Рубрика: AI

Databricks VP of GenAI recently praised Sam Altman for turning OpenAI into a profitable product company.

Contrarian view: @sama is actually doing the right things with Open AI. He has the impossible job of taking an ideologically driven research group and trying to make it into a product company. Goals aside, if he didn't do this, investment would stop.
Arguably, some of this issue…

— Naveen Rao (@NaveenGRao) June 24, 2024

“Contrarian view: Sam is actually doing the right things with Open AI. He has the impossible job of taking an ideologically driven research group and trying to make it into a product company. Goals aside, if he didn’t do this, investment would stop,” he posted on X.

“See, this is the issue…everyone is angry about the fact it was “open” AI. The problem is, that wasn’t going to sustain. It had to become a product company at some point to survive,” he added.

Altman recently told some shareholders that the company is considering changing its governance structure to a for-profit business that the firm’s nonprofit board doesn’t control. One scenario Altman said the board is considering is a for-profit benefit corporation, which rivals such as Anthropic and xAI are using.

OpenAI has more than doubled its annualised revenue to $3.4 billion in the past six months or so, Altman has told staff, a sign that growth in the ChatGPT developer’s business is accelerating despite intensifying competition.

Databricks recently hosted its Data+ AI Summit 2024. The company announced a host of new generative AI capabilities and a major push to its open-source strategy

The new offerings, such as Mosaic AI Model Training, Mosaic AI for RAG, and Mosaic AI Gateway, in addition to open-sourcing their Unity Catalog, aim to help enterprises build high-quality, domain-specific AI applications.

“We want to help people get the best quality possible in their domain for their GenAI application,” said CTO and co-founder Matei Zaharia in an exclusive interview with AIM. “And to do that, we see a lot of companies are building what we call compound AI systems.”

The compound AI systems involve multiple components, such as calls to different models, retrieval of relevant data, use of external APIs and databases, and breaking problems into smaller steps.

At the same time, Databricks is also focusing on open-source models. Last year, Databricks acquired Mosaic ML, and earlier this year, the company released its first open-source model, DBRX. “DBRX was the best model in the world for two whole weeks, and thank you, Mark Zuckerberg, for releasing that model two weeks later,” said Databricks chief Ali Ghodsi, jokingly, about the release of Llama 3.

The post Databricks’ Naveen Rao Praises Sam Altman for Making OpenAI a Product Company appeared first on AIM.

At the recently held World Tour Essentials event in Mumbai, Salesforce, the leading CRM software provider, announced its entry into the public sector domain.

In a press conference, Salesforce India CEO and Chairperson Arundhati Bhattacharya shared the company’s desire to bring its cloud and AI solutions to the government to help them deliver better citizen services to India’s populace.

Bhattacharya revealed that the company has formed a new division focusing on selling its solution to the Union government, the state governments, and Quasi-government agencies.

Under Bhattacharya, who was the first woman to be the Chairperson of the State Bank of India (SBI), Salesforce has grown significantly, recording a 50% revenue jump to INR 6,000 crore in FY 2023. The company’s employee count has also increased from around 2500 to over 10,000 in the country.

Given Salesforce India’s above-average performance, venturing into the public sector seems the next logical step. Nonetheless, announcing the public sector division was the easier part. The biggest test for Salesforce remains convincing the government to leverage its solutions.

( Salesforce press conference held in Jio World Convention Centre, Mumbai)

Government’s Appetite for Public Cloud

In an exclusive interview with AIM at the sidelines of the World Tour Essentials event, Bhattacharya said that Salesforce’s biggest challenge is that the majority of its critical services are on the cloud.

“We do have some on-prem solutions, such as the integration layer in Mulesoft or the graphic analytics layer in Tableau. But other than that, much of our solutions are cloud-based. We need to find out what is the appetite of the government in order to use the public cloud,” she said.

While major government agencies and departments are in dire need of digital transformation, the government has shown an appetite to leverage the cloud and has formed the National Cloud Initiative of India. The initiative, called Meghraj, aims to host various government applications and services on the cloud.

Moreover, the Ministry of Electronic and Information Technology (MeitY) has provided Cloud Service Provider (CSP) empanelment to public cloud companies like AWS. Notably, Salesforce leverages AWS’ cloud infrastructure.

Even though the company refrained from revealing much about this new division, Bhattacharya did say that it is still in its nascent stage and could take a couple of years before we see any possible outcomes. She also said that Salesforce has initiated talks with a few parties.

“The other thing, of course, is that we have to learn how to participate and respond to a request for proposals (RFPs) on time. These documents are often extensive and detailed, requiring a learning curve to understand the response process, which also involves collaborating with partners,” Bhattacharya said.

Data Cloud Could be a Seller for Salesforce

One of the biggest hindrances for many organisations is that most of the data is in silos, and this is even true for many government organisations.

Bringing scattered data together into a cohesive source and deriving meaningful insights remains a significant challenge. This is where Salesforce’s Data Cloud could prove to be a valuable tool when it tries to woo the government.

“Data Cloud isn’t something you need ages and ages to build. We have customers that have actually done it in the span of just 90 days and these are not small customers. These are large-scale customers, nothing in the scale of the government, but definitely large scale.

“When you leverage Data Cloud, it’s something where you are not really storing any data; you are merely getting intelligence from the pools of data that you already have and presenting it in real-time. That’s what it does and that’s something the government could possibly use or be interested in using,” she said.

Data Cloud is one of the fastest, if not the fastest growing, organic innovation to come out of Salesforce. In fact, the company is projected to make half a billion dollars in revenue from Data Cloud in just two years.

Moreover, it also does not matter where the data is sitting. “It’s irrelevant whether the data is stored in Snowflake, DataBricks, or elsewhere. Organisations have a wealth of additional data beyond their typical data warehouse, such as client email interactions, contact centre chats, social media profiles, and ongoing online interactions,” she pointed out.

The data, however, do need to be in digital form. In some government offices, some data could be in physical form. “If the data is not digital, then it becomes a challenge. “ If it is not digitised, we can help you digitise it and ensure that the data that’s coming in in the future stays in a digital format.”

Salesforce Public Sector Solutions go beyond Data Cloud, offering out-of-the-box apps tailored for rapid digital service delivery in government agencies.

Leveraging the Salesforce Einstein Platform, these solutions empower government organisations to future-proof their IT investments with citizen developer tools, AI, and automation. This approach enhances service agility and efficiency and supports digital transformation efforts.

Building Trust in AI

Over Time, multiple thinkers in the AI space, as well as many Salesforce spokespersons, have stressed the importance of building trust in AI.

The primary concerns, be it public or private sector enterprises, regarding trust in AI primarily revolve around the data. According to Bhattacharya, catering to the customer’s data requirement and ensuring there are no mishaps is the best way to build trust in AI.

“If we commit to ensuring data privacy and confidentiality, we must uphold that promise without compromise. It is imperative that our offerings reflect this commitment.

“Moreover, the government will have its own laws and regulations as to where it wants the data, how it wants it to be treated, what data can be accessed, and what data should not be accessed. How should it be masked when at rest? How should it be encrypted when in flight, so all of that has to be very assiduously met,” she said.

AI Needs a Global Compact

Bhattacharya also opines that AI today is mostly self-regulated. While some jurisdictions have some regulations, some don’t have any regulations at all. However, it requires a global compact.

“All countries should agree that for the good of mankind, this is akin to the nuclear non-proliferation treaty. We must now agree on principles and actions that benefit humanity collectively,” she said.

This does not have to be a regulatory body but a collective agreement where countries agree that ‘this is what we will do and this is what we will not do with AI.’

Interestingly, this is not the first time an idea like this has been conceived. Sam Altman, the CEO of OpenAI, also shared similar thoughts.

He urged the US to establish a global organisation similar to the International Atomic Energy Agency (IAEA), focusing specifically on AI regulation.

His comments raised some concerns in India as technology’s impact and implications are shaped by human choices, values, and societal contexts. India also has unique socio-economic challenges, such as income inequality, poverty, and access to basic services.

Furthermore, if such a regulatory body is established, does India secure a seat at the table? If so, does it wield a significant voice in decision-making? Or will it be dominated by Western bureaucrats who will regulate the technology based on their country’s risks and concerns?

Bhattacharya believes India already has a strong voice in the global space. Also, today, the world itself is very interconnected. “You should go to Silicon Valley and see how many Indians are working there. Our chief engineer happens to be from Andhra. He’s obviously an expat and lives in the US, but he’s from India,” she pointed out.

Moreover, we can’t take an approach like China, which bans Western technologies. What is required is a collective approach in which all parties work together for the betterment of humanity.

At the recently held World Tour Essentials event in Mumbai, Salesforce, the leading CRM software provider, announced its entry into the public sector domain.

In a press conference, Salesforce India CEO and Chairperson Arundhati Bhattacharya shared the company’s desire to bring its cloud and AI solutions to the government to help them deliver better citizen services to India’s populace.

Bhattacharya revealed that the company has formed a new division focusing on selling its solution to the Union government, the state governments, and Quasi-government agencies.

Under Bhattacharya, who was the first woman to be the Chairperson of the State Bank of India (SBI), Salesforce has grown significantly, recording a 50% revenue jump to INR 6,000 crore in FY 2023. The company’s employee count has also increased from around 2500 to over 10,000 in the country.

Given Salesforce India’s above-average performance, venturing into the public sector seems the next logical step. Nonetheless, announcing the public sector division was the easier part. The biggest test for Salesforce remains convincing the government to leverage its solutions.

( Salesforce press conference held in Jio World Convention Centre, Mumbai)

Government’s Appetite for Public Cloud

In an exclusive interview with AIM at the sidelines of the World Tour Essentials event, Bhattacharya said that Salesforce’s biggest challenge is that the majority of its critical services are on the cloud.

“We do have some on-prem solutions, such as the integration layer in Mulesoft or the graphic analytics layer in Tableau. But other than that, much of our solutions are cloud-based. We need to find out what is the appetite of the government in order to use the public cloud,” she said.

While major government agencies and departments are in dire need of digital transformation, the government has shown an appetite to leverage the cloud and has formed the National Cloud Initiative of India. The initiative, called Meghraj, aims to host various government applications and services on the cloud.

Moreover, the Ministry of Electronic and Information Technology (MeitY) has provided Cloud Service Provider (CSP) empanelment to public cloud companies like AWS. Notably, Salesforce leverages AWS’ cloud infrastructure.

Even though the company refrained from revealing much about this new division, Bhattacharya did say that it is still in its nascent stage and could take a couple of years before we see any possible outcomes. She also said that Salesforce has initiated talks with a few parties.

“The other thing, of course, is that we have to learn how to participate and respond to a request for proposals (RFPs) on time. These documents are often extensive and detailed, requiring a learning curve to understand the response process, which also involves collaborating with partners,” Bhattacharya said.

Data Cloud Could be a Seller for Salesforce

One of the biggest hindrances for many organisations is that most of the data is in silos, and this is even true for many government organisations.

Bringing scattered data together into a cohesive source and deriving meaningful insights remains a significant challenge. This is where Salesforce’s Data Cloud could prove to be a valuable tool when it tries to woo the government.

“Data Cloud isn’t something you need ages and ages to build. We have customers that have actually done it in the span of just 90 days and these are not small customers. These are large-scale customers, nothing in the scale of the government, but definitely large scale.

“When you leverage Data Cloud, it’s something where you are not really storing any data; you are merely getting intelligence from the pools of data that you already have and presenting it in real-time. That’s what it does and that’s something the government could possibly use or be interested in using,” she said.

Data Cloud is one of the fastest, if not the fastest growing, organic innovation to come out of Salesforce. In fact, the company is projected to make half a billion dollars in revenue from Data Cloud in just two years.

Moreover, it also does not matter where the data is sitting. “It’s irrelevant whether the data is stored in Snowflake, DataBricks, or elsewhere. Organisations have a wealth of additional data beyond their typical data warehouse, such as client email interactions, contact centre chats, social media profiles, and ongoing online interactions,” she pointed out.

The data, however, do need to be in digital form. In some government offices, some data could be in physical form. “If the data is not digital, then it becomes a challenge. “ If it is not digitised, we can help you digitise it and ensure that the data that’s coming in in the future stays in a digital format.”

Salesforce Public Sector Solutions go beyond Data Cloud, offering out-of-the-box apps tailored for rapid digital service delivery in government agencies.

Leveraging the Salesforce Einstein Platform, these solutions empower government organisations to future-proof their IT investments with citizen developer tools, AI, and automation. This approach enhances service agility and efficiency and supports digital transformation efforts.

Building Trust in AI

Over Time, multiple thinkers in the AI space, as well as many Salesforce spokespersons, have stressed the importance of building trust in AI.

The primary concerns, be it public or private sector enterprises, regarding trust in AI primarily revolve around the data. According to Bhattacharya, catering to the customer’s data requirement and ensuring there are no mishaps is the best way to build trust in AI.

“If we commit to ensuring data privacy and confidentiality, we must uphold that promise without compromise. It is imperative that our offerings reflect this commitment.

“Moreover, the government will have its own laws and regulations as to where it wants the data, how it wants it to be treated, what data can be accessed, and what data should not be accessed. How should it be masked when at rest? How should it be encrypted when in flight, so all of that has to be very assiduously met,” she said.

AI Needs a Global Compact

Bhattacharya also opines that AI today is mostly self-regulated. While some jurisdictions have some regulations, some don’t have any regulations at all. However, it requires a global compact.

“All countries should agree that for the good of mankind, this is akin to the nuclear non-proliferation treaty. We must now agree on principles and actions that benefit humanity collectively,” she said.

This does not have to be a regulatory body but a collective agreement where countries agree that ‘this is what we will do and this is what we will not do with AI.’

Interestingly, this is not the first time an idea like this has been conceived. Sam Altman, the CEO of OpenAI, also shared similar thoughts.

He urged the US to establish a global organisation similar to the International Atomic Energy Agency (IAEA), focusing specifically on AI regulation.

His comments raised some concerns in India as technology’s impact and implications are shaped by human choices, values, and societal contexts. India also has unique socio-economic challenges, such as income inequality, poverty, and access to basic services.

Furthermore, if such a regulatory body is established, does India secure a seat at the table? If so, does it wield a significant voice in decision-making? Or will it be dominated by Western bureaucrats who will regulate the technology based on their country’s risks and concerns?

Bhattacharya believes India already has a strong voice in the global space. Also, today, the world itself is very interconnected. “You should go to Silicon Valley and see how many Indians are working there. Our chief engineer happens to be from Andhra. He’s obviously an expat and lives in the US, but he’s from India,” she pointed out.

Moreover, we can’t take an approach like China, which bans Western technologies. What is required is a collective approach in which all parties work together for the betterment of humanity.

The Chief Data Officer (CDO) role has evolved significantly over the past five to seven years. It has transitioned from a CIO Mini-me, focused on managing data infrastructure, to a business executive tasked with deriving value from the organization’s data. Senior management now recognizes the potential of data to optimize operations, mitigate risks, generate new revenue streams, and create a more compelling, differentiated customer experience. As a result, there is an increased demand for CDOs with economics training, highlighting the importance of an interdisciplinary approach to data management, advanced analytics, and business strategy.

Research shows that CDOs with economic backgrounds are increasingly successful[1]. Economics training equips CDOs with a robust framework for creating value and aligning data initiatives with organizational business and operational goals. These economics-equipped CDOs blend technical data expertise with strategic economic principles, enhancing their ability to interpret complex data and optimize operational decision-making. Their understanding of causal relationships and business implications further aligns data and AI strategies with broader organizational goals.

“The CDO Toolkit”

The original title for my book, “The Economics of Data, Analytics, and Digital Transformation,” was the “CDO Toolkit.” Given the CDO’s rapidly changing role and charter, I envisioned the book as a comprehensive resource to help CDOs navigate their evolving responsibilities. The CDO could use a toolkit to help them apply modern data economic concepts to unleash the value of their data by optimizing or re-engineering key business processes, reducing operational and compliance risks, creating a more compelling and engaging customer experience, and driving data-driven innovation. This toolkit would equip CDOs with economic-based strategies and methodologies to manage and leverage data more effectively as a strategic business asset (Figure 1).

Figure 1: Source: “Is God An Economist?”

Organizations increasingly recognize the importance of leveraging data for a competitive edge. This has led to a growing requirement for Chief Data Officers (CDOs) who have a firm grasp of business and economic principles and the ability to “more effectively leverage data and analytics to power the organization’s business and operational models” (Figure 2).

Figure 2: Data & Analytics Business Model Maturity Index

Let’s consider a situation where you need to interact with the Chief Data Officer (CDO) of a major theme park. Let’s explore how modern economic concepts can help facilitate a more relevant and meaningful discussion with the CDO.

Engaging a CDO: Theme Parks Scenario

Leading theme parks concentrate intensively on guest experience, utilizing sophisticated metrics to measure customer satisfaction and engagement. These metrics can encompass four core indicators: Safety, Courtesy, Show, and Efficiency, each contributing to exceptional service and memorable guest experiences.

To effectively engage the theme park’s CDO, aligning with the business and operational initiatives that the CDO is seeking to optimize is crucial. Start by thoroughly researching how data and analytics can enhance guest experiences in the abovementioned areas. Tools like the “Dean of Big Data” GPT can be invaluable in identifying, exploring, and understanding potential use cases (it’s free if you have ChatGPT4). This preparation will help you shape your conversation around initiatives to improve guest experiences, thus resonating with the CDO’s objectives and demonstrating your commitment to their key concerns. Table 1 lists some use cases that the “Dean of Big Data” GPT identified that might be relevant to the CDO conversation.

Guest Satisfaction Aspects	Potential Use Cases
Safety: Ensure guests feel safe and fully enjoy their experience without worry.	Predictive maintenance for rides and attractions to prevent accidents and unplanned downtime
Real-time monitoring, alerts and recommendations for crowd control and emergency response
Facial recognition for enhanced security and lost child recovery
Health monitoring and safety compliance tracking for staff
Courtesy: Accentuate the importance of friendly, polite, and helpful employees (cast members)	Real-time sentiment analysis of guest feedback to customize staff interactions and improve guest experience
Personalized guest services using AI chatbots and virtual assistants to optimize park flow and guest experience
Employee training programs using AI-driven simulations for better service
Automated language translation services for international guests
Show: Create an immersive experience that contributes to the overall brand magic.	AI-driven content creation for immersive shows and attractions
Personalized Virtual reality (VR) and augmented reality (AR) experiences
Real-time data analytics to adjust shows and in-park promotions based on guest engagement
Efficiency: Eliminate inefficiencies and streamline operations to minimize guest frustration.	Optimizing park operations through AI-driven scheduling and staffing
Queue management systems using real-time data to reduce guest wait times
Optimizing energy, water, and waste management systems for cost savings and sustainability
Predictive analytics for inventory and supply chain management

Table 1: Potential “Improve Guest Experience” Use Cases

While the data and analytic technologies enabling these use cases are essential, the CDO will probably find the economic aspects of the data and analytics conversation even more compelling. Focusing on how to leverage data and analytics to create new sources of value for guests, the park, and individual attractions will better align with the CDO’s strategic goals. Highlighting the economic impact of data initiatives – such as increased revenue, improved operational efficiency, and enhanced guest satisfaction – can make your conversation more relevant, meaningful, and impactful.

Table 2 outlines how various economic data concepts can assist the CDO in unlocking the economic value of their data.

Data Economic Concept	Benefits to Guest Experience Satisfaction Initiative
Data Economic Multiplier Effect	The accumulation of attributable and quantifiable value from using a data set across multiple use cases. For the CDO, this means investing in the most important data sets that can be used to support multiple use cases, such as reducing ride wait times, personalizing guest interactions, promoting in-park activities, and increasing guest advocacy.
Economies of Learning	Improves decision-making and innovation over time, leading to continuously better guest experiences. By learning from customer, service, operational, and market data patterns, the CDO can enable the continuous refinement of attractions, services, and interactions to meet evolving guest expectations and preferences.
Nanoeconomics	Optimizes individual customer experiences by tailoring interactions, increasing guest satisfaction. Personalized recommendations, targeted marketing, and customized experiences based on micro-level data analysis can make each guest feel uniquely valued and catered to, driving higher engagement and loyalty.
Marginal Propensity to Reuse	Maximizes return on data investment by reusing data and analytics across use cases, leading to consistent improvements in guest experiences. Data collected in one area (e.g., dining preferences) can be applied to enhance other areas (e.g., in-park personalized dining recommendations), ensuring a cohesive and enriched guest journey.
Schmarzo Economic Digital Asset Valuation Theorem	Prioritizes data investments based on potential business outcomes, ensuring data-driven strategies enhance guest experiences. The CDO can allocate resources to the most impactful data initiatives, such as predictive maintenance for attractions or enhanced guest service training programs, ultimately boosting the overall quality of the guest experience.

Table 2: Mastering the Economics of Data and Analytics

CDO: Thinking Like an Economist Summary

If you want to communicate effectively with the Chief Data Officer (CDO), it’s essential to understand that their role has evolved significantly. They are no longer just responsible for managing data systems; instead, they now focus on using data to create business, operational, and customer value. This shift means there is a growing demand for CDOs with a background in economics. Research shows that CDOs with expertise in economics are successful because they can combine their technical knowledge of data with strategic economic principles. This unique blend of skills allows them to lead their organization in leveraging data and analytics to achieve better business and operational outcomes.

[1] PWC “The Chief Data Officer: the key to business strategies, innovation and culture that deliver value”, Deloitte Insight “Organizing to drive change” and Splunk “The Chief Data Officer Role: Responsibilities, Skills & Career Path”.

In a recent interview, Kai-Fu Lee, Taiwanese computer scientist, entrepreneur, and author of ‘AI Superpower’ said that China will catch up with the US in AI capabilities. Speaking about his company, 01.ai, which Lee founded, he said that it is the first company in China to achieve GPT-4 performance.

“If there are breakthroughs, the US will continue to lead. If most technologies are known and it’s about execution, then China with its tenacious, hardworking culture and tremendous market opportunity will be able to catch up,” said Lee.

China’s Advantage

Lee went on to explain how 14 months ago, they had nothing, including no GPT. “At that time, we were six or seven years behind, and at this moment we are six to nine months behind. So the catch up has already been happening and going forward, we hope that it can continue,” said Lee.

China;s advantage lies in execution. Lee believes that building a robust model involves more than just invention; it requires others to replicate, like Mistral and similar companies have.hina’s strategy involves extensive data collection, cleansing, and prioritisation for effective training. They utilise advanced infrastructure and distributed computing to maximise GPU performance.

The country excels in application development, emphasising practical implementation above all else.“China’s advantage is doing whatever it takes to catch up,” said Lee.

Interestingly, such conversations have been gaining traction. Recently, investor and entrepreneur Chamath Palihapititya, also posted about how China is zooming ahead in advancements across varied fields. “China is absolutely destroying us in discovering the future,” he said.

Source: X

While Lee believes China is only months behind the US in AI capabilities, a few months ago, former Google CEO Eric Schmidt believes that the US still maintains a significant lead in AI.

“In the case of AI, we are well ahead two or three years, probably, of China, which in my world is an eternity,” said Schmidt.