

By

Sahil Pawar

-

June 13, 2023

Goa Board AI subject Class 9 onwards — Image Credits: GBHSE

From Class 9 forward, the Goa Board of Secondary and Higher Secondary Education (GBSHSE) has introduced Artificial Intelligence as a new subject. In accordance with the state board’s official announcement, AI will be introduced in all schools that do not provide NSQF subjects.

According to the official circular, “With reference to the aforementioned subject, the new subject Artificial Intelligence (AI) has been introduced as an additional subject in accordance with NSQF Scheme, as 7% subject for Class 9 from the Academic year 2023-2024.”

All schools that do not offer NSQF subjects are required to provide AI as a subject. The syllabus for Artificial Intelligence (AI)/ICT for Classes 9 and 11 would be the same for the Academic Year 2023–2024 under school assessment, according to a partial modification to circular No. 33, dated 28TM June 2022. For Schools/Students opting for AI as the 7th subject at Class 9, the existing School Assessment Subject, Information and Communication Technology, will be exempted.

Students in class 9th who choose AI will have a theory assignment of 40 points (20 points – first term and 20 points – second term). They will also have a practical assignment of 60 points (30 points – first term and 30 points – second term).

For Class 9th and Class 11th, the Goa Board has suggested that schools conduct three sessions of 33 minutes each per week on AI. The Goa Board also revised the current ICT curriculum for Class 9th and Class 11tg in addition to adding AI as a new subject.

Sam Altman Calls For Collaboration With China To Counter AI Risks

By

Sahil Pawar

-

June 12, 2023

Sam Altman Calls Collaboration China Counter AI Risks — Image Credits: AD

According to Sam Altman, CEO of OpenAI, China should have a major influence on the guardrails for artificial intelligence that are needed to guarantee the security of revolutionary new technologies. “The stakes for international cooperation have never been higher,” said Altman, speaking via video link on Saturday to a conference in Beijing.

Talent and capital are pouring into AI, a crucial field that could help define the escalating tech rivalry between the world’s two greatest economies in Silicon Valley as well as China. The development of the AI technology has also brought to light difficulties in government attempts to regulate the industry, which, according to China’s President Xi Jinping, needs more state control to reduce hazards to national security.

Altman, at the event organized by the Beijing Academy of Artificial Intelligence, said, “China has some of the top AI expertise in the world, and fundamentally, given the challenges in resolving alignment for advanced AI systems, this requires the best minds from around the world.”

Since the academy has established a significant presence in the AI industry in China, Altman’s lecture at the Beijing conference was noteworthy in itself. Brad Smith, the president of Microsoft, has mentioned this Chinese non-profit as one of the three leaders in AI innovation. It is funded by the Ministry of Science and Technology of the nation and the municipal government of Beijing.

In China, where stringent data and censorship laws have long barred services from Western internet giants like Alphabet’s Google and Meta Platform’s Facebook, OpenAI’s ChatGPT is not yet accessible. Complex data and algorithm restrictions, according to experts, will also make it challenging for foreign corporations to advance AI in the nation.

Sam Altman Says His Remarks About Possibility of Indian Model Like ChatGPT Taken Out of Context

By

Sahil Pawar

-

June 12, 2023

Sam Altman remarks about Indian model ChatGPT taken out of context — Image Credits: AD

Recently, OpenAI CEO Sam Altman spoke at the Economic Times Conversations event in Delhi, where Rajan Anandan, the previous CEO of Google India inquired about his advice on how Indians could develop something akin to ChatGPT in India. He said, “Sam, we have a very vibrant ecosystem in India, but specifically focusing on AI, are there spaces where you see an Indian startup building foundational AI models? How should we think about that?”

“Where is it that a team from India, with three super-smart engineers, having not 100, but USD 10 million each, could actually build something truly substantial?” he added.

Sam responded by adding that while one could attempt to construct something similar in India, it would be “pretty hopeless.” During the interview, the CEO of OpenAI said, “We’re going to teach you how this works. Competing with us in the area of training foundation models is just futile. Even though you shouldn’t try, it’s your responsibility to enjoy trying. And I firmly believe both of those. It seems rather hopeless to me.”

Thank you @sama for the clear answer. As you said, "it is hopeless, but you will try anyway". 5000 years of Indian entrepreneurship has shown us that we should never underestimate the Indian entrepreneur. We do intend to try 🚀@sama @OpenAI https://t.co/gQkxOZAFCW
— Rajan Anandan (@RajanAnandan) June 8, 2023

After that, Anandan tweeted a clip of the exchange with the comment, “Thank you Sam for the clear answer. As you said, ‘it is hopeless, but you will try anyway.” 5000 years of Indian entrepreneurship has shown us that we should never underestimate the Indian entrepreneur. We do intend to try.” Even though Anandan responded positively to Altman’s response, Twitter users weren’t as kind to the OpenAI CEO. Social media soon became a hive of activity as netizens discussed Altman’s statement.

OpenAI founder Sam Altman said it’s pretty hopeless for Indian companies to try and compete with them.

Dear @sama, From one CEO to another..

CHALLENGE ACCEPTED. pic.twitter.com/67FDUtLNq0
— CP Gurnani (@C_P_Gurnani) June 9, 2023

When Tech Mahindra CEO CP Gurnani interpreted Altman’s statement as a “challenge,” the video attracted even more attention. “OpenAI founder Sam Altman said it’s pretty hopeless for Indian companies to try and compete with them,” he commented on Twitter after sharing the video. “CHALLENGE ACCEPTED,” he added.

this is really taken out of context! the question was about competing with us with $10 million, which i really do think is not going to work. but i still said try! however, i think it’s the wrong question.
— Sam Altman (@sama) June 10, 2023

The CEO of OpenAI responded to Gurnani’s tweet on Saturday, claiming that the question he was given was the incorrect one and that his response had been taken out of context. He tweeted, “This is really taken out of context! The question was about competing with us with $10 million, which I really do think is not going to work. But I still said try! However, I think it’s the wrong question.”

the right question is what a startup can do that’s never been done before, that will contribute a new thing to the world. i have no doubt indian startups can and will do that! and no one but the builders can answer that question.
— Sam Altman (@sama) June 10, 2023

He later added, “The right question is what a startup can do that’s never been done before, that will contribute a new thing to the world. I have no doubt Indian startups can and will do that! And no one but the builders can answer that question.”