Blog

One more mode of transport could be added to urban transport soon — electric vertical take-off and landing vehicle (eVTOL). It is an aircraft which can lift off the ground like a helicopter and can land also easily. It can fly at speeds of 322 kms per hour with a range of 161 kms. The aircraft does not produce excessive noise.

Private companies in the US have raised billions of dollars in funding and are working hard to convert their dream of flying taxis into reality. Some of these companies are being backed by Big Tech such as Google and Boeing. Some have tied up with existing airlines.

They are working closely with regulatory authorities. They have created a new aircraft category called ‘powered lift’. Such a lift has been introduced when helicopters were introduced for civilian use in the 1940s. There are still some regulatory hurdles to be cleared.

Dubai is most likely to be the place where eVTOLs will take commercial flight, perhaps by the end of 2025.

It is going to be like a crawl, walk and run situation. They are still crawling at present.

China is also pursuing the idea of flying cars. This could motivate the US administration to make such vehicles a priority.

Such taxis can serve the passengers at airports of New York and Los Angelas. Electric taxis can fly four passengers from say New York airport, to say Manhattan in about 10 minutes or less.

In the beginning, they could charge a higher price. They would be costlier than the cab rides. The difference between air taxi and cab ride could narrow over time, as they would be able to transport a higher volume of passengers than ground vehicles stuck in city traffic.

Surely, eVTOLs are going to transform the way we move. It is better to see the world from the air than being stuck in the traffic on the roads.

AGI is not just one new tool. It is a new phase of civilization. OpenAI is introducing progressively powerful reasoning models. These ferret out knowledge and information from a vast base of accumulated reservoir of knowledge. They have an additional ability to think further and solve complex problems. The implications of this leap have not yet fully sunk in by penetrating human consciousness. However, this has profound implications.

GPT o1 model scored 83 per cent on IMO. GPT o3 model achieved an unprecedented score of 87.5 per cent on the ARC-AGI benchmark. It measures a model’s ability to solve completely novel problems without depending on pre-trained knowledge. ARC-AGI tests conceptual reasoning and adaptive intelligence. These areas are traditionally the forte of human beings.

Till now, AI systems demonstrated narrow intelligence — writing copy, diagnosing diseases from symptoms. optimizing logistics. These have narrow limits. General intelligence is fundamentally different. It has the ability to adapt, reason and solve problems across domains.

LLMs and muti-modal models have already shown proto-AGI traits — generalization across tasks, multi-modal reasoning and adaptability. Let us call these glimmers of AGI. These capabilities are improving iteratively — better architecture, larger datasets and improved training methods.

OpenAI has redefined AGI. It is first of all an autonomous system. It secondly outperforms humans at most economically valuable work. That creates a shifting endpoint. Microsoft and OpenAI has created a linkage between AGI and profits — an AI system that generates $ 100 billion in profits.

AGI challenges a basic human trait — intelligence. It is no longer exclusive to humans. There is an issue of integrating AGI to our lives.

AGI puts us on the road to ASI — artificial superintelligence. AGI systems will become self-learning and will surpass collective human intelligence.

The goal is ambitious. There should be machines that not only think but evolve.

The emergence of AGI will not be sudden. It will unfold gradually. There will be journey from AGI to ASI.

Microsoft proposes to invest $ 3 billion in India in cloud and AI in frastructure and skilling over the next two years. This also includes the establishment of new data centers.

The company will also help train 10 million people over the next five years with AI skills.

India is emerging as a leader in AI innovation. Microsoft wants to make India AI-first.

Microsoft has 3 data center regions and the fourth will go live in 2026. The investment aims to develop a scalable AI computing ecosystem to meet the growing demands of India’s rapidly expanding AI startups and research community.

Indian professionals are early adopters of a new technology, and many professionals are adding AI skills to their profiles.

Nvidia has unveiled its first desktop computer — called Project DIGITS. It is a computer designed for programmers, rather than regular consumers. It costs $3000 and runs Nvidia operating system based on Linux.

Nvidia’s data center AI chips will now power PCs and laptops. Nvidia has also introduced Cosmos foundational models that generate photo-realistic video. It can be used to train robots and self-driving cars at a much lower cost than using conventional data.

Cosmos will be made available on open license, similar to Meta’s Llama3 language models. Nvidia’s Cosmos will do for the world of robotics and industrial AI what Llama3 has done for enterprise AI.

Nvidia also unveils its gaming chips that use its Blackwell AI technology. Nvidia calls them RTX 50 series. It will give video games movie-like graphics.

The new chips will help the game developers to generate more accurate human faces.

Sam Altman writes in a blog post that his team knows ‘how to build AGI’ and is hopeful that by the end of 2025 the world is likely to witness ‘the first AGI agents.’ It will have an impact on every facet of human life — from economy to education.

The effect will be felt in every industry and every home.

AGI is highly autonomous system that outperforms humans at most economically valuable work. In his blog, Altman calls this the ‘most impactful technology in human history.’ It is both promising and disruptive. Though AGI integration will take time, AGI will ‘materially change the output of companies.’

AGI will accelerate scientific discovery and innovation. It will redefine what it means to ‘learn’ and ‘teach’. Students will have to foster critical thinking and resilience. Education must prepare students for ethical challenges posed by AGI. AI is evolving very fast. It shows accelerating pace of innovation. There should be now a culture of lifelong learning. There should be commitment to ensure the benefits of AGI are broadly shared.

Altman looks beyond AGI to the promise of superintelligence.

There is intersection of technology and education. AGI is not just a technological advance but a societal transformation. It involves engagement of educators, policy makers and citizens.

Educators should welcome AI as a partner in learning.

Quantum computing attracts a lot of attention. However, computing is just one element of quantum technologies. In fact, we are entering a new quantum era where we leverage quantum states of matter down to the level of individual particles. Just entangle two photons of light and we create a communication channel that cannot be subjected to eavesdropping. Or the sensitivity of quantum particles, could be availed of to detect the phenomena that have never been sensed before.

The previous 100 years development could be called quantum 1.0. This led us to transistors and laser. Further, we got microprocessor and internet. We are now witnessing Quantum 2.0 redefining the way we communicate, compute and sense the world.

There is quantum networking. Quantum networks carry quantum information. This will pave the way for quantum internet. It could create secure communications systems. Our definition of communication will change. We can expand the amount of information we can pack into a single photon of light. These optical networks can transmit large volumes of data at low power. The networks will extend to our solar system.

A powerful quantum computer will be able to crack the widely used security measures. A cryptographically relevant quantum computer (CRQC) will take more than a decade to arrive. We have to think how information is to be protected from quantum attacks.

Quantum sensing will allow us to detect natural phenomena that we have never been able to measure so far e.g. electrical impulses in CNS or fluctuations of gravity on the earth.

Applied quantum mechanics will accelerate the pace of innovation since we will be able to harness matter at the quantum level. The earliest quantum innovation was a transistor developed at Nokia Bell Labs in 1947. It led to microprocessor, the Internet and the digital world.

Quantum computing provides tremendous data processing power. Quantum computing is not going to replace classical computers, but the amazing computing power could be leveraged in the field of medicine, chemistry, material science and other fields.

Classical computers store and process information in the form of binary numbers or bits. A single bit is represented as 0 or 1. The basic unit of a quantum chip is a qubit — these are subatomic particles (electrons or photons) controlled and manipulated by specially designed electric and magnetic fields.

Qubits can be made in different ways — use of superconductors, semiconductors, photonics and other approaches.

It is not the quantity of qubits that a chip has is important. It is the quality of qubits. A quantum chip with thousands of low-quality qubits will not be able to perform any useful function. Qubits are sensitive to unwanted disturbances from many sources. This reduces the reliability of a qubit (known as fidelity.). A computer chip must have high-fidelity qubits.

It is not necessary to build such perfect chips. There are techniques to have low-fidelity qubits encoding abstract logical qubits which are immune to errors and therefore, of high-fidelity. A useful quantum processor must be based on many logical qubits.

Quantum chips consisting of over 100 qubits are available. Till now, developers have only made single logical qubits. It will take some time to put several logical qubits together on a quantum chip.

The aim is to reach quantum supremacy — a quantum processor solves a problem that would take a classical computer a very long time to solve.

Quantum hardware (processors) are progressing. At the same time, researchers are developing and testing various quantum algorithms.

A full-scale quantum computer is still a daunting task where so many things will have to fall in place– number of qubits on a chip, improving fidelity of qubits, improved error correction, quantum software, quantum algorithms and various other sub-fields of quantum computing.

OpenAI is re-entering the field of humanoid robots. In past, they had a dedicated robotics team, which was shut down three years ago. The reason advanced was lack of sufficient training data.

OpenAI has chosen humanoid robotics — a humanoid robot has features similar to humans in skeletal form — arms, legs, head, etc. It is trained to do human-like tasks.

Even when the robotics division was shut up, the company remained invested in the field. They had invested in figure AI, a US-based startup.

They want to back the humanoid models by powering them with multimodal LLMs. The company had also invested in 1X Technologies, a company that produced humanoid robots. In 2024, they also invested in robotics centered startup called Physical Intelligence.

Humanoid robots have been used to test AI in workspaces, warehouses and other physical spaces. As OpenAI has introduced o3 models, they may be thinking of humanoids with advanced reasoning capabilities to perform complex tasks.

These robots would be useful in manufacturing, healthcare, hospitality and caregiving fields.

X

It is difficult to predict about the success of a Hindi movie in 2024. The box office collections have dropped by 30-40 per cent. There are no footfalls at the theatres.

It was with the advent of freedom in 1947 that the Golden Age of Indian cinema began. It lasted for 20 years. In the 1960s, politics became cynical, and there was loss of idealism. During this period, the cinema did not reflect the social concerns and became an escapist fantasy. The volume of film production shot up by 1970s. The official financial channels were reluctant to fund the films. It is during this period that a nexus between the underworld and film industry emerged. The funds came from the profits made out of smuggling. In 1980s and 1990s, small time criminals became money launderers and drug and arms traffickers. These anti-social elements went beyond funding. They entered the film industry as producers. They bought overseas rights for film and music distribution. Along with these anti-social elements, the film industry too experienced the harmful effects. There was assassination of the head of a music company in 1997.

In this millennium, the corporatization that began in the previous century continued. The films have been recognized as industry and nationalized and private financial institutions are now ready to finance them. The shadow of the dark past still continues to trouble film industry, as the biggest superstar receives a death threat from a gangster that is in prison.

In the internet age, there are streaming OTT channels such as NetFlix and Amazon Prime. Movies are released here too.

Though it is an open market now, the audiences are still not drawn in large members to theatres.

On Jan 6, 2025, Google DeepMind announced the formation of a new team to work on massive generative AI which would ‘simulate the world.’ In fact, it is the next step in the development of AI.

World models, as Google calls it, are computational frameworks that enable the AI system to understand and simulate the real or virtual world. AI systems would learn to navigate the environment. These models will be useful in gaming, autonomous systems and robotics. Autonomous cars make use of world models to simulate traffic and road conditions. AI robots which are generalist can be trained to operate in different environments.

Such scaling of AI models will evolve the technology further. Scaling related to video and multi-modal data brings us closer to AGI. World models will power different domains — visual reasoning and simulation, planning for embodied agents, real-time interactive environment.

Former OpenAI employee Tim Brooks who left in October 2024 has come on the rolls of DeepMind and would lead the world models team. It will work further on Gemini, Veo and Genie. DeepMind has already developed world models of Genie1 and Genie 2.