ChatGPT could be used to produce content for a book by giving suitable prompts. Some of these e-books available on e-commerce platforms attribute the authorship to ChatGPT, whereas some do not. ChatGPT could be either an author or co-author. Some books are exclusively produced by ChatGPT. It is difficult to account for ChatGPT-written books because of the nature of ChatGPT and the failure of the parties to disclose the authorship.

ChatGPT has ruffled feathers of the tech firms. Microsoft and Google are trying to incorporate AI into their existing products. ChatGPT shows rapid adoption. It has spurred frenzied activity at tech companies. Investors pour money into AI-focused startups. There is a possibility of huge lay-offs.

Several Indian companies have offered metaverse offerings. Infosys has offered metaverse foundry that enables customers to create their own metaverse environments. The customers can be given experiences in an existing metaverse. Tech Mahindra has set up TehMVerse to deliver interactive and immersive experiences in metaverse. TCS is working on verticals such as online shopping and workplace experience. Metaverse is in its early stages and it has yet to establish itself as an alternative channel of business. Wipro too has invested in metaverse and Web.3.0.

The adoption of metaverse may vary from industry to industry and depends on viability, digital maturity, trust, infrastructure and creator economy.

Nvidia’s high end GPUs used in AI and gaming are in short supply. The demand for such GPUs soared up in 2023 on account of AI’s spread, the crypto boom and the popularity of Nvidia’s RTX30 series.

There are reasons for this shortage. There is a shortage of computer chips all over the world. There is US-China trade war. The production capacity of Nvidia’s partners is limited.

Some companies obtain Nvidia GPUs using them as collaterals for loans.

Nvidia RTX 4090, 4080, 4070 cost around 1 lac plus rupees in India. 4060 and 3060 models cost around Rs.40 thousand. 1660 model cost about Rs.21000.

Two types of models are available — RTX and GTX. These are graphics cards. There are some similarities and some differences.

GTX stands for Giga Texel Sharder eXtreme designed to handle high performance gaming with fast and smooth graphics. They use rasterization technique. This technique converts a 3D model into 2D pixels on the screen.

RTX stands for Ray Tracing Texel eXtreme, also designed for gaming and use ray tracing. This technique simulates how light behaves in the real world. It produces reflections, refractions, shadows. illumination and ambient occlusion. It has dedicated hardware to perform ray tracing cores. The games become more realistic. These cards are, therefore, expensive.

Nvidia’s GeForce RTX 4090 is the most powerful and expensive RTX card available. The 4080 is slightly less powerful and less expensive. The choice depends on the budget, monitor’s resolution, the desired frame rate and the games you would like to play.

AMD Radeon RX 7900 is AMD’s flagship GPU. Slightly less powerful and less expensive is AMD Radeon RX 7900 XT. Nvidia’s GeForce RTX 4070 Ti is a mid-range RTX card. The Ti version is more powerful.

ChatGPT runs at a high cost. It cost around $7lac per day to run it. Microsoft and other investors in OpenAI cover these costs out of their pockets. If OpenAI does not show profits soon, the investors could turn off the tap. The company could go bankrupt, sooner rather then later.

It is the Microsoft investment that keeps the company going. The losses are mounting. San Altman, the CEO does not own any equity in OpenAI. The company was non-profit to begin with, but later became profit company. The user visits are shrinking — from 1.9 billion to1.7 billion to 1.5 billion. It could be on account of API cannibalisation . They use API to leverage the LLM. The shrinkage can also be attributed to the rise of open source LLM models such as Llama 2 in partnershing with Microsoft.

There is also a shortage of the graphics processing units (GPUs). It affects OpenAIs capacity to improve and develop new models.

There could be competition from Musk promoted rival chat bot TruthGPT which has already purchased 10000 Nvidia GPUs to get going.

It is to be seen how the shift to paid version works financially for OpenAI. It is possible the company can go in for an IPO. It could then be taken over by bigger companies.

It will take some time to have large fabs in India. However, India is witnessing startups who are building up on what they can in the semiconductor eco-system. India’s startups are fabless units. There are companies which build in-house intellectual properties (IPs). Again there are companies that aggregate the IPs into what is called SoCs — systems on chips. SoC companies source IPs from various players, connect them into a combination that address a particular feature or a problem or a particular application.

There are physical design teams that enable the designs to be taken to fabs for manufacturing chips. The chips are designed across geographies like the US, Europe, Israel and India. These are utilised by the foundries located in Taiwan, Korea, Germany, US, Japan, Singapore and China. Indian startups can use outside foundries partners, and after production the chips can be brought back to India.

Fabless startups add a significant value to the total supply chain. The owners of the chips are essentially those who design the chips.

There is a growing automobile and electronics industry in India. It is an incentive for setting up fabless startups. Today, a car has electronic components of almost 50 per cent of its cost, as against 20 per cent a few years back — ranging from power windows to cruise control.

There is also a growing market for fabless startups consisting of IoTs, energy meters, toy controllers, embedded systems and others. It is a high volume but low margin market.

As generative AI has appeared as a disruptor , the computational requirements would increase — both on the cloud and edge computing. The same would require robust designs for future requirements. Moore’s law has reached its limits and further miniaturization to accomodate transistors is not possible. Here the accelerators are required to speed up the power of transistors.

VCs are funding these startups in India. Fabless startups have a long gestation period. The government too is supportive. There is design-linked-incentive (DLI). It is 50% of the eligible expenditure subject to a ceiling of Rs.15 crore per application. There is deployment-linked-incentive of 6-4% of net sales turnover over five years, subject to a ceiling of Rs.30 crore per application.

The challenge is to attract skilled talent. There is a talent dearth. Startups must be ready to face problems of funding. There should be assembling teams with the necessary expertise.

Established semiconductor giants outbid the startups in paying compensation to the manpower.

The other side of supply chain belongs to foundries and ATMPs — Assembly, Testing, Marking and Packaging.

WPP is the world’s biggest advertising group. It is open to acquire companies in India to strengthen its AI capabilities. They are after AI talent. India has both creative and technological talents. WPP’s workforce consists of 10 per cent people from technology background.

In August 2021, WPP acquired Satalia, a tech company offering AI solutions. WPP wants to strengthen its capability in solving supply chain problems. The supply chain of WPP includes creation of content, production and dissemination. Each component of this process has AI applications.

AI has introduced data-driven decisions in the advertising industry. There is an excitement about the application of generative AI to produce content ideas which can be pushed across multiple channels to maximise the desired advertising goals. As we know, generative AI generates text, sound and image content, and will soon extend to video. The process of idea generation and production can be expedited with the use of generative AI. Generative AI can also be used to assess how the audiences perceive the content.

Despite the benefits of generative AI, there are issues of IP, its infringement and data privacy. It is necessary to create walled gardens around the data to make it secure. The models have been trained on the clients data. The same data then becomes available to others.

AI may replace the bulk of manpower in the US advertising companies (Forrester Analysis). It is to be appreciated that manpower will be free from performing mundane and routine tasks, and such manpower can be exploited for creative and strategic work.

The Advertising Standards Council of India (ASCI) released fresh guidelines in August 2023 for celebrities. It has now broadened definition of celebrities to include those who have 5 lac plus followers on social media. These celebrities will have to exercise due diligence before endorsing the brands.

The claims made by the brand and the descriptions in the ad must be valid. Celebrities cannot endorse any product or brand that has a health warning.

Due diligence consists of making sure that description, claims and comparisons in ads the celebrities appear in or endorse are verifiable, and do not mislead or be deceptive.

ASCI has redefined celebrities to include famous and well-known people who get compensation of 40 lac or more per annum for appearing in ads or campaigns in any media.

The rise of social media necessacitated the change of definition of a celebrity.

Celebrities do influence the spending habits of consumers who trust them. It is necessary to protect consumers.

The guidelines prohibit celebrities to endorse anything covered by the Drugs and Magic Remedies (Objectionable Advertisements) Act, 1954 and the Drugs and Cosmetics Act, 1940 and Rules, 1945.

Testimonials, endorsements, or representations opinions or preferences of celebrities must align with their current thinking and be based on information about or experience with the product being advertised.

Many clients of IT companies expect IT companies to go beyond technology to solve their problems, and that open up the field of consulting capabilities and that differentiates these companies. To begin with consulting was offered by IT companies as an add-on to drive technology business, these days it is offered as an integrated bundle.

Management consulting is the foundation since one who has the decision making power gets the power to execute the agenda. The consulting firms and IT firms were distinct entities, but the lines between them are blurring. The Big Four — EY, Deloitte, PwC and KPMG — represent this trend. These traditional consulting firms try to scale up their tech capabilities. On the other hand, the IT firms acquire specialised firms to build their consulting capabilities, e.g. Infosys acquires Lodestone in 2012 and Noah in 2015. Similarly Wipro acquired Capco for providing financial services. HCL aquired StrongBridge and Symplicity DWS. Cognizant is in the process of acquiring AustinCSL.

IT consulting market is growing fast. IT service providers should focus on professional consulting services at reasonable prices, rather than offering consultancy either free or at throw-away prices. IT service providers should build a special cadre of manpower talent who can work on cross-disciplinary projects involving business strategy, technology and industry context or domain expertise. IT firms must also consider the client context since they are familiar with their operations, technology and data landscape. This can be used to showcase more practical roadmaps for the clients.

E-pharmacies representatives and health ministry officials are holding a meeting to address the concerns about the regulations which would govern them.

E-pharmacies accept e-prescriptions but here the government is cautious as it has wider implications.

There are data privacy issues as these online pharmacies have data of the consumption of medicines by the patients as well as the diagnostic tests they have been subjected to. There is a possibility of misuse of the data.

Online pharmacies practice predatory pricing to gain customers, and this adversely affects the offline retail chemists — there are 8 lac chemist shops in India.

The health ministry should formulate guidelines by holding discussions with all the stakeholders, especially the Chemists and Druggist Associations. The retail chemists are outpriced.

Offline pharmacies cannot deliver medicines to residences, whereas online pharmacies can do so.

Online pharmacies should not be allowed to sell the habit forming drugs, anti-TB medications and higher antibiotics.

Prompt engineers are employed to train the large language models (LLM) of generative AI to deliver more accurate and relevant responses to questions people are likely to ask.

They can be utilised to generate code suggestions and identify coding errors and bugs. They can used for improving code performance. Prompt engineers design and develop prompts for ChatGPT. They use methods such as prefix tuning or prompt tuning. It works for pretrained LLM models. Prompt engineers use Chain-of-Thought (CoT) prompting to improve the reasoning ability of LLM Models, especially when the problem is complex and multi-step. Here the prompt is broken into a series of intermediate steps.

Prompt engineers must have creativity and must show language precision. He need not be a hardcore computer engineer. He can be drawn from humantities stream. Prompt engineer must be able to imitate human thought. Of course, an understanding of programming and AI and ML and NLP often helps. Tech skills must be combined with the understanding of user needs. A prompt engineer must have better linguistic skills.

There are short-term courses available for learning prompt engineering. Practice makes a prompt engineer perfect.

A prompt engineer cannot afford to be too literal or too fuzzy. He should combine reality and imagination. Essentially, it is an exercise of precision thinking. You should be at generative AI terminal for hours and figure out what prompt generates what response.