In less than a year since the advent of generative AI, LLMs have taken the enterprises by storm. It has immense possibilities. IT firms have infused AI into most of its operations. In code development, engineers use GPT, say at the stage code analysis, code migration, documentation and code’s automatic testing. AI has been incorporated into management systems. It has been integrated to project delivery. AI improves project delivery by 20-35 per cent.

In many projects, Co-pilot is being used — AI that is used in code development.

There are industry specific LLMs now. There is work-in-progress for healthcare LLM and LLM for production operations. Companies would like to have products which are AI-assisted. Education is another sector which could have its own LLM.

Microsoft has developed Nuance with ambient documentation for doctors. A doctor can get his speech transcribed by speaking on mic. It can create a medical record using Gen AI by summarizing what the doctor has to say. The transcribed data goes into record . It saves both time and costs. It improves a doctors productivity — he can attend to more patients in a day.

It is not necessary always to create a company specific LLM. A company can get a pre-trained model, say LLM of OpenAI. They can feed their own data into this LLM to customize it to their own needs.

AI has been used in radiology. It can detect anomaly in X-ray images. Its accuracy is high. LLMs can explain what a medical scan means in plain language, avoiding medical jargon.

Doctors can seek second medical opinion on diseases prevalent in the area they work in. AI can create synthetic images of impaired Covid-infected lungs. Doctors can distinguish between pneumonia infected and lower respiratory tract infected lungs and Covid infected lungs.

The concept of AI will be extensively used in different business models in near future.

Manpower working in this field must have tech knowledge plus domain knowledge. The idea is to apply tech to a problem on hand.

Startups and amateurs have a chance to use their talents to compete with the Goliaths in this area. A small firm has the capability to disrupt a big company.

Gen AI will make our lives better. There is need for innovation. If you do not, you may be left behind. Gen AI has become a pervasive force.

Indian IT firms rely upon new tech spends by global retailers. They are likely to get more deals from the US and Europe-based retail players. Walmart, Home Depot, Carrefour and Target have scaled up spending on digital technologies and generative AI. Generative AI is used to enhance shopping, making it easier and more convenient for customers.

Larger retailers are exploring the use of AI and LLMs for predictive analytics. Generative AI solutions such as Topaz by Infosys are being offered to the clients.

Indian IT firms get about 12-18% of their revenue from retail verticals.

A robot could be dismantled for excesses it committed. However, there could be defense on the basis of its rights. Can it have the same rights as a human being? These days the legal right of generative AI are on the anvil.

There are successive improvements of AI system. Have you ever wondered what happens to the previous versions? Of late, a food delivery robot was assaulted by two students at University of Tennessee. The students were charged. Was the charge for cost involved — replacement cost of this robot was $5500? Or was it for acting against a robot?

AI has not yet been treated as something with a ‘new’ life. AI systems are killed by simply unplugging them or disconnecting them from internet. In future, the same could amount to a murder. The time is ripe for deciding the rights of an AI system.

AI systems, all said and done, lack emotion. They at times cause harm to humans. Some philosophers talk about their utilitarian value, and hence their having a moral standing. AI can also think of human beings as smart and exploitative. In that case, a premptive strike by them is justified. Here we are conceiving a conflict between humans and AIs.

There are now endeavours to develop AI capable of thought and conciousness. There are going to be relationships between AI and humans, and some of these could go up to marrying them. How can these two marry? And under what law? What is going to be definition of this marriage? Maybe, such a marriage does not remain a solitary case. In that case, there should be rules to legitimize such marriages. Later, there should be rules for separation too.

In future, there can be an ‘AI justice’, sitting in judgement over such issues.

Computers can be taught to translate brain signals into text. Two new studies in Nature have documented this. It is an exciting development. One team was from Stanford and the other from California. They used different types of implants to collect electrical signals from the brains of volunteers. They used different algorithms to interpret these signals. Still, these led to similar results. Brain signals can be converted to text at the rate of 60-70 words per minute. Though it does not approach a typical person’s cadence, it is an improvement over previous interfaces. Surprisingly, computer in both the experiments decoded the speech with reasonable accuracy.

Researchers created a personalised avatar for a paralyzed volunteer. The same voice taken from a previous clip of voice was used to make the avatar speak. The avatar was integrated to the facial expressions of the volunteer. This experiment proved that such avatars are a boon for people who are unable to speak. BCI or brain computer interface is a very useful technology. This experiment is a turning point in the development of BCI technology.

These technologies must be translated into useful products. There is a niche market for these products. The products could be expensive. They require training before being used. The early marketers will face a difficult task.

As we know, a new computing era has begun. Corporates around the world are shifting to more powerful computing that can handle generative AI. Nvidia’s chips are at the forefront of an industry-wide AI race. Its shares rose 4.8 per cent to $493.70 at NY. It has become a key beneficiary of AI computing boom. Data centres are stocking up the company’s processors. These chips are adept at handling heavy workloads required by AI. As we know there is global chip slump and Nvidia has witnessed accelerated sales growth.

Nvidia has become the first-ever semiconductor company to get a $1 trillion market valuation. It has become the main supplier of infrastructure required to support the graving use of AI systems.

Generative AI is being criticized for using data to train its models because the data does not belong to the AI companies. Many websites block the AI companies from scraping their content. A short code tells the AI company to go away. Many websites shut the door this way. Many websites add the code to their ‘robot.txt’ file to regulate bots or spiders that scour the internet.

How can an AI company take the content that does not belong to them? It will take much more than just a line of code to stop this happening.

There are other online resources to exercise the block on crawlers. StackOverflow blocks the crawler but it does not block GitHub, owned by Microsoft.

OpenAI’s offer giving sites the ability to prevent siphoning off the content by using a code is to be juxtaposed to a long data gathering exercise it has done for quite sometime. It is trying to be helpful when the horses have bolted. The burglary is already over.

At the same time, it must be appreciated that OpenAI expects other companies to follow suit.

Web crawlers are not the only method used by AI companies to collect the data. These companies use bulk datasets provided by third parties, say Books3 has a dataset of 2 lac books. Authors should sue over the use of books this way.

There is no way we can do something about data already scooped. There is no guarantee that despite the code to block, data will not be collected by alternative methods. OpenAI has acknowledged in future consent will be vital for undertaking future scraping.

There are so many bots out there. They do not provide any kind of safeguard and the sites have no option to opt out.

Bard of Google is interested in a discussion to develop a mechanism for administering consent on AI.

What has already been done cannot be undone. The data is in the digital blender.

AI has the capacity to auto-generate ads on platforms such as Google Ads. AI can leverage landing pages, successful search queries, and approved headlines to create entirely new creative content.

Google’s innovative feature which automatically generates additional assets such as headlines, descriptions and images, enhance the search ads. AI-powered Performance Max when allowed on websites learns about relevant things of the brand and facilitates a better campaign.

Meta has introduced AI Sandbox (a suite of generative AI tools) for text variation, background creation and image cropping. It improves the overall layout and text of the ad. MetaAdvantage is a portfolio of automation products. It uses AI and ML to optimise campaign outputs. Meta’s distinctive technology of text-to-video produces videos on prompts. It is promising for advertisers. Google may roll out a similar technology soon.

AI tools optimise and expedite creative process. They are time saving. The top tools used by agencies are Jasper which creates copy, Hemingway App which checks grammar and enriches text, Albert.ai which gives variations, INK which provides co-writing abilities and is SEO assistant, Fullstory which gives data to build better digital experiences and Headlime which is a copywriting tool.

Such tools help advertisers for pitching. Instead of explaining everything verbally, and in text form, they now use AI tools to craft mock-ups. It is so effortless.

PDF stands for Portable Document Format. It is a file format developed by Adobe in 1993 to present documents (both text formatting and images) independent of the app software, hardware and OS.

PDF files can be viewed, navigated, printed and edited with Adobe Acrobat Reader software.

PDFs are generally used to distribute electronic documents. They can also be used to create interactive documents (online forms and presentations).

As PDFs are platform independent, they can be viewed on any device. They are scalable, are password protected and can be embedded with links, buttons and such other interactive elements.

John Warnock, founder of Adobe systems, developed PDF. He died in August 2023 at the age of 82 at his home in Los Altos, California. Prior to PDF that was released in 1990, the company had developed postscript. Acrobat and PDF were not immediately successful even after Adobe made its Acrobat Reader free to download. The PDF eventually became standard. It made paperless office a reality.

Though Adobe is known for its PDF, it has designed other programmers such as InDesign, Photoshop and Illustrator. All these have made our computing experience aesthetic engagement which formerly consisted of obscure commands and monochromatic images.

Adobe’s market capitalisation in 2023 is $235 billion. It is thus one of the largest IT companies in the world. Despite this, it has faced much larger companies such as Microsoft.

So far, India did not have any crash tests for the car, and relied mainly on foreign tests. Bharat NCAP or New Car Assessment rates the car on its crash performance. These are crash tests regulations, and these are aligned with international testing protocols. The programme will be implemented from October 1, 2023. It is a voluntary programme, and is not mandatory to get BNCAP certification. The crash cars will be selected randomly from the carmaker. These will be crash-tested at the testing centre. Depending on the results, certification and start rating will be issued. The certificate will be issued by the Central Institute of Road Transport (CIRT).

The testing will cost the car-maker Rs.60 lac, as against Rs.2.5 crore abroad. Indian car makers have made testing requests — as many as 30 requests are pending. The safest car is assigned five stars.

The programme is applicable to M1 category of vehicles — carrying 8 seats including the driver, and weighing less than 3.5 tonnes gross.

Rockets have been operating for a century in a similar manner. There is a tank that stores chemical fuels. These are ignited and they spit out a nozzle at high speed to create thrust. Space journeys require a lot of fuel as there are no refills available in outer space. Major portion of the tonnage of rockets putting weather and communication satellites in space is accounted by fuel.

In the 1950s, scientists thought of an alternative fuel. There can be a small nuclear reactor to heat up a propellant, say liquid hydrogen, to high temperatures (higher than what could be achieved in a chemical rocket). That will make rocket engine efficient — twice as efficient as a traditional rocket and faster too. A chemical engine will burn out at such high temperatures.

This is in essence nuclear thermal propulsion (NTP). There was research on this till the early 1970s. Such rockets cannot be used as launch rockets, since an NTP system lack the thrust to leave the earth’s surface. Instead, an NTP rocket could be carried into space on a traditional rocket and then make it operate there. So far no such experiment has been conducted but there are tests on the ground which demonstrate that the concept could work.

NTP as a concept has never been forgotten. There are issues of safety. However, tech advancements have placed it on the agenda of space authorities.

National Aeronautics and Space Administration (NASA) has partnered with the department of defense. It has given Lockheed Martin $500 million to build and test superfast nuclear-powered engine by 2027. If this is not done, NASA cannot compete with private companies like SpaceX and cannot realise its dream of putting its mark on more parts of the solar system and make US the leader in outer space.

The immediate objective for NASA is the journey to Mars. By 2030s, there should be manned flight to the Red Planet. Traditional rockets take seven months to reach Mars. A round trip mission would take two to three years. If an efficient rocket is used, Mars could be reached in 45 days. It augurs well for the physical and psychological well-being of the astronauts. It is an incentive to make more trips.

It is a competition among the US, China and Russia. The other countries try to neutralise the US.

A nuclear powered rocket is an untested entity in space. There are issues of its performance in extreme conditions. If there is an accident at the time of launch, there could be an environmental and health catastrophe.

The comforting part is that such a rocket is easier to build new than it was in the past. Material science has made rapid advances. It has come to a stage where an entity could be engineered that can withstand outer space and a heat of a nuclear reactor. Computers have advanced, so that complex reactor designs can be subjected to simulations, and redesigns very quickly.

Success is not guaranteed but the effort is worth.

The Associated Press has issued guidelines on AI. It has set rules on how to integrate tech tools like ChatGPT into their work.

The aim of the news organisation is to give a good story. As we are aware, generative AI has the ability to create text, images, audio and video on command. However, AI is not yet fully capable of distinguishing between fact and fiction. AP therefore, advised that AI-generated material should be vetted carefully ( like any other material from any other source). AP advises that a photo, video or audio segment generated by AI should not be used (unless such altered material is itself the subject of the story.) This aligns well with the tech magazine Wired’s policy.

There are instances of ‘hallucinations’ or made-up facts in AI-generated content. Consumers must know that there are standards in place so as to ensure that the content they are reading is verified, credible and fair.

Generative AI can facilitate publishing in a number of ways. Stories can be summarized or be put in the form of a digest. AI can be used to create headlines. AI can generate story ideas. AI can be used to make a story concise, and more readable. AI can suggest possible questions for an interview.

AP has simply used AI for a decade. It has created short news stories out of sports box scores or financial data. They want to enter the new phase of generative AI with caution. They are careful about their credibility. ChatGPT and OpenAI has struck to deal with AP of licensing news stories for training purposes.

News organisations are also concerned about the use of their content by AI companies without permission or payment. Their associations arrive at a deal with AI and Big Tech companies to protect the IP rights of their members.