In a strategic announcement, Google Cloud collaborates with ParallelDots, one of the leaders in Retail Image Recognition solution, to deliver advanced, real-time AI solutions to global Consumer Packaged Goods (CPG) manufacturers and retailers. The goal is to employ the strengths of both the companies and unlock enhanced data accuracy and simplified AI training for CPGs. As a consequence of this, the in-store execution, customer satisfaction, and sales would increase. 

AI in Retail

Google Cloud ensures end-users with reliable performance and security due to its secure infrastructure. This integration will allow customers to deploy ParallelDots’ solutions quickly and easily. Customers will be able to  integrate ParallelDots’ shelf data within the Google Cloud platforms, eliminating the costly and complex task of manual integration.

This partnership helps combat losses related to the retail industry which is currently plummeting with a loss of 25% of sales annually due to poor in-store execution. Due to the limitations of manual store audits, there is also a dearth of  real-time data, efficient audits, and timely reporting. In such a context, advanced AI & IR solutions are gaining significance as they address problems like missing SKUs and price labels or incorrect product placement. 

Commenting on the partnership, Bikram Singh Bedi, Vice President and Country MD, Google Cloud India said, “Our aim is to offer our customers a secured infrastructure and capabilities to seamlessly perform complex tasks. With this collaboration we aim to empower ParallelDots to deliver unparalleled solutions to their customers. Our advanced cloud capabilities combined with their innovative technologies will simplify tasks, efficiently run complex problems and enhance cost efficiency.”

Ankit Singh, Co-founder & CTO, from ParallelDots is confident with this partnership, “This milestone marks a significant advance in delivering a world-class Retail Image Recognition solution to the global CPGs and retailers. Our partnership with Google Cloud enhances the reliability, security, and speed of our solutions, dramatically reducing deployment time, scaling our Image Recognition solution ShelfWatch, and boosting platform reliability and cost-effectiveness. This is a pivotal moment in our mission to create the world’s foremost retail shelf insights platform,” he said. Google Cloud’s strong presence in the retail space strengthens this collaboration in the marketplace. This joint solution stands as it allows end-customers to adopt ParallelDots’ technology, especially due to its platform security and robust infrastructure.

The post Google Cloud Partners with ParallelDots to Enhance Retail Shelf Monitoring with AI  appeared first on AIM.

Jaspreet Bindra, a tech evangelist and author, launched his AI literacy initiative “AI&Beyond” with the goal to make AI more accessible and understandable. It is in collaboration with Anuj Magazine, an AI & cybersecurity expert, it is spread across industries. 

“At AI&Beyond, we believe that AI is no longer the future – it is very much the present. Our goal is to ensure that AI literacy becomes as fundamental as reading and arithmetic, especially in large organisations where AI’s impact will be profound. Through AI&Beyond, we aim to bridge the gap between AI’s capabilities and its practical application across various sectors,” said Jaspreet Bindra. 

Inspired by CLR James’ book Beyond A Boundary, AI&Beyond aims to simplify and distill the concept of AI beyond the technical aspects of it. 

One of the initiative’s flagship programme, The Generative AI Bootcamp,  is designed to equip participants with the necessary and most important AI skills. Additionally, the platform’s Ethics Bootcamp is crucial in today’s AI-driven world, helping organisations not only adopt AI, but do so with a strong ethical foundation. 

The underlying idea is to make the learning experiential and immersive – that form the foundation of future AI businesses. These resources – including workshops, briefings, webinars, and consulting services – would empower organisations to become more agile, innovative and competitive in the digital age.

Jaspreet Bindra on GenAI

Bindra has been a staunch advocate of Make in India, and to that end, proposed that India should consider building generative AI as a digital public good – known as JanAI or GenAI for the people. “If you look at ChatGPT or Bard, they are all trained on the internet, where almost 80 to 90% of the data is English, and West-oriented. It doesn’t have vernacular data nor an Indian context,” said Bindra, while emphasising on why in the current LLM market, he wanted to bring an Indian context to the training model. 

Also, at MLDS 2024 – India’s biggest generative AI summit – hosted by AIM, Jasprit Bindra spoke about the history of AI and the concept of singularity in today’s generative trends. 

The post Tech Veteran Jaspreet Bindra launches ‘AI&Beyond’ to Democratise AI Literacy  appeared first on AIM.

HARMAN, a Connecticut-based Samsung subsidiary, specialising in audio electronics, launched HARMAN ForecastGPT, a predictive analytics platform that enables organisations to become more efficient through future forecasting and optimum resource allocation. It is aimed at understanding complex data patterns, and some of its chief features are advanced AI capabilities, real-time adaptability and the ability to seamlessly integrate with various platforms. It is compatible with any data format and source including CSV, Excel, SQL, and API.

HARMAN ForecastGPT’s Versatility 

It can benefit businesses in a plethora of ways including sales forecasting, supply chain forecasting, financial planning, and marketing even. 

“Embracing AI is imperative for business success and at HARMAN DTS, we are pioneering the application of AI to deliver tangible, bottom-line results. By understanding the unique challenges and aspirations of each client, we’re crafting AI solutions that go beyond generic predictions. The ForecastGPT platform is a testament to our commitment to equipping businesses with the tools to move past challenging roadblocks and fully capitalise on the potential of AI,” said Nick Parrotta, President – Digital Transformation Solutions & Chief Digital and Information officer at HARMAN.  

HARMAN & its Growing Capabilities 

HARMAN has a wide portfolio including audio and video systems, car audio, connected car solutions, professional audio and lighting equipment, and more. Additionally, it is also the parent company to brands like JBL, Harman Kardon, AKG, Mark Levinson, and Infinity Systems. 

“HARMAN’s data science team has made significant contributions by incorporating machine learning and deep learning models into a range of applications, such as predictive analytics, computer vision, NLP, and graph analytics,” said Dr Jai Ganesh, chief product officer of HARMAN, in an exclusive interview with AIM previously. 

The data science team of HARMAN employs both open source as well as commercial tools such as Python, TensorFlow, PyTorch, AWS, Azure, or Google Cloud Platform, Java, C++, Git, Jenkins, Docker, Kubernetes, R, Jupyter, SAS, MongoDB, Spark, Kafka, MySQL, RStudio, KNIME, RapidMiner, H2O etc.

The post HARMAN Introduces ForecastGPT, a GenAI Platform for Enterprises appeared first on AIM.

OpenAI’s Mira Murati shared that the first time her mother used ChatGPT, she asked, “When will Mira get married?”. 

In a recent interview at Cannes Lion 2024, Murati recalled that when she first introduced her mother to ChatGPT in 2022, she was with her sister in Italy. She explained to her mother that she could ask any question she wanted,even in Albanian. Her mother’s first question to ChatGPT was about Mira’s marriage. To this, Mira’s sister humorously responded, “Mom, it’s not magic, it’s artificial intelligence”.

“She was having such a natural interaction with it that she thought she could ask anything,” said Murati. 

Paytm founder Vijay Shekhar Sharma took to X and said, “#MomsEverywhere.”

#MomsEverywhere https://t.co/8bIcdVYFlE

— Vijay Shekhar Sharma (@vijayshekhar) September 3, 2024

“didn’t know she was Indian,” quipped a user on X. 

Another user humorously remarked, “The most Albanian mother interaction ever. Recalling OpenAI’s leadership transition when Mira Murati was appointed Interim CEO during the Sam Altman kerfuffle, the user shared,”I told my mom, “Look, an Albanian woman is the CEO of one of the most innovative companies in the world.”

His mother’s response? “Message her, ask her on a date, and for a job.”

The most Albanian mother interaction ever.

During @OpenAI's Sam kerfuffle where Mira was appointed Interim CEO, I told my mom "look an Albanian woman is the CEO of one of the most innovative companies in the world."

She Replied:
Message her asking her on a date & for a job. https://t.co/uZ5niCpg7a

— Ar₿en (@arben777) September 3, 2024

Since joining OpenAI in 2018, Murati has been instrumental in leading the development of ChatGPT, DALL-E, and the GPT series. Her leadership has been recognised by industry leaders, including Microsoft CEO Satya Nadella, for her ability to blend technical expertise with a deep appreciation for OpenAI’s mission of responsible AI development.

Murati briefly served as interim CEO of OpenAI in November 2023, following the unexpected removal of Sam Altman. Her tenure as interim CEO was short-lived, as Emmett Shear replaced her three days later, and Altman was reinstated five days after his removal. 

The post “When Will Mira Get Married?” OpenAI CTO’s Mother Asked ChatGPT appeared first on AIM.

With AI code editors like Cursor, Zed, and Codeium, gathering all the attention, there is an open source alternative recently launched in the market. Meet Melty, an open source AI code editor, which is specifically designed for 10x engineers.

Register for NVIDIA AI Summit India

Started by Charlie Holtz and Jackson de Campos, Melty is backed by Y Combinator and is part of the S24 batch. The founders tout it is the first editor that understands what developers are coding from the terminal to GitHub and collaborates for writing production ready code.

Click here to check out the GitHub repository for Melty. 

In a thread on X, Holtz said that during the first weeks of YC they were still figuring out a bunch of ideas when they started working on AI developer tools as suggested by Aaron Epstein. “We’re big fans of the tools that are out there. But we still find ourselves copy-pasting from Claude, juggling ten chats for the same task, and committing buggy code that comes back to bite us later. Nothing is quite ‘it’ yet,” said Holtz.

After working on Melty for 28 days, Holtz said that Melty is already writing about half of its own code.

“Charlie and Jackson build so fast you’d think they’re a team of 10 stacked engineers, rather than just 2 co-founders. Melty’s going to give every engineer their superpowers,” said Epstein. 

The goal that the team explains of building Melty is to help people understand the code by allowing them to watch every change like a pair programmer and it also learns to adapt to the developer building the code. 

The post Meet Melty, Open Source Alternative to Cursor  appeared first on AIM.

Microsoft has announced the appointment of Vaishali Kasture as the General Manager for its Small, Medium, and Corporate (SMC) business in India and South Asia. With over 25 years of leadership experience, Kasture will oversee operations at a time when technology adoption is accelerating across both corporate and SMB sectors. 

Kasture previously served as the Head of AWS India and South Asia. She will be the second high-ranking executive to leave AWS for Microsoft within the past year, as the companies vie for a larger share of the cloud market in India. Her predecessor at AWS, Puneet Chandok, became the head of Microsoft India and South Asia in September last year.

“I cannot think of a better time to take up the role as GM for the SMC business in India and South Asia,”said Kasture on her appointment to the new role.  She highlighted Microsoft’s comprehensive approach to technology, focusing on areas such as cloud computing, artificial intelligence, data, and security.

“Microsoft is singularly positioned as a full-service technology company, focusing on an end-to-end technology stack,” Kasture added, citing the company’s ability to enhance productivity through generative AI.

Kasture also referenced Microsoft CEO Satya Nadella’s vision, quoting him: “Microsoft is like the United Nations of Software. Helping people and organizations to do more, to achieve more.”

India, described by Kasture as “the fastest-growing large economy in the world,” presents significant opportunities for Microsoft, which has bolstered its presence in the country over the past three decades. 

Kasture will collaborate with key figures including Chandok and Rachel Bondi to advance Microsoft’s vision in the region. She expressed gratitude towards Ahmed Mazhari and Kevin Peesker for their support and partnership during her transition into the role.

The appointment underscores Microsoft’s commitment to expanding its footprint in India and South Asia, aligning with the growing demand for technology-driven solutions in the region.

The post Microsoft Appoints Former AWS Head Vaishali Kasture as GM for India and South Asia appeared first on AIM.

Revrag, a startup developing AI agents, has announced the launch of its first AI-powered solution, Emma.

Emma is designed to revolutionise the initial stages of customer outreach and lead generation. Seamlessly integrating with popular tools like Slack and HubSpot, Emma streamlines the sales process, making it faster and more efficient for teams. With Emma, sales teams can easily:

• Set up campaigns to target potential customers.
• Customise outreach efforts based on specific business needs.
• Save valuable time by automating the initial contact stages.
• Potentially reach a larger pool of leads in less time.

“The SaaS world is evolving rapidly, and we see enormous potential for Generative AI in the sales industry, a sentiment echoed by recent market trends. AI agents like Emma represent a transformative technology that’s reshaping business operations globally. We aim to fill existing gaps and lead the market towards greater innovation. For small businesses and startups, AI agents like Emma offer a unique opportunity to level the playing field,” Ashutosh Singh, CEO of Revrag, said.

Revrag’s AI agents are sophisticated systems designed to function autonomously, learn from their environment, and achieve specific goals with minimal human intervention.

Revrag’s journey has been bolstered by an impressive $600k in pre-seed funding, reflecting investor confidence in the company’s vision. As the industry eagerly anticipates the launch of Emma, Revrag is set to challenge traditional sales processes and potentially usher in a new era in enterprise sales.

Backed by prominent angel investors and a powerhouse venture, Revrag is poised to make a significant impact in the competitive SaaS landscape.

The Bengaluru-based startup has attracted the attention of industry heavyweights, including Viral Bajaria, Co-founder of 6sense, Kunal Shah, Founder of Cred, Deepal Anchala, Founder of Slintel, Vetri Vellore, Founder of Rhythms, and 20 other renowned investors.

The post Revrag Unveils Its First AI Agent Emma appeared first on AIM.

Finacus Solutions, a company specialising in banking technology, and pi-labs.ai, a startup focused on AI-powered deepfake detection, have joined forces to develop a new e-KYC (electronic Know Your Customer) solution. This collaboration aims to improve the security of e-KYC by addressing the growing concern of deepfake fraud within the financial industry.

This partnership marks a significant step forward in the e-KYC landscape. The new solution integrates pi-labs.ai’s deepfake detection technology into Finacus Solutions’ existing e-KYC framework.

Addressing the Rise of Deepfake Fraud

The Reserve Bank of India (RBI) has mandated live video for KYC procedures, supplemented by enhanced security measures. However, the emergence of deepfake videos in Video KYC presents a formidable challenge, particularly for credit and loan applications.

This partnership signifies a pivotal moment in the application of artificial intelligence within the eKYC process. pi-labs.ai’s AI tools will complement the existing manual authentication process, creating a sophisticated hybrid system that merges human expertise with AI-driven insights to ensure the highest level of security. This collaboration not only strengthens the integrity of the eKYC process but also paves the way for further automation, ultimately reducing operational costs.

“The integration of Aadhar-based eKYC has drastically reduced KYC costs,” said Rahul Ayyappan, co-Founder and CTO of Finacus Solutions. “The financial industry has a vested interest in maintaining the integrity of the eKYC process. The AI-based detection capabilities provided by pi-labs.ai will offer our banking clients enhanced security and peace of mind.”

pi-labs.ai’s deepfake detection technology, ‘Authentify,’ is an AI-powered platform designed to detect deepfakes across various media formats. Utilising advanced AI algorithms, pi-labs.ai empowers enterprises to identify and mitigate fraud attempts involving deepfakes.

“Our deepfake detection technology is poised to revolutionise the eKYC process,” said Ankush Tiwari, Founder and CEO of pi-labs.ai. “The collaboration with Finacus Solutions will greatly enhance the reliability and security of eKYC, ensuring that the financial sector remains resilient against emerging challenges.”

Also Read: Banks are Working on Tech to Counter Deepfakes

The post Finacus Solutions and pi-labs Develops World’s First eKYC Solution Resistant to Deepfake Frauds appeared first on AIM.

AI educator and DeepLearning.AI founder, Andrew Ng and Meta AI chief Yann LeCun joined Korea’s National AI Committee as advisors. The announcement came following a meeting between Ng and the President of Korea and the Minister of Technology. This collaboration marks a pivotal moment for the country’s AI ambitions, as it leverages the expertise of two of the most influential figures in the field.

South Korea has been witnessing rapid development in the AI landscape marked by substantial government investment, a flourishing startup ecosystem and breakthroughs in AI chip technology. Andrew Ng’s involvement with the National AI Committee will undoubtedly accelerate these advancements.

Andrew Ng’s recent appointment to the National AI Committee comes on the heels of a series of educational initiatives launched through DeepLearning.AI. These courses cover a broad spectrum of AI topics, equipping learners with the skills to build robust AI applications, train models securely on private data, and explore efficient large language model pre-training techniques. 

South Korea’s AI Ambitions

This educational focus aligns with South Korea’s aspiration to become a global AI leader. Ng’s involvement with the National AI Committee is likely to further this goal, potentially leading to the development of new educational programs that nurture a strong domestic AI talent pool.

The country’s efforts towards AI based developments have started yielding results, South Korea has witnessed a surge in AI startups, exceeding 1,100 in 2024 alone. These startups are making waves in the field, with companies like Upstage developing sophisticated AI models such as Solar, a powerful Korean language LLM that even surpassed models from established tech giants. Align AI, another South Korean startup, has garnered recognition for its innovative AI chatbot solutions.

Andrew Ng’s involvement with the National AI Committee will undoubtedly accelerate these advancements. Ng’s extensive experience and educational initiatives through DeepLearning.AI align perfectly with South Korea’s goals. This collaboration is likely to lead to the development of new educational programs and foster a culture of AI innovation.

The post Andrew Ng and Yann LeCun Joins Korean National AI Committee As Advisors appeared first on AIM.

Reliance Industries Limited (RIL) chairman and managing director (CMD) Mukesh Ambani announced the launch of the Jio AI-Cloud Welcome Offer, and Jio Phone Call AI at the company’s 47th annual general meeting. 

Jio Phone Call AI is a no app based feature that allows users to record calls on their server and transcribe it too, it also allows translation for the AI generated transcription all by calling on their AI number.

“Today, to support our AI Everywhere For Everyone vision using connected intelligence, I am thrilled to announce the Jio AI-Cloud Welcome Offer,” Ambani said, with user data safely stored on the cloud, AI will be able to deliver intelligent personalised services over the network, he added.

Set to go live in Diwali, this offer will allow its Jio users 100 GB of free cloud storage space to securely store and access photos, videos, documents and other digital content and data. 

Earlier this year, Jio launched ‘Jio Brain’, positioned as the Industry’s first 5G-integrated ML platform aimed to empower telecom networks, enterprise networks and industry-specific IT environments to seamlessly incorporate ML tools into their day-to-day operations.

Akash Ambani, Chairman of Reliance Jio Infocomm Limited (RJIL) also announced the upgrade to its HelloJio, Jio STB’s voice assistant using Gen AI technology, improving its natural language processing making it feel more human-like to understand Indian dialects better.

Jio Goes Big on AI

Jio is making significant strides in the field of AI, looking at the broader Indian AI landscape, the JioGenNext cohort highlights the potential of AI in various sectors like healthcare, banking, and agriculture. Startups like Medhini-Arficus and Dista are developing innovative solutions that address real-world problems. Additionally, the collaboration between NVIDIA, Tata Communications, and Jio Platforms will provide the necessary infrastructure for AI development in India.

Akash Ambani, the chair of Jio also announced that it is partnering with IIT Bombay for the BharatGPT initiative aimed to address the lack of large language models (LLMs) for Indic languages. BharatGPT’s open-source approach fosters collaboration and knowledge sharing, which is essential for overcoming these hurdles and establishing India as a leader in Indic LLMs.

The post Reliance Jio Announces Free Cloud Storage with Jio AI-Cloud appeared first on AIM.

OpenAI, ChatGPT maker, is expected to raise funding from Apple and NVIDIA at a valuation of $100 billion. Per reports, Thrive Capital, an NYC based VC firm, will lead the round. OpenAI is currently valued at $86 billion. 

Earlier this year, Microsoft, OpenAI’s oldest, and most significant strategic partner, valued them at $30 billion. While they are still in the talks to participate in the upcoming fundraise, the details are unclear. 

Apple and OpenAI could see this fundraise as a new way to strengthen their existing relationship that saw a new height with the “Apple Intelligence” initiative where OpenAI’s AI features were made a core part of apple’s products. This acquisition is also telling about OpenAI’s value within the apple ecosystem, and how Tim Cook, CEO of Apple, wants an inside on OpenAI’s future innovations.  

NVIDIA is optimistic in OpenAI commitment to “Humanity, Computing and AI.”  In line with this, earlier this year, Jensen Huang, NVIDIA’s CEO, personally hand-delivered the first NVIDIA DGX H200 to OpenAI. 

What does Open AI’s future hold?

OpenAI’s growth since the launch of 2022 has been record breaking. It has sparked a reckoning in the AI community, especially among investors and VC firms, who are constantly in the race of staying  on top of cutting edge AI driven technologies and capturing market share. 

On 29th August, OpenAI reported that ChatGPT has surpassed more than 200 million active users.  Fortune 500 companies are heavily dependent on OpenAI, with around 92% of them using the latter’s products, per an OpenAI spokesperson. According to her, the API usage has also doubled since the launch of GPT-4o mini. OpenAI’s recent collaboration with AI Safety Institute  also highlights their commitment to advance safe and trustworthy AI innovation for all.

The post Apple, NVIDIA in talks to Raise Funding for OpenAI appeared first on AIM.

Infosys has announced an expansion of its partnership with NVIDIA to bring AI-driven, customer-centric solutions to the telecommunications industry. 

The collaboration will integrate Infosys Topaz, an AI-first suite of services and platforms, with NVIDIA’s advanced generative AI technologies to deliver innovative solutions.

Anand Swaminathan, EVP & global industry leader – communications, media and technology, Infosys, expressed enthusiasm about the partnership and said, “Through our collaboration with NVIDIA, we are poised to unlock cutting-edge enterprise AI capabilities, helping global organisations improve business efficiency.”

As part of this partnership, Infosys has developed three generative AI solutions powered by Infosys Topaz. These solutions leverage NVIDIA’s NIM inference microservices, NeMo Retriever embedding models, and NeMo Guardrails to customise and deploy telco domain-specific large language models (LLMs). 

Additionally, Infosys Topaz utilises NVIDIA Riva for building real-time conversational AI pipelines, enabling real-time transcription and translation for call centre agents.

Highlighting the growing importance of generative AI in the telecommunications sector, Ronnie Vasishta, Senior Vice President of Telecom at NVIDIA, said, “Telcos are increasingly adopting generative AI solutions to improve the productivity of their businesses with smarter networks, more efficient operations, and enhanced customer service.”

One of the key innovations emerging from this partnership is the TOSCA Network Service Design solution, which employs NVIDIA’s generative AI to streamline and customise TOSCA (Topology and Orchestration Specification for Cloud Applications) templates. 

This has resulted in a 28% reduction in latency and a 15% improvement in accuracy. Similarly, the NVIDIA-powered smart networking operations centre has demonstrated remarkable efficiency, reducing issue resolution times by 61% and improving accuracy by 22%.

Another highlight of the collaboration is Infosys Cortex, a generative AI-driven contact centre transformation platform. By leveraging Infosys Topaz and Infosys’ cloud suite, Infosys Cobalt, Cortex creates realistic customer simulations for trainee agents, offering an immersive training environment. 

These solutions are expected to enhance operational efficiencies, boost productivity, and optimise generative AI pipeline and inference costs.

The post Infosys, NVIDIA Introduce Generative AI Powered Telco Solutions appeared first on AIM.

While the debate around GitHub Copilot and Cursor AI is increasing, Codeium, an AI-powered code acceleration platform, announced that it has raised $150 million in a Series C funding round, propelling its valuation to $1.25 billion, becoming a unicorn.

The round was led by General Catalyst, with continued participation from existing investors Kleiner Perkins and Greenoaks. This funding milestone marks Codeium’s ascent to Unicorn status in less than two years since its inception.

“The future of coding isn’t just about writing lines of code faster—it’s about enabling developers to think bigger, push boundaries, and achieve the extraordinary,” said Varun Mohan, CEO of Codeium. “This fresh funding means we’re more equipped to help developers turn those ‘what ifs’ into ‘what’s next,’ having the freedom to innovate without limits and turn challenges into opportunities for growth.”

Codeium’s platform leverages proprietary code-based LLMs to streamline software development and enhance developer productivity. With the newly secured funds, the company plans to accelerate the development of new features, expand its product offerings, and increase its workforce. The focus will also be on strengthening partnerships to maximise AI strategies and broaden its market impact.

Quentin Clark, Managing Director of General Catalyst, emphasised Codeium’s rapid adoption in real-world production environments, stating, “Codeium isn’t just an idea—it’s a fully scaling business with widespread enterprise adoption. Their genAI tools for software development are proving their worth in real production environments, where reliability is key.”

The funding announcement follows a period of significant growth for Codeium. The company has expanded its team to 80 professionals and grown its user base to over 700,000 active developers. In 2024, the enterprise product achieved eight figures in annual recurring revenue, with ARR increasing by over 500%. Additionally, Codeium now processes more than 100 billion tokens daily and has been integrated into production workflows at major companies such as Zillow, Dell, and Anduril.

Codeium has proactively removed “non-permissively” licensed code, such as copyrighted code, from the datasets used to train its AI models. This step addresses a common issue with some code-generating tools that, when trained on restrictively licensed or copyrighted code, can inadvertently reproduce that code, leading to potential legal risks for developers. According to Mohan, Codeium avoids this problem through its careful preparation and filtering of training data.

Recent technological advancements from Codeium include the launch of Cortex, an AI-powered reasoning engine for managing complex coding tasks, and Forge, an AI-assisted tool that enhances code review efficiency and culture. These innovations are part of Codeium’s mission to transform software development by making coding faster, smarter, and more intuitive.

In similar news, another AI coding assistant platform, Magic, closed a round of $320 million, including participation from Eric Schmidt, Atlassian, Jane Street, and Sequoia.

The post Cursor Rival Codeium Raises $150 Mn, Becomes Unicorn with $1.25 Bn Valuation appeared first on AIM.

NPCI, IRCTC, and CoRover have introduced “Conversational Voice Payments” for UPI at the Global Fintech Fest 2024, marking a significant leap in the ease of digital transactions. This new feature, integrated with payment gateways and powered by BharatGPT, allows customers to complete transactions using their voice or by typing their UPI ID or mobile number. 

When a mobile number is provided, the system automatically retrieves the corresponding UPI ID and initiates the payment request via the user’s default UPI app.

The feature offers flexibility, enabling users to update their mobile number or UPI ID within the transaction time limit. This groundbreaking technology, the first of its kind for UPI payments, aims to break down language barriers and make transactions more human-centric and accessible. CoRover’s voice-enabled BharatGPT enhances the multilingual system, supporting Hindi, Gujarati, and other languages, ensuring inclusivity in the payment process.

Vishal Anand Kanvaty, CTO of NPCI, expressed his enthusiasm, stating, “I am really excited for this launch because it addresses India and Bharat. Also, I think it helps all the citizens of this country to make the payment very seamless. I am really excited, we launched UPI123 last year, but the real use cases have started now, and I think I am really excited about this one because we will see this scale happening on this.”

Sanjay Kumar Jain, CMD of IRCTC, added, “As I speak, your payment will not get deducted. You have to speak; you have to raise your voice to get your payment deducted and get the tickets. That’s what we are offering.”

Ankush Sabharwal, Founder and CEO of CoRover.ai, highlighted the natural evolution of technology, saying, “What’s the natural way to interact? The natural way is that we see and talk, so why should technology not talk? In this pursuit, with the help of NPCI and IRCTC, you will now be able to talk to Disha and ask her to get your ticket booked, including payment. It is just a start; you will see everything, even hardware products talking.”

The integration of Conversational Voice Payments into AskDISHA, the AI virtual assistant for IRCTC and Indian Railways, allows users to book tickets and make payments using just their voice. This launch represents a pivotal moment in the evolution of AI, digital payments, and eCommerce, ushering in a more intuitive and convenient booking experience.

The post CoRover, IRCTC, NPCI, Launch Conversational Voice Payments for UPI with BharatGPT appeared first on AIM.

In a collaborated effort with Waymo & University of Southern California, Meta FAIR released its research on the importance of multi-modal generative models. Transfusion aims to unite and simplify the gap between discrete sequence modeling and continuous media generation. 

The Transfusion Model 

The model is trained equally on text and image. Per Meta, Transfusion is more advanced than quantising images and training a language model over discrete image tokens. The model’s performance can be enhanced through “modality-specific” encoding and decoding layers. The model predicts the next word in a sequence. Trained on improving predictions, it reduces the difference between guessing and actual words. It is imperative to note that with 7 billion parameters and 2 trillion multi modal tokens, Transfusion is at par with other larger models that create image and text – and outperforms models like DALL-E 2 and SDXL. It works better than Chameleon as it takes lesser computing power and generates better results. 

One limitation is perhaps that diffusion models do not perform at par with traditional language models. A lot of research is yet to be done in this area to improve overall performance.  

Transformer’s Uniqueness & the Future of Innovation in AI Research 

What differentiates Transformer from the rest is its unified architecture that runs end to end to generate text and images. Existing models like Flamingo, LLaVA, GILL, and DreamLLM combine separate architectures for different types of data, which are trained separately. 

The goal of this Transfusion is to synergise two modalities in a single joint model – with each of them fulfilling their objective. The incentives are that these are versatile, resource efficient, and cost effective for handling different types of data without any additional costs. 

The post Meta FAIR Releases Transfusion for Multimodal AI Training appeared first on AIM.

Semiconductor Fabless Accelerator Lab (SFAL) CoE for fabless semiconductors, has incubated 43 startups and benefited over 90 companies in the last five years. 

At a recent media briefing at the Vikasa Soudha, Karnataka IT/BT Minister Priyank Kharge highlighted the remarkable achievement and said, “With INR 21.56 crore investment, we’ve supported 200+ innovations, creating 800 ESDM jobs. SFAL startups have raised INR 140 crore, valued at INR 500 crore collectively.”

Success stories from SFAL include Calligo Technologies’ groundbreaking POSIT chip, Morphing Machines’ impressive $2.76 million funding round, and AB Circuits and Research Labs (ABCRL)’s pioneering automotive chip innovations.

As India becomes the world’s 2nd largest semiconductor consumer, Karnataka is positioning itself as a hub for semiconductor innovation and skilling, driving India’s transformation from a service-based to a product-driven economy.

India’s Semiconductor Plan

The Union Budget 2024-25 allocated Rs 21,936 crore to the Ministry of Electronics and Information Technology (MeitY), a 52% increase from the revised estimates of FY24, which were at Rs 14,421 crore.

This boost in funding underscores the government’s commitment to run multiple incentive schemes and programmes in key areas including semiconductor manufacturing, electronics production, and the India AI Mission. 

One of the focuses of this budget is the initiative to establish advanced semiconductor facilities in India, including compound semiconductors, silicon photonics, sensors fab; discrete semiconductor fab; semiconductor assembly, testing, marking and packaging (ATMP), and outsourced semiconductor assembly and test (OSAT) facilities. 

This initiative alone saw its allocation nearly triple to INR 4,203 crore, up from INR 1,424 crore in FY24.

The post ‘Karnataka Will Lead India’s Semiconductor Revolution,’ says State IT/BT Minister Priyank Kharge appeared first on AIM.

Chennai based global technology company Zoho enters the payments sector with the launch of Zoho Payments, a comprehensive solution designed to facilitate online transactions for businesses. Announced during the Global Fintech Fest, the announcement highlights Zoho’s commitment to enhancing financial management for its users.

This unified payment solution aims to streamline the payment process, reducing cart abandonment rates enhancing overall customer experience. With Zoho Payments, businesses ensuring compliance with PCI-DSS level 1 standards can now accept various payment methods, including UPI, net banking, and credit cards.

Sivaramakrishnan Iswaran, Zoho’s Global Head of Finance and Operations, emphasised the importance of integrating banking and payments within business finance “At Zoho, we strongly believe that business finance, banking and payments should work together”, he said. The new platform, Zoho Payments, integrates with leading banks and offers a range of financial management features, including B2B payments and invoice financing.

B2B Payments Powered by BBPS

A notable feature of Zoho Payments is its B2B payments capabilities, powered by the Bharat Bill Payment System (BBPS) from NPCI Bharat BillPay Limited. This integration allows businesses to manage invoice presentment and payments seamlessly. Invoices generated through Zoho Books can be automatically recorded in the buyer’s accounting system, streamlining the reconciliation process and reducing manual intervention. Additionally, businesses can access financing for unpaid invoices through the TReDS platform, further enhancing cash flow management.

Market Growth

As the Indian B2B payments market is projected to reach USD $11 trillion by 2026, Zoho Payments offers a holistic financial management solution for businesses and aims to capitalise on this growth by providing businesses with a robust platform to integrate payments into their existing workflows.

The post Zoho Unveils Unified Payment Solution with Zoho Payments appeared first on AIM.

Ola Electric will make its entire range of products available on the Open Network for Digital Commerce (ONDC) platform starting next week. This move aligns with ONDC’s vision to revolutionise the e-commerce landscape by promoting open and inclusive digital marketplaces.

“All Ola Electric products will be available on ONDC from next week onwards. ONDC is the future of commerce,” posted Ola chief Bhavish Aggarwal on X. 

The ONDC initiative, backed by government support, aims to create a unified network for various commerce platforms, enabling greater transparency and reducing the dominance of major e-commerce players. By joining ONDC, OlaElectric seeks to leverage this emerging infrastructure to expand its market reach and facilitate easier access to its products for consumers across India.

Aggarwal has been bullish on ONDC and recently said, “At Ola, we’re building the future of commerce with ONDC, which will enable kiranas and small merchants to reach consumers through digital networks. ONDC is the future.”

This comes after he declared ONDC “the UPI moment for e-commerce” at the recent Ola Sankalp 2024 event.

At the same event, he also announced that the company would change its name from Ola Cabs to Ola Consumer, as it expands its services beyond cabs. “We are going to offer a much broader suite of consumer services. Many of you have already used some of these. As I said, our ambition is to truly make commerce accessible, affordable, and efficient,” said Aggarwal.

For the same, Ola Consumer is integrating with ONDC to reduce the cost of commerce and expand service categories. Currently, Ola users can order food and beverages via ONDC through the Ola app, with plans to expand to groceries and fashion items. 

The company has also integrated a food delivery plugin within its app, enabling users to access a variety of restaurants and food brands listed on ONDC. This feature is in the pilot phase, available to Ola employees and a select group of consumers.

Aggarwal also plans to leverage the power of ONDC to take on the likes of Zepto and Blinkit. The company plans to implement fully automated dark stores and fulfilment centres to revolutionise warehousing and improve the commerce supply chain.

In addition, Ola is focusing on sustainable logistics by electrifying its delivery operations, which is expected to reduce logistics costs by about 50% and create additional jobs.

The post Ola Electric to Launch Products on ONDC Platform Next Week appeared first on AIM.

Andrew Ng’s DeepLearning.AI in collaboration with Google Cloud launched a new course named ‘Large Multimodal Model Prompting with Gemini’ aiming to provide learners with essential skills in using text embeddings for various applications in AI. Unlike Large Language Models (LLMs) that accept text prompts as input, this course aims at teaching how Large Multimodal Models (LMMs) like Gemini can integrate text, images and video as input prompts to deliver more comprehensive and accurate outputs. 

Andrew Ng with the offering of this course aims to teach effective techniques for multimodal promotion, the differences and use cases for Gemini Nano, Pro, Flash, and Ultra models. Taught by Erwin Huizenga, developer advocate for Gen AI on Google Cloud, it focuses on how to integrate Gemini with external APIs using function calling and best practices for creating multimodal applications.

Check out the course here

Innovation First

The ability to have a model that can reason across text and image is quite news. Before LMMs such as Gemini, one effective way to work with image and text simultaneously, would be to use an image captioning model to describe the image feeding that caption to a LLM. But with LMMs these images can be interpreted directly as inputs along with text by the AI.

DeepLearning.AI is offering a flurry of new courses, this month alone the company has unveiled a course on federated learning, allowing secure training on private data, alongside a partnership with Flower Labs. A significant collaboration with Upstage is helping students to efficiently pretrain large language models, including cost-saving techniques like depth up-scaling.

With this diverse range of offerings, DeepLearning.AI is providing a valuable resource for those seeking to advance their skills in the ever-evolving field of AI. Rounding out the recent launches is a course on optimising retrieval augmented generation (RAG) in partnership with MongoDB. This course equips learners with the skills for building efficient and scalable RAG applications, covering techniques like vector search and prompt compression.

Also Read: Andrew Ng’s DeepLearning.AI Unveils New Course on Building AI Applications with Haystack

The post DeepLearning Offers New Course on Large Multimodal Model Prompting with Gemini appeared first on AIM.

Imagine a world where traffic lights don’t just blink on and off, but work together as a team, making real-time decisions to keep traffic flowing smoothly and safely. This is the idea that XTraffic is bringing to life. 

By upgrading traffic lights to operate as an intelligent, coordinated system, XTraffic is not only reducing traffic jams but also giving city managers and citizens back their most valuable resource: time.

Behind this solution are the minds of Luke Adams, Everett Ivy, and Brian Payne, the founders of XTraffic. Their vision is about more than just reducing traffic; it’s about making smarter use of infrastructure and creating cities that are better equipped to handle the challenges of the future.

XTraffic is at the forefront of this movement, offering cities a smart, scalable solution that turns ordinary traffic lights into a self-optimising network. These upgraded lights communicate and coordinate with each other, adjusting to traffic conditions, emergencies, and other factors in real time. 

What makes XTraffic truly revolutionary is its approach to affordability and ease of implementation. Instead of requiring cities to spend massive amounts on new infrastructure, XTraffic’s system integrates seamlessly with existing technology. This lowers the barrier to entry, making it possible for cities to quickly deploy the system at a fraction of the cost—whether it’s for a single intersection or an entire metropolitan area.

Now, as part of Y Combinator, XTraffic is ready to connect with cities, investors, and partners who share their vision for smarter, safer urban environments. 

Project Green Light

Google’s Project Green Light is also working on similar lines to help cities, improve traffic flow at intersections and reduce stop-and-go emissions. 

With AI and insights from Google Maps driving trends, Green Light demonstrates a deep understanding of global road networks. This enables it to model traffic patterns and build intelligent recommendations for city traffic engineers to optimise traffic flow. 

Currently, Project Green Light operates at 70 intersections in 12 cities, including Bengaluru, Hyderabad, and Kolkata. 

The post XTraffic is Turning Ordinary Traffic Lights into Smart, Coordinated System for Safer, Smoother Cities appeared first on AIM.

On Wednesday, California lawmakers approved a contentious AI safety bill, which now requires a final procedural vote. Afterward, the decision will rest with Governor Gavin Newsom, who has until September 30 to either sign the bill into law or veto it.

California’s Senate Bill 1047, a proposed AI regulation, has sparked intense debate in Silicon Valley, drawing both praise and criticism from tech leaders, lawmakers, and AI experts. 

xAI chief Elon Musk has also voiced support, emphasising the need for regulation to prevent AI misuse. “This is a tough call and will make some people upset, but, all things considered, I think California should probably pass the SB 1047 AI safety bill,” he posted on X. 

Musk said that for more than 20 years, he has supported AI regulation, drawing parallels to how society regulates any technology or product that could potentially pose a risk to the public.

The bill, introduced by State Senator Scott Wiener, seeks to implement strict safety measures for large-scale AI models, aiming to prevent potential catastrophes, but critics argue it could hinder innovation.

The legislation requires developers of significant AI models—those costing over $100 million to train—to perform comprehensive safety testing before releasing them to the public. 

It also mandates an “emergency stop” feature to shut down AI systems in critical situations and obligates developers to report any safety incidents to California’s Attorney General within 72 hours. A new state agency, the Frontier Model Division, would oversee compliance, with penalties of up to $30 million for repeated violations.

Proponents, including AI luminaries Geoffrey Hinton and Yoshua Bengio, believe the bill is crucial to addressing AI risks comparable to pandemics or nuclear threats, potentially setting a national standard for AI safety. 

Conversely, the bill has faced strong opposition from major tech companies like Google, OpenAI, and Meta, who argue it could stifle innovation and drive talent out of California. Critics, such as U.S. Representative Nancy Pelosi and AI expert Fei-Fei Li, caution that the bill’s requirements could disproportionately burden smaller companies and slow technological advancement. They advocate for federal regulation to avoid inconsistencies across states.

OpenAI has publicly opposed SB 1047, stating that it poses a threat to AI’s growth and could push entrepreneurs and engineers to relocate.

The bill has passed the California Appropriations Committee with amendments and is awaiting a final vote in the state assembly. If signed into law by Governor Gavin Newsom, it would be the first AI regulation of its kind in the U.S., potentially shaping AI governance nationwide.

As California, a key hub for AI innovation, seeks to balance technological advancement with safety, the fate of SB 1047 could have significant implications for the global tech industry and the future of AI regulation.

The post California’s SB 1047 Bill Sparks AI Civil War appeared first on AIM.

IT giant, Infosys, has announced a collaboration with the Stanford University Institute for Human-Centered Artificial Intelligence (Stanford HAI), a leading organisation shaping the future of AI. As a new member, Infosys will work closely with the institute’s esteemed faculty and researchers on pioneering AI initiatives.

James Landay, Stanford HAI Co-Director, expressed enthusiasm about the collaboration and said, “Together, we hope to push the boundaries of AI research and develop solutions that are ethical, trustworthy, and impactful.”

This partnership will focus on key areas including responsible AI, addressing the complex technical, policy, and governance challenges that enterprises face. Additionally, Infosys aims to enhance business processes through AI/ML and optimise AI models for cost efficiency and scalability using a specialised transformer-based approach that requires minimal data.

Mohammed Rafee Tarafdar, CTO of Infosys, added,“ By incorporating the latest AI research and proven practices into Infosys Topaz, we aim to drive innovation and deliver transformative AI solutions for our clients and industry.”

Earlier in May, Infosys and ServiceNow announced a strengthened collaboration to transform customer experiences with generative AI‑powered industry solutions, at ServiceNow’s annual customer and partner event Knowledge 2024.

And in June, GitHub partnered with Infosys to launch the first GitHub Center of Excellence in Bangalore. This initiative aimed to leverage AI and advanced software solutions to drive global economic growth.

Infosys Collaborates with Microsoft

In early 2023, Infosys and Microsoft joined forces to push the boundaries of what’s possible with AI. 

Nicole Dezen, Chief Partner Officer, Microsoft Corp, stated, “We’re pleased to expand our collaboration with Infosys to deliver innovative solutions, utilising Azure OpenAI Service and Azure Cognitive Services, that will help customers develop new business models, and realise new revenue streams.”

The post Infosys Joins Stanford HAI Affiliate Program to Accelerate Enterprise AI Research Initiatives appeared first on AIM.

On 27th August 2024, Google released a paper detailing the capabilities of their new model GameNGen. This research was authored by Dani Valevski (Researcher at Google Research), Yaniv Leviathan (Engineer at Google Research), Moab Arar (PhD candidate at Tel Aviv University), and Shlomi Fruchter (Engineer at Google DeepMind). 

GameNGen’s Capabilities 

Powered entirely by a neural model, GameNGen is one of the pioneer search engines that allows high-quality, real-time interactions with complex environments over long periods. “GameNGen can interactively simulate the classic game DOOM at over 20 frames per second on a single TPU. Next frame prediction achieves a PSNR of 29.4, comparable to lossy JPEG compression”, per Google. The training occurs primarily in two phases:  (1) an RL agent learns to play the game, with training sessions being recorded, and (2) a diffusion model is trained to predict the next frame using the sequence of previous frames and actions. 

This approach, while still in its nascent stage, is testament of a new paradigm where games are weights of a neural model, not lines of code, in an era where video games are programmed by humans. GameNGen reveals that a neural model can interactively run a complex game like DOOM on existing hardware. 

Architecture & How It Works

In order to collect data efficiently, an RL agent plays the game, and its actions and observations are recorded as training data. For more enhancement and consistency, the Stable Diffusion model is trained to predict the next game frame using previous frames and actions. Lastly, the model’s decoder is fine-tuned to improve image quality and reduce visual artefacts. 

AI in Gaming

Jack O Brien, ex Google, remarked on how this research is seminal and opens up new possibilities for the world of video generation and the use of AI.  Jensen Huang, Founder of NVIDIA, has been bullish on the inevitability of the world of gaming and the role played by AI in making it a reality. “We used AI to revolutionise computer graphics. Graphics enabled AI, and now AI is saving graphics”, Huang said. Last year, OpenAI’s acquisition of Global Illumination, a gaming and design startup and the creators of Biomes, highlights OpenAIs gaming focus for AI simulation and use. As AI research advances, companies will increasingly test their models in game worlds before real-world deployment.

The post Google Unveils GameNGen: Integrating Generative AI into Gaming  appeared first on AIM.

AI4Bharat is seeking contributors for Chitralekha, an open-source Video Transcreation platform supported by the EkStep Foundation. The platform, initially developed for video annotation, allows users to auto-generate and edit audio transcripts in Indic languages.

“We’re thrilled to invite you to contribute to Chitralekha, an innovative open-source video transcreation platform built by AI4Bhārat and funded by the EkStep Foundation,” said Ishvinder Sethi, Project Officer at AI4Bhārat.

Chitralekha’s features include subtitle generation and download, audio/video dubbing, and video translation across various Indic languages. The project, built on AI models developed in-house by AI4Bhārat, is open to enhancements and new feature integrations.

The initiative offers contributors the chance to gain experience in a live project environment, build their portfolios, and contribute to the Indian Language AI landscape. AI4Bhārat encourages participation in community discussions on the platform’s GitHub page, focusing on potential use cases and desired features.

This is an unpaid opportunity for open-source enthusiasts to help expand Chitralekha’s capabilities.

Chitralekha is an open-source platform designed for video transcreation across various Indic languages, leveraging machine learning models for ASR (Automatic Speech Recognition) for transcription, NMT (Neural Machine Translation) for translation, and TTS (Text-to-Speech) for voice-over.

The platform supports multiple input sources, such as YouTube and local files, and offers various options for transcription generation, including models, source captions, custom subtitle files, and manual creation.

For translation and voice-over generation, users can utilise models or manually created content. Currently, Chitralekha supports voice-over for single-speaker videos, with multi-speaker support under development.For more details or to contribute, visit the GitHub page here or try out the platform here.

The post AI4Bharat Invites Contributors to Chitralekha Open-Source Project appeared first on AIM.

NTT DATA today announced the launch of its new Accelerated AI Platform in India today. This solution is designed to simplify the adoption of artificial intelligence (AI) for businesses of all sizes, from startups to large enterprises.

The platform will address the challenges faced by organisations looking to leverage AI, including the complexity of implementation, resource-intensive nature of AI projects, and the need for specialised expertise. 

“We understand that AI initiatives can be complex and resource-intensive,” said Avinash Joshi, Chief Executive Officer, India, NTT DATA. “The Accelerated AI Platform tackles these challenges directly. This launch comes at a time when there is a growing demand for high-performance computing resources, and India’s current growth rate in AI adoption is impressive.”

To meet the infrastructure requirements for the platform, NTT DATA partnered with Neysa, an AI cloud and platform-as-a-service (PaaS) company funded by NTT DATA. Neysa provides the high-end GPUs necessary to run advanced AI models and the essential components to deliver the infrastructure and associated services as a service. This includes identifying and delivering the right GPU mix for training and inference workloads.

NTT DATA had also launched a new Edge AI platform in July designed to speed up the integration of IT and operational technology. The platform brings AI processing closer to where data is generated, enabling faster decision-making and more efficient operations. By unifying various IoT devices, systems, and data, this fully managed solution supports real-time analytics, secure AI deployment, and the advancement of Industry 4.0 technologies across different industries.

The post NTT DATA Launches Accelerated AI Platform to Simplify Complex Projects in India  appeared first on AIM.

Dell Technologies recently announced that it is accelerating skills development in India by offering its GenAI Foundations course at upcoming forums in Mumbai and Bengaluru, equipping professionals with AI/ML expertise and providing certifications to drive transformative business outcomes in the AI era.

The specialised training program will aim to equip individuals with the necessary knowledge and expertise to leverage AI and machine learning technologies effectively driving transformative business outcomes. By focusing on GenAI, Dell Technologies is addressing the growing demand for skilled professionals in this rapidly evolving field.

Check: GenAI Courses by ADasci

The first 250 attendees who register for the Dell Tech Forums in Mumbai and Bangalore will have the opportunity to take an exam and receive a Dell Proven Professional certification, demonstrating their proficiency in GenAI. The tech solutions group also offers a wide range of technical courses including large language models, data engineering and cybersecurity. Coming this fall will be a GenAI governance and security course.

Alok Ohrie, Dell Technologies India president and managing director laid emphasis on how Dell  Technologies is committed to empowering organisations with the skills they need to thrive in the AI era saying, “By investing in skills development, we are contributing to the broader digital transformation journey across India. At Dell, we ensure our customers have the skills and knowledge they need to lead in the age of AI”

Pioneering Enterprise AI 

Dell Technologies has been at the forefront of  driving generative AI initiatives across healthcare, transportation, and public services by partnering with organisations like Northwestern Medicine, Duos Technologies, and the City of Amarillo to deploy AI solutions that improve patient outcomes, enhance railway safety, and democratise access to city services.

Recently Wipro and Dell Technologies partnered to accelerate enterprise AI by offering localised AI processing solutions that cater to stringent data privacy requirements, alongside democratising access to advanced AI capabilities and enabling flexible, scalable deployment across cloud, data center, and edge environments.

In April 2024, Dell Technologies and NVIDIA joined forces to deliver end-to-end generative AI solutions, combining advanced infrastructure and professional services to accelerate AI adoption across industries, unlocking a potential $4.4 trillion annual impact as businesses harness the transformative power of data-driven AI.

Discover: List of Top 10 GenAI Courses

The post Dell Technologies to Offer GenAI Foundations Course and Professional Certifications in India appeared first on AIM.

Cerebras Systems today announced its new AI inference solution, Cerebras Inference, which it claims is the fastest in the world. The solution delivers 1,800 tokens per second for the Llama 3.1 8B model and 450 tokens per second for the Llama 3.1 70B model, making it 20 times faster than NVIDIA GPU-based hyperscale clouds. 

Cerebras Inference is priced at 10 cents per million tokens for Llama 3.1 8B and 60 cents per million tokens for Llama 3.1 70B, and is available to developers through API access.

The solution is powered by the third-generation Wafer Scale Engine (WSE-3), which enables it to run Llama 3.1 models 20 times faster than GPU solutions at one-fifth the cost. The WSE-3 integrates 44GB of SRAM on a single chip, eliminating the need for external memory and providing 21 petabytes per second of aggregate memory bandwidth, which is 7,000 times greater than that of an NVIDIA H100 GPU.

Cerebras addresses the inherent memory bandwidth limitations of GPUs, which require models to be moved from memory to compute cores for every output token. This process results in slow inference speeds, particularly for large language models like Llama 3.1-70B, which has 70 billion parameters and requires 140GB of memory.

Cerebras Inference supports models from billions to trillions of parameters. For models exceeding the memory capacity of a single wafer, Cerebras splits them at layer boundaries and maps them to multiple CS-3 systems. Larger models, such as Llama3-405B and Mistral Large, are expected to be supported in the coming weeks.

Cerebras Inference operates using 16-bit model weights, preserving the accuracy of the original Llama 3.1 models released by Meta. The solution offers developers chat and API access, with an initial offering of 1 million free tokens daily.

The company emphasises that the speed of inference is not just a matter of performance metrics but also enables more complex AI workflows and real-time large language model intelligence. Cerebras notes that techniques like scaffolding, which require significantly more tokens at runtime, are only feasible with its hardware.

Cerebras Inference aims to set a new standard for large language model development and deployment, offering both high-speed training and inference capabilities. The company anticipates that the speed and cost advantages will open up new possibilities for AI applications.

The post Cerebras Launches Fastest AI Inference Solution, Claims 20x Speed Advantage Over NVIDIA appeared first on AIM.

SambaNova Systems has achieved a new performance milestone, setting a world speed record with Meta’s Llama 3.1 405B model, processing 114 tokens per second. The performance, verified by Artificial Analysis, outpaces other providers by over four times, positioning SambaNova as a leader in AI speed and efficiency.

“I’ve been playing with SambaNova Systems‘s API serving fast Llama 3.1 405B tokens. Really cool to see the leading model running at speed. Congrats to Samba Nova for hitting a 114 tokens/sec speed record,” said DeepLearning.ai founder Andrew Ng.  

The benchmark was set using a single 16-socket node, operating with full 16-bit precision on SambaNova’s custom RDU chips. This advancement addresses the challenge of balancing quality and speed in large models like Llama 3.1 405B, enabling the deployment of the model in more speed-sensitive applications, such as customer support and AI agents.

George Cameron, Co-Founder of Artificial Analysis, confirmed the record, saying that SambaNova’s platform reduces the trade-off between model size and operational speed, making it viable for real-time applications.

SambaNova’s fourth-generation RDU chip, the SN40L, plays a critical role in this achievement, facilitating real-time processing that opens up new enterprise use cases. These include intelligent document processing, real-time AI copilots, and explainable AI, all of which benefit from the platform’s speed.

The company is offering a demo of the Llama 3.1 405B model on its website and is inviting developers to access its APIs for building enterprise-level AI applications.

SambaNova Systems is a technology company specializing in artificial intelligence (AI) hardware and software solutions. Founded in 2017 in Palo Alto, California, by Kunle Olukotun, Rodrigo Liang, and Christopher Ré, the company provides purpose-built solutions for deep learning and AI applications.

The company’s technology is built around the SN40L chip, which features a reconfigurable dataflow architecture. This design optimizes data movement and reduces latency, making it highly efficient for AI tasks compared to traditional GPU-based systems.

The post SambaNova’s Llama 3.1 405B Model Hits 114 Tokens Per Second, Setting Speed Record appeared first on AIM.

Amazon CEO Andy Jassy recently revealed that by leveraging Amazon Q, the company was able to save 4,500 developers-years of work. “Yes, the number is crazy, but real,” he posted on X.

With Amazon Q, the company has significantly cut down the time needed to update Java applications. “The average time to upgrade an application to Java 17 plummeted from what’s typically 50 developer days to just a few hours,” he said. 

He added that in under six months, the company has been able to upgrade more than 50% of its production Java systems to modernised Java versions at a fraction of the usual time and effort. “Our developers shipped 79% of the auto-generated code reviews without any additional changes.”

“The benefits go beyond how much effort we’ve saved developers. The upgrades have enhanced security and reduced infrastructure costs, providing an estimated $260 million in annualised efficiency gains,” he claimed.

Amazon plans to expand the transformation capabilities of Amazon Q, with further developments aimed at additional developer tools. This advancement marks a notable shift in how enterprises handle software updates and maintenance.

Amazon Web Services announced the general availability of Amazon Q, earlier this year. The chatbot is available in three forms: Amazon Q for developers, Amazon Q for businesses, and Amazon Q apps. 

Amazon Q not only generates highly accurate code, it also tests, debugs, and has multi-step planning and reasoning capabilities that can transform (e.g., perform java version upgrades) and implement new code generated from developer requests. 

The chatbot also makes it easier for employees to get answers to questions across business data such as company policies, product information, business results, code base, employees, and many other topics by connecting to enterprise data repositories to summarize the data logically, analyze trends, and engage in dialog about the data.

The post Amazon Claims to have Saved 4,500 Developer-Year of Work with Q appeared first on AIM.

Rufus, a generative AI expert, aimed at enhancing customer experience by making shopping more easier and convenient, is now launched in India on 27th August 2024. With its initial rollout in the US earlier this year, Rufus is now open to only select customers for now. “Customers in the U.S. have already asked Rufus tens of millions of questions, and we’re excited to introduce it in India.” said Amazon in a blog post.

Rufus’ Generative AI Capabilities 

In a blog post, Amazon detailed how Rufus seamlessly transforms the customer’s shopping experience through a wide range of queries. The aim is to maximise AI to effectively meet the customers’ needs. 

Per Amazon, customers using Rufus on Amazon.in can enhance their shopping experience in several ways. They are offered tailored information, product comparisons, personalised recommendations, among other things.

Rufus & India’s E-commerce Space

Rufus is not new to this market. In 2023, Myntra collaborated with Microsoft to launch “My Stylist” a feature that recommends complete, shoppable looks tailored for fashion-forward customers. In the same year, they launched MyFashionGPT, a tool that enables users to search using natural language to give relevant suggestions. Even Flipkart leveraged LLMs, through AI-driven ‘Flippi’ to simplify shopping experience. As of last week, Seematti, a retail brand, launched India’s inaugural AI fashion ambassador, Isha Ravi, who as an Indian young woman with a passion for colour and travel, represents the brand in the Indian landscape. 

In the past 25 years, Amazon has built their suite of AI backed customer offerings. Personalised recommendations, like Alexa, have been successful for Amazon in the past. 

Amazon’s AI generated review highlights and product listings have enabled customers to shop better in the past. Rufus is at the heart of this AI expansion. Speaking on the potential of Generative AI  in the e-commerce space, Amazon recognises that while Generative AI is still in its early stages, they are committed to improving their models and make Rufus better over time. In the upcoming weeks, the beta version of Rufus will be extended to more customers. 

The post Rufus, Amazon’s AI Assistant, is Now Live in India appeared first on AIM.