Will AI Take My Job? Why AI Will Create MORE Jobs, Not Less

The Core Argument

Guided by host Sanjay Saini, veteran Agile Coach Srikanth argued that the fear of massive AI job replacement is largely overblown. Srikanth explained that as the massive infrastructure investments in AI demand payback, token costs will rise until using an LLM essentially equals the cost of a human worker. Organizations attempting to use AI purely for cost-cutting will lose to companies using it to 10x their output. The future belongs to professionals who transition away from "doers" into "orchestrators"—managing a hybrid team of artificial agents and humans to deliver stable, secure, end-to-end solutions.

“You have to know more than what the AI knows. You employ the agent to work for you; it's got no imagination on its own.” Srikanth Ramanujam (22:23)

Detailed Answers

1. Could you give a brief about your professional experience and background?

Srikanth has spent decades mirroring the massive paradigm shifts in the tech industry, beginning with MS DOS 1.0 in the 1980s. He transitioned from writing software and selling computers in India to global outsourcing across banking, supply chain, and government sectors. Roughly 20 years ago, he organically shifted into Agile coaching after successfully scaling a delayed software project, and currently coaches product teams in the financial space.

2. Why do you think AI is going to create more work, not less?

Srikanth believes looking at AI purely through an "efficiency and job-cutting" lens is deeply flawed. The trillions of dollars invested by massive tech companies require immense financial payback. To achieve profitability, the cost of tokens will inevitably rise over the next decade. Eventually, the cost of running large language models for complex work will rival the cost of human labor. This creates a scenario where human-AI hybrid teams are absolutely necessary to justify the investment.

3. How effective is AI currently at writing production-ready code?

Citing industry data from Henrik Kniberg, Srikanth noted that only about 20% of AI-generated code is accepted without changes. Roughly 40% requires significant review, 30% demands a complete rewrite, and 10% is absolute garbage. Organizations require stability and predictability; if humans cannot understand why an AI's code broke production, the business cost is catastrophic. Because of this, AI will remain an assistive tool rather than a fully autonomous developer.

4. Will the increasing cost of tokens impact overall AI adoption?

Yes. The era of the free LLM is dying. Srikanth pointed out that OpenAI had to strictly limit free access, conceding the personal space largely to Google’s Gemini integration. In the corporate sector, AI is becoming a mandatory "tax"—much like an AWS or Azure cloud subscription. Organizations will have no choice but to pay for AI services to stay competitive, regardless of whether their profit margins immediately support it.

5. Will there be a moment that makes AI super cheap and accessible to everyone?

Rather than a pure software revolution, the real expansion of AI will be local and physical. Srikanth pointed out that upcoming smartphone generations will run smaller LLMs entirely on-device, bypassing costly cloud token processing. More importantly, massive physical investments—like Tesla replacing car manufacturing lines to build Optimus robots—signal that AI will soon govern physical labor and home services, opening up completely new economic sectors beyond coding.

6. How will AI alter the amount of work available for agile teams?

Companies attempting to cut a 200-person department down to 50 using AI will be destroyed by competitors who keep all 200 people and use AI to multiply their innovation by 10x. Srikanth reminded listeners that writing code is only 20-30% of the end-to-end product lifecycle. The future requires "human orchestration"—managers who direct artificial agents to handle coding while humans manage edge cases, security, and complex integration.

7. If organizations failed to adopt Agile, can they successfully adopt AI?

Srikanth is highly skeptical. He noted that 90% of companies failed to truly adopt Agile over the last 25 years, instead building fake "DevOps" pipelines that still rely on slow, large-batch releases. If an organization lacks the discipline to adapt its human culture to rapid delivery, it will fundamentally lack the agility required to utilize an LLM effectively. These slow, legacy companies will simply be outcompeted and die.

8. What is your advice to professionals who are scared about job losses?

Shed your fear and master the tool. Srikanth advised professionals to treat an LLM like a highly capable, $20/month employee working under their command. Because AI has zero imagination and cannot initiate pure innovation, it relies entirely on a human orchestrator. He also suggested using the AI to teach itself—instead of buying a book on prompting, simply ask the LLM to build you a structured, 3-week curriculum on how to master its capabilities.

Srikanth's shift from "doer" to orchestrator is a skills change, not a title change — directing models, checking their output and owning the edge cases they miss is the working pattern our Artificial Intelligence training program builds.