ChatGPT Enriches the Rich

We are celebrating the second birth anniversary of ChatGPT which was introduced in November 2022. It quickly became the fastest growing app in history with about 200 million as active users today. In its early days, it was hailed as a big leap forward — it had varied capabilities to generate text, images, code. It had the capacity to influence film making and the education system. It has fulfilled some of these expectations — say it is a good educational tool, but its influence on other areas of life and business is still an open question.

In these two years, a handful of tech firms have benefited greatly — the big six tech firms grew in aggregate more than $8 trillion since ChatGPT’s birth. Enterprise level adoption, though slow, has been growing continuously. The app must facilitate transformation of the business in general, rather than enriching the big tech.

Consultancy firms promoted generative AI to their clients. The AI business of IT service firms too is going up. We expect to see the benefits of AI percolate down to small businesses and startups. It is expensive to be in the business of building ‘foundational models’. These have to compete with OpenAI, Google and Facebook models. The smaller businesses focus on building services that act as wrapper around existing models. Even here big tech can introduce a model that snuffs out models of the smaller firms, e.g. speech recognition models suffered when OpenAI introduced Whisper in 2022.

Maybe, the stranglehold of big tech loosens in future, as for specific needs there could be a shift for smaller models. These smaller models are easier and economical.

Of late, AI has plateaued. It is, in fact, an opportunity for businesses to introduce generative AI into their processes and assess it for a return on investment.

ChatGPT has left a legacy two years after its appearance of enriching the big tech. It should lead to the next phase of AI — it should become democratic. It should benefit smaller business. It should have smaller models to remove entry barriers. Till now, the revolutionary technology has just extended the wealth and influence of big tech.

Satcom and Mobile Services

Satellite communication services and terrestrial networks (mobile services) are pitted against one another. There is an issue of spectrum allocation. The mobile services companies expect the spectrum to be allocated by auction, while satcom companies expect it to be allocated administratively.

Are both these companies on par? Terrestrial networks can serve both hand-held and rooftop devices. Satellite services face significant hurdles for mobile devices. They cannot match the terrestrial companies. Both of them are on par only when fixed wireless access (FWA) solutions or broadband services.

Satellite companies cannot replicate the services of terrestrial networks for cellphones which cannot efficiently receive satellite signals at high frequencies. Thus there is no direct competition between with traditional mobile networks.

The area of overlap of services is fixed wirless access (FWA). Here too satcom companies face significant challenges as compared to their terrestrial counterparts. They have to track fast-moving satellites and hence their terminals are complex. Cellular base trans receiver stations (BTS) are stationary.

In addition, satellites operate 600-1200 kms away. Base stations or BTS are just a few or even a few hundred kms. away. Thus, satcoms offer lower data speeds than their terrestrial counterparts. Satellite terminals are thus more expensive. These are used where terrestrial networks are not available. India has 29 million BTSs, and 8 lacs towers. Starlink, by contrast, has only 7000 satellites globally and could expand to 40000 satellites.

Mobile operators work on the model of high volumes and low average revenue per user (ARPU). It is the opposite for satellite players. Starlink’s ARPU is $100 as compared to that of mobile operators which is $ 10-15 for FWA services in India.

Both these companies are complementary and do not compete with each other. The issue of level playing field is not relevant.

Regulatory provisions cannot stifle innovation. There should not be lobbying on the basis of level playing field.

AI-driven Drug Discoveries

Earlier, pharma companies largely relied on time-consuming and expensive methods of drug discovery. The advent of AI changed this. AI now supports research to achieve breakthroughs. It helps in manufacturing and testing novel therapies. Merck has developed an in-house clone of ChatGPT called myGPT with 30000 users. It has been used for automation and efficiency in day-to-day activities, e.g. content writing, translation, content revision, coding assistance and so on.

Merck also uses SYNTHIA, an AI-powered drug discovery platform. It also uses ADDISON, first AI-powered software-as-a-service platform that facilitates screening of billions of molecules. It accelerates the process of drug discovery and reduces its cost.

A Fateful Knock

These are the days of large language models (LLMs). This is the story of two important men in the field of AI and how they started working together.

Geoffrey Hinton, now a Nobel laureate, was teaching in 2007 at University of Toronto. He describes how he first met IIaya Sutskever, formerly working in OpenAI as Chief Scientist. Ilaya had completed his master’s in computer science. In Hinton’s office, on a Sunday, Ilaya knocked at the door. Hinton answered the knock. Ilaya requested for working in Hinton’s lab. This is a common practice amongst the students — they approach the faculty for lab work. Hinton passed on a paper on backpropagation to Ilaya and set another meeting with him a week later. In the next meeting, Ilaya came back and said he did not understand the paper. Hinton was disappointed. In fact, Ilaya had understood it thoroughly but had an issue with not giving the gradient to a sensible function optimizer. He wanted to improve on the paper. Hinton realized Ilaya was special. Ilaya did research under Hinton to get his PhD in 2013.

As we have observed in a previous blog, Alex, Hinton and Sutskever developed AlexNet (a CNN network). AlexNet became a precursor to modern AI models.

In 2013, Hinton, Ilaya began working at Google Brain (after a company they had started was acquired by Google). Hinton continued with Google, while Ilaya joined Greg Brockman and Sam Altman to launch OpenAI in 2015.

We know OpenAI initiated the AI revolution by launching ChatGPT in late 2022. Ilaya was overseeing the research at OpenAI as the Chief Scientist.

Hinton has been awarded Nobel in physics in 2024 for inventions in ML. Ilaya left OpenAI and is now working on SSI.

It was a knock at the door of Hinton on a placid Sunday that changed the world for ever.

Vanilla RAG and Agentic RAG

We have already discussed RAG — retrieval augmented generation. It accesses external knowledge sources to respond to user queries. However, we do need a more nuanced, complex and adaptive RAG. The traditional vanilla RAG has its limitations. Agentic RAG has now emerged. It is an advanced architecture — it combines the foundational principles of RAG with autonomy and flexibility of AI agents.

Vanilla RAG is a linear pipeline. The user queries are processed through retrieval. It struggles with flexibility. There is no iterative refinement. Agentic RAG addresses these shortcomings. Agents act autonomously. They coordinate complex task — planning, reasoning with multiple steps and tool utilization. The retrieval system becomes dynamic.

Agents are incorporated at various stages of RAG pipeline. Agents decide whether external knowledge is required. They select apt retrieval tools — vector search, web search, APIs. It formulates queries customized for the task. Agents after retrieving data validate the data. Agents can resolve queries with accuracy and speed from internal sources, documentation and community fora. It is similar to the fine-tuning of an LLM.

The architecture is not confined to a single agent. It can use multiple agents.

RAG has its limitations. An agentic RAG may not respond since information is not available in database. It is a waste of compute. In addition, it does not scale with more compute.

Google has moved to RIG retrieval interleaved generation.

Google’s Ad Tech Monopoly Case

Google faces the charge of monopoly in search engine. It is also fighting another alleged monopolistic conduct case over technology that puts online advertising for the audience. The closing arguments will be heard in November 2024. The ruling is expected by the end of the year. If the court finds that Google engages in monopolistic conduct, there will be further hearings to explore the remedies.

The Justice department, along with some states believe Google should be forced to sell off its ad tech business. The justice department contends Google has built and maintained a monopoly in ‘open web display’ advertising. Essentially these are rectangular ads that appear on the top and right-hand side of the page when one browses the websites.

Google dominates all facets of the market — DoubleClick technology is used by news sites and other online publishers. Google Ads maintain a cache of advertisers looking to place their ads on right webpage for the targeted audience.

Google also operates AdExchange which conducts instant auctions to match the advertisers to the publishers.

Thus, it is a trifecta of monopolies which google tries to preserve, rather than serving its customers — both publishers and advertisers and winning the bids on merits.

Google brokers transactions between advertisers and publishers by extracting heavy fees, and as a result the content providers and new sites do not generate enough revenues.

In defense, Google says the government focuses only on online advertising which is a narrow niche. There is online advertising on social media, streaming TV services and app-based advertising. According to Google, it controls only 25 per cent of the market, and this share too dwindles in the face of competition.

Google also says it has invested billions in technology that facilitates the efficient match of advertisers and target audience. It says it should not be forced to share its technology and success with competitors.

The ad monopoly case is runs in Virginia and is separate from search engine monopoly case in Columbia.

During the proceedings, Google lawyer raised the issue of AmEx case related to anti-trust activity, but the judge disagreed with the relevance of that 2018 case to the present case.

In fact, the case pertains to monopolizing three separate markets for ad technology: sell-side tools used by websites ( called and servers ), ad exchanges and buy-side tools used by advertisers (called and networks). Google claims that it is wrong to put its tools in three buckets and it’s business be understood as a single market where websites publishers and advertisers transact.

Transformative AI Agents

AI sooner will undertake hands-on roles with the help of AI agents which are automated tools to do the specific tasks to bring an element of efficiency across industries. There is constant expansion of the usage of generative AI from year to year. The advent of AI agents in 2025 will escalate the productivity to superhuman levels. AI so far was a novelty. It will then become a necessity.

LLMs no doubt have occupied the center stage, and rightly so. Still, these have limitations. There are only incremental improvements in their functioning. Some organizations leverage a stack of technologies to maximize the benefits. Some companies use smaller models. Such combinations enhance productivity.

AI agents can be used for the least glamorous tasks. They are best suited for mundane tasks. AI can facilitate repetitive tasks leaving high judgement tasks in human hands.

One such example is the review of the legal contracts. These agents go beyond handling routine tasks. They facilitate collaboration by being active partners. They can scan unstructured documents. They can manage customer inquiries. Specialized AI agents can tackle specific tasks.

They provide conversational interfaces. They can do document analysis. They can review code and do security monitoring. They can be used in healthcare and financial services. AI can be democratized. AI teams will manage their own data. Technology will no longer serve a select few, but will be available to everyone, everywhere. Industries will be transformed — productivity will reach new heights and accessibility will be much wider.

New Era of Agentic AI

We use software, and then we are called users. However, AI agents are the software and the software is the user. AI agents bring about a fundamental shift in the role AI plays in organisations. OpenAI’s agent called Operator is expected to be launched in January, 2025. It will serve as a personal assistant which will take multi-step action on its own. Operator will be able write code, do travel bookings and manage daily schedules. It will accomplish this by using apps already functioning on our devices and by using cloud services.

Anthropic has released Computer Use that enables Claude 3.5 Sonnet to perform complex tasks on its own.

This heralds the coming age of agentic AI.

Agentic AI understands the user’s objective and comes up with component parts to achieve the goal for the user. It plans and executes that plan. It usually uses other software and cloud services.

Three Abilities of AI Agents

1. AI agents possess the power of reasoning since at its core is an LLM that can plan and reason. It is the LLM that breaks down the problem and creates plans to solve it. It gives reasoning for each step of the process.

2. AI agents have the ability to interact with external programmes — web searches, database queries, calculators, code execution and with other AI models. The LLM decides how and when to use these tools.

3. Agents can access a memory of past events — internal logs of the agent’s thought process and history of conversations with users.

The interactions are thus personalized and context specific.

Reasoning and Acting (ReACT) are the key differences between AI chatbots and AI agents. The Acting part is really different.

AI’s future is linked to AI agents. AI agents will take over many tasks in business which are currently automated.

Agents will create new opportunities for training agentic systems.

AI agents and AI smart glasses will support each other.

AI agents have the potential to improve and degrade human capability.

AI for Enterprises

Enterprise level AI could either use discriminative AI or generative AI. Discriminative AI focuses on data classification. It undertakes customer sentiment analysis, fraud detection or automation. Here classification accuracy is crucial. Discriminative AI enhances operational efficiency and decision making.

Generative AI creates new content. It also does predictive simulations. It can generate new text, image or music. It facilitates innovation and can engage customers.

While using AI, a business defines its goals. It assesses data availability. It assesses need for accuracy. It considers resource constraints. It has to be cost effective. It should have flexibility.

US Agencies Seek Google Breakup

In November 2024, the US Justice Department and a group or states asked a federal court to force Google to sell Chrome, its popular web browser.

The request follows a landmark ruling in August 2024 by the US district court for District of Columbia which found Google had illegally maintained a monopoly in online search. The court asked the Justice Department and the states that brought anti-trust case to submit solutions to correct the search monopoly.

Apart from selling Chrome. the government asked the court to give Google a choice: either sell Android, its operating system for smartphones or bar Google from making its services mandatory on phones that use Android. If this does not work, in future, the government could force the company to sell Android.

The government asked the court to stop Google from entering into paid agreements with Apple and others to be the automatically selected search engine on smartphones and in browsers.

The government asked the court to allow the rival search engines to access Google’s data for a decade.

Since the breakup of Microsoft in 2000, these proposals are the most significant remedies requested in a tech anti-trust case. If these proposals are adopted, it will set the tone for a string of anti-trust cases against other Big Tech companies such as Apple, Amazon and Facebook.

Google’s worst nightmare would be an order to sell Chrome and Android. Chrome is the most popular browser introduced in 2008, having an estimated market share of 67 per cent of the global browser market.

Android operating system that is open source enjoys 71% of market share. It’s being open source means that mobile phone companies use it free, but then it comes bundled with Google apps already installed. The government contends that there is no level playing field.

Legal experts feel that the solution that Google should be forced to sell Chrome may not be accepted by the court, since the same break up remedy in Microsoft’s case in 2000 was overturned by an appeals court.

Google has yet to file its suggestions for fixing the search monopoly by December 2024.

Both sides can modify their suggestions before the arguments open before the court. The ruling is expected by the end of summer.