The US and Iran war of 2026 was a battle between the AI Goliath and the NI (Natural Intelligence) David. The US pivoted from the conventional body count that defined the 1960s US wars to infrastructure damage that defined the 2020s wars. Should we see this as a precursor of the impact of machine thinking on human thinking? For a society that has always believed that the best thinking is a gadgetless one, will NI David’s stones worry the AI Goliath any more? <\/p>\n\n\n\n
Are we doomed to a ‘Matrix’ like existence? Will our future reputation be captured solely in numeric scores created by machines that constantly capture data about our behavior? We always felt that design and creativity would be the last bastions of human work, but they have been the first to crumble. Are we going to see a reversal of roles – where the only jobs that survive are the blue collared ‘routine’ ones – waiters, housemaids, cooks, personal groomers, plumbers and electricians – with everything else delegated to agents and machines?<\/p>\n\n\n\n
Prof Sanjiv Vaidya, who taught at IIMC for over three decades, gave us some ground for optimism. Technology and environment are opposing forces – every time tech converts unstructured tasks to structured, the environment hits back – creating more unstructured tasks in the process. Think Covid and healthcare, for instance. Another nugget from the Prof. The author of the initial ‘Friends’ episodes, who passed away recently, mentioned that AI will never be able to understand ‘the pain behind humor.’ And I would add – engage with live audiences in concerts and dramas.<\/p>\n\n\n\n
We had two participants who were in grades 8 and 9. And a lot who were in their twenties and thirties. Aditya had just finished his BA in German from SP College. His worry: With most translation taken over by tech, will there be a future for graduates in foreign languages? A few college going types were worried: How will college admission committees get to know whether college applications are AI fabricated? The common questions that all of them had: What sectors will see dominance of AI in future? What are sectors that AI will have lesser impact on?<\/p>\n\n\n\n
Coding will soon be as contemporary as using a slide rule. If your job is recycling existing content to produce more – whether it is a website, a blog, a podcast, or a PPT, then you need to start hunting for a new job asap. If you work in a market research agency, and you scrape the internet to make sense of all that stuff for a client – rest assured the client has a much cheaper alternative now, and that too working 24 by 7. If you are into mass production – stay away from production and quality – those jobs will be gone in the blink of an eye. The best jobs are going to be in maintenance and sales. If you are in production, it would help if your firm focuses on low volume customised solutions. <\/p>\n\n\n\n
The participants in the fifties and sixties, on expected lines, talked about deteriorating educational standards. Mandar, who runs the Pune center of Bulls Eye, talked of reduced attention spans – as evidenced in SAT reading passages. They are down from a page to 6 lines, from 10 questions to 2. Prof Vaidya recounted a dinner in Coorg, where the restaurant was playing a song mix, each song would only play for one stanza. It seems that Gen Z has lost its abilities to focus on a task for more than 5 minutes. How will humanity face the challenge of deskilling and cognitive degradation that is expected when a generation becomes slaves to AI even in their learning? Walking with crutches leads to muscle atrophy, can we say AI will soon lead to mental atrophy?<\/p>\n\n\n\n
As is the norm, in all things education in India, the conversation soon shifted to engineering and medicine. Ashutosh, our grade 9 friend, wanted to know if CS engineering programs will be redundant by the time he finishes his graduation? Mandar mentioned a preference shift to branches like Civil \/ Mech \/ Electrical. Though CS is better than fields like AI \/ ML – courses which focus on fundamentals are better than specialised ones. You can navigate implementation challenges better with a focus on fundamentals. <\/p>\n\n\n\n
The outlook is brighter for medicine, thanks to the ethical and legal moats that surround healthcare. Vaibhav, an IISER and IIMC alum – who comes from a biology background and is currently doing some AI consulting in the pharma industry, believes that in the healthcare field, diagnostics will see fast adoption of AI, but prescriptions will still require a human touch. The jury is still out on surgeons. For a country like India, robotic surgeries may see reduced costs and a democratization of access to specialised healthcare. But, we will still need to resolve trust issues – in all of surgery, diagnostics, and prescriptions. Vaibhav comes from a Pravaranagar agricultural family – and it would have been good to get insights from him on what agricultural jobs he thinks are under threat from AI. Next time, maybe.<\/p>\n\n\n\n
What will happen for sure is the use of AI in decision support systems in healthcare. We are already seeing something similar in law, which like medicine, has a nice legal moat that protects it. So do areas where the government has a monopoly – think regulators, policing and the army. Dhruba, a civil servant turned music teacher who also advises on taxation, talked about the supreme court’s draft policy on AI, where it talks of AI collated case notes to judicial officers. <\/p>\n\n\n\n
Another insight from Dhruba – who uses a lot of tech in his online music classes – a first time music learner should not be bombarded with too much information. I am reminded of the great marketer at Apple, Steve Jobs, who ensured that his kids did not have any digital devices around when they were growing up. I am not too sure what roles teachers will have to play. It surely is not going to be checking students’ homework – that is in any case being done by ChatGPT and the like. Unless teachers transition to become mentors, the future is bleak for them. <\/p>\n\n\n\n
In an era of big data, Garbage In Garbage Out is a big problem. We have to be patient. Almost all of us are using AI – even browsers start by giving you AI generated answers and AI generated news headlines. To use AI well, you need to know how it works, what assumptions the models have made. Shreyas, who designs products in wireless communication, tests new AI models with a standard set of prompts – to understand hallucinations to decide on reliability and trust. We need to understand the output and know how to shape it. <\/p>\n\n\n\n
We should create prompts on your own – though we may use machine help for future iterations of the prompt. Shreyas is a great coder – and is sad that AI does not allow him to spend as much time coding as he used to earlier. I wonder whether the topdown AI push in tech companies has been overdone. <\/p>\n\n\n\n
Every time tech changes, there are job losses. Bullock cart drivers had to look for other jobs when motor vehicles came on the roads. With AI, it is easier to start businesses as the cost of product development has been slashed. The corollary – expect more competition now. The mantra: Build fast, Go To Market Fast, Learn Fast, Fail Fast. Entrepreneurs can now throw out a lot of ideas, with risk of failure mitigated, at the market and find out what works. When Varun, who works in the tech healthcare space, presented his mouth cancer treatment device to Gautam Adani, the one question he had – how will you scale this? Can Varun get the price down from 700,000 to 700?<\/p>\n\n\n\n
Simran talked about her interaction with ChatGPT’s Sam Altman at IIT Delhi – and his idea of zero person companies. How should the world respond to zero person companies? Will society become a collection of entrepreneurs running zero person companies? How do we prepare for this? Gaurav, who has his AI Reddit agent working for his clients, believes that the only moat in the AI world is going to be ‘proprietary data.’ A question that I would have wanted Gaurav to answer is about the ethics of bots which want humans to believe that they are also humans.<\/p>\n\n\n\n
How do we find the golden mean between productivity gains and quality loss? There are opportunities in making AI more efficient, in reducing the damage to the planet with all the energy that AI is using up in its thinking of the world’s problems. What would be efficient use of AI? Should there be a cap on AI? Dhruba talked about his law firm’s practice of sharing tokens – it is easier done in smaller groups. RK, a COEP alum, who is working in building products for defence intelligence, talked of local LLMs that will be much less energy and capital intensive. COEP is working on one, with a budget of a couple of crores, using the likes of DeepSeek, to mitigate token cost.<\/p>\n\n\n\n
There will always be a ‘premium’ for human services. My personal bet would be on ‘human’ call centers to handle support and complaints. There is always the hope that problems will actually be solved when you talk to a human versus today’s bots. To survive in an AI world, humans will need to develop judgmental qualities: curiosity, empathy and responsibility. Though bots will soon be getting better at handling human emotions, they will only be superficially constrained by rule based ethics and accountability. <\/p>\n\n\n\n
Thinking is always iterative. Prompts should work the same way. Keep in mind, AI thinking is probabilistic, not deterministic. Learn to live with constraints – having limited resources makes you creative. The ‘what to build’ will still be a human decision. In refining an AI generated design, decisions on human factors are best validated by humans. AI is commoditising intelligence, thankfully judgment is not yet commoditised!<\/p>\n","protected":false},"excerpt":{"rendered":"