#035 AWS’s GenAI Strategy, IBM CTO On AI, Dubai's 90% discount for AI,Fine-tuning GPT-3.5 Turbo
Fresh & Hot curated AI happenings in one snack. Never miss a byte 🍔
This snack byte will take approx 7 minutes to consume.
AI BYTE # 1 - AWS’s GenAI Strategy: A Three-Layered Approach to Cloud Computing
⭐ AWS, Amazon’s cloud computing division has been strategically leveraging Generative AI (GenAI) to reinvigorate its cloud business and strengthen its leadership position in the industry.
GenAI is a branch of AI that can generate novel and realistic content, such as text, images, audio, and video. GenAI has transformative applications across various industries, such as e-commerce, entertainment, education, and healthcare.
AWS has invested in three core layers of the GenAI opportunity: AI chips, AI models, and AI-powered applications.
AI Chips
The first layer focuses on the computational power required for training and running AI models.
AWS has developed its own AI chips, Trainium and Inferentia, to cater to the soaring demand for AI processing. These chips offer a competitive price-performance proposition and could potentially rival Nvidia’s GPUs.
AI Models
The second layer involves offering AI models as a service, allowing businesses to build applications atop these models.
AWS’s own AI model, Bedrock, can be customized with proprietary data and managed by AWS. Bedrock can also tap into AI models from various companies, such as OpenAI’s GPT-4 and DeepMind’s Gato, amplifying its potential impact.
AI Applications
The third layer entails creating valuable AI-powered applications that cater to customer needs. One successful example is Amazon CodeWhisperer, an AI-backed coding assistant that aids developers in boosting productivity.
AWS aims to further enhance customer experience by developing more AI applications that span personalized product recommendations, customer service enhancement, new product ideation, fraud detection, content creation, and beyond.
AWS faces fierce competition across all three layers of the GenAI landscape from traditional manufacturers like Nvidia and AMD, newcomers like Google and Microsoft, and rivals like GitHub Copilot and OpenAI.
However, AWS’s focus on GenAI could be the catalyst for its resurgence and sustained growth in an ever-evolving market.
As AI continues to redefine industries, AWS’s strategic integration of AI technologies positions it as a frontrunner in the cloud computing landscape.
By investing in AI chips, offering AI models as services, and crafting innovative AI-powered applications, AWS is poised to shape the future of AI and cloud computing.
AI BYTE # 2 - The Future Of Generative AI: IBM’s CTO Shares His Insights On Cost, Regulation, And Innovation
⭐ Large language models (LLMs) such as ChatGPT have been making headlines in the AI community for their ability to generate realistic text and images.
However, these models also come with a hefty price tag and a host of regulatory concerns.
In a recent interview with CGTN, Xie Dong, the CTO of IBM Greater China Group, revealed the financial sustainability issues of ChatGPT-like models.
He cited a report that estimated the operational costs of maintaining ChatGPT at around $700,000 per day, regardless of user engagement. This staggering figure highlights the resource-intensive nature of training and deploying LLMs, which require a massive number of GPUs.
Xie also acknowledged the revolutionary impact of ChatGPT and similar technologies, which he described as a tipping point for generative AI. However, he also cautioned that generative AI poses ethical and regulatory challenges, such as data protection, privacy, and reliability.
To address these challenges, Xie shared IBM’s strategic approach to AI development. Instead of focusing solely on creating ChatGPT-like models, IBM’s researchers have been exploring foundation models. These models are trained on vast sets of unlabeled data, enabling them to cater to various tasks and domains.
Building on foundation models, IBM aims to craft specialized and sophisticated models tailored to specific enterprise use cases. Xie emphasized that this approach allows IBM to provide more value to its clients, enabling them to address domain-specific challenges effectively.
IBM’s new platform, powered by foundation models, also introduces generative AI capabilities.
The company anticipates that the Generative AI market will burgeon, projecting a value of $8 billion by 2030. Xie cited an impressive figure of over 85 million job vacancies in this burgeoning sector.
IBM’s vision for AI is not only innovative but also responsible. Xie highlighted the importance of compliance with data protection regulations and safeguarding privacy. He stressed the significance of reliability in both the data used to train AI models and the models themselves.
IBM’s CTO has offered a comprehensive overview of the challenges and opportunities posed by LLMs and generative AI.
His insights reveal how IBM is tackling financial and ethical issues while delivering specialized and sophisticated AI solutions to its clients.
AI BYTE # 3 - How Dubai Aims to Become a Tech Hub with 90% License Discount?
⭐ Dubai, one of the most dynamic and innovative cities in the world, has launched a bold strategy to attract AI and Web3 businesses to its shores.
The city is offering a 90% discount on commercial licenses for these sectors, as well as providing them with a state-of-the-art facility called the Dubai AI and Web 3.0 Campus.
This initiative is part of Dubai’s vision to become a major tech hub in the Middle East and North Africa (MENA) region.
The Dubai AI and Web 3.0 Campus is not just a place to get a license, but also a place to grow and thrive. The campus boasts of cutting-edge AI labs, transformative training programs, and the latest in tech hardware.
The campus aims to foster a culture of innovation and excellence among next-gen tech companies and to help them scale their businesses globally.
Dubai is also embracing the cryptocurrency wave, as it has endorsed several crypto-based ventures and granted them operational licenses.
One of these ventures is Laser Digital Middle East, a branch of Nomura, which has received a license from Dubai’s Virtual Asset Regulatory Authority. This license allows Laser Digital to provide broker-dealer services and oversee virtual assets, marking a milestone in UAE’s digital journey.
The city is witnessing a surge in crypto and blockchain enterprises, with over 400 current businesses expected to grow to more than 1,000 by the end of the year.
This remarkable growth signifies a growing trust and interest from investors in Dubai’s crypto and blockchain initiatives6. As Web 3.0 technologies continue to shape the future, Dubai is positioned to witness a boom in startups harnessing this potential.
The ensuing wave promises to usher in a plethora of new job opportunities and unique investment ventures in the region, reinforcing Dubai’s image as a frontrunner in the digital age.
Dubai is not only offering attractive incentives for AI and Web3 businesses but also creating an ecosystem that supports their development and success. Dubai is setting an example for other cities to follow, as it aims to become a leader in the next-gen tech revolution.
AI BYTE # 4 - Fine-tuning GPT-3.5 Turbo With Your Own Data: OpenAI’s Latest Offering
⭐ OpenAI has announced a new feature that allows enterprises to fine-tune its GPT-3.5 Turbo model with their own data.
GPT-3.5 Turbo is a large language model (LLM) that can generate realistic text and images, and is optimized for chat and other tasks. By fine-tuning the model on company-specific data, developers can improve its performance and create unique and differentiated experiences for their customers.
Fine-tuning is a process of training a pre-trained model on a new dataset to adapt it to a specific domain or task. OpenAI claims that fine-tuning GPT-3.5 Turbo can make it more capable and cost-effective than the flagship GPT-4 model in certain narrow tasks.
It also says that fine-tuning can help the model follow instructions better, format responses in a given way, and use a specific tone or language.
To fine-tune GPT-3.5 Turbo, developers need to prepare their data, upload the files, and create a fine-tuning job. Once the job is finished, the model is ready to be used in production with the same rate limits as the underlying model.
OpenAI charges $0.0080 per 1,000 tokens for training, $0.0120 per 1,000 tokens for input usage, and $0.0120 per 1,000 tokens for outputs.
OpenAI assures that the fine-tuning process is safe and respects the data privacy of the users. It uses its Moderation API and a GPT-4 powered moderation system to detect unsafe training data that conflict with its safety standards.
It also states that the data sent in and out of the fine-tuning APIs and systems is owned by the user and is not used for training any other model besides the user’s own.
OpenAI plans to open GPT-4, its most advanced generative model that can also understand images, for fine-tuning later this fall.
It also plans to launch a fine-tuning interface to give developers easier access to information about their fine-tuning jobs and models.
OpenAI’s move to offer fine-tuning for GPT-3.5 Turbo is aimed at making its LLM more accessible and useful for enterprises.
However, it also faces competition from other startups and established players that offer their own LLM fine-tuning solutions.
Be considerate of your fellow human beings, and don’t be stupid