The Talent Barbell: What Arun Varadarajan Says Builders Should Do Now
Where we are right now in the world is simple: people are anxious about artificial intelligence and their livelihoods. In this interview, I sit down with Arun Varadarajan, Chief Commercial Officer and co-founder at Ascendion, to get straight into the reality on the ground: the shifting engineering landscape, the economics driving these decisions, and where the future is actually headed.
Picture the scene.
It is three minutes to eight in the morning. You are sitting at your desk, the coffee still untouched on the coaster. Outside, the dog next door will not stop barking. Then the notification pings.
An email from Human Resources.
Your heart hits your throat. What am I going to do? How am I going to live? Who is going to pay the bills?
Then comes the twist. You open the attachment, and the severance figure stops you cold. It is large enough that you could do nothing for two years. For ten seconds, you jump up and down with relief.
And then the cold question lands: and then what?
That scene is my own dramatisation, not something Arun said. But the question is real, and it is haunting every engineer, architect, and technology professional right now.
The Layoff Narrative
Corporate boards love a clean public relations story. When leadership announces rolling workforce reductions, executives often point to artificial intelligence as the justification. It sounds forward-thinking, decisive, and inevitable.
To explore what is really going on, I spoke with Arun Varadarajan. He is Chief Commercial Officer and co-founder at Ascendion, an AI-native software engineering company serving Global 2000 enterprises, with three decades across telecoms, cloud, and enterprise AI. He argues from the engineering floor, and his focus is the economics of what happens when AI stops being a co-pilot and starts doing the build.
Here is the heart of his thesis. When the cost of building collapses, everyone can become a builder. For fifty years, software was gated by syntax, degrees, budgets, and teams, and most ideas died in the gap between seeing a problem and being able to build the solution. That gap is closing. Arun uses the printing press as his analogy: it did not create more scribes, it created more authors. In his view, AI will not create more coders. It will create more builders, and those are not the same thing.
On the layoffs themselves, his explanation is pressure rather than inevitability. Leaders have promised the street large operating savings from AI, but top-line growth has not kept pace, so attention turns to the bottom line. In his view, that is where the problem starts. If he were running one of these companies, he would focus on keeping people with deep domain knowledge. And he pointed out that talent which leaves may well start something new that takes the incumbents head on.
Below is the playbook drawn from our conversation. Where an idea is Arun's, I say so. Where it is my own extension, I say that too.
1. The Talent Barbell (Arun's Framework)
For decades, software engineering ran on an apprenticeship pyramid: many junior people doing repetitive tasks, slowly absorbing how systems work, and eventually maturing into senior roles. Arun's view is that AI breaks that model and replaces it with a barbell.
The Top: Subject matter experts. Arun compares them to super surgeons who have done thousands of operations and know exactly where a process will fail, where it will succeed, and where the risk sits. Building has become cheaper, but knowing where the problems lie has become more valuable as systems grow more complex. His advice is to nurture this talent and not let it go. These are the system designers of the future.
The Bottom: Fresh graduates. Arun takes a contrarian view here. Most people think the bottom of the ladder is collapsing. He thinks this is the time to bring in fresh minds with no baggage. They learn like medical residents, who do not learn by operating on day one but by watching the best surgeons work. They observe agents doing the toil, learn to review the agents' output, and watch how the experts think. The apprenticeship or bootcamp model for new joiners has to fundamentally change.
The Middle: The thinning middle. Arun's concern is for layers of the organisation whose role has drifted into coordination rather than domain expertise, forwarding emails and moving paper rather than contributing deep knowledge. That is the group he sees at greatest risk.
My own extension (not Arun's): the same logic applies to individual engineers. If your day is passively taking tickets and translating specifications into standard code, that is exactly the work agents absorb first. Two directions look stronger: going deep on foundations (distributed systems, databases, security, systemic risk), or going deep on a business domain while directing agents to deliver working systems. Arun's own emphasis is domain over syntax, which leads to the next point.
2. Domain Over Syntax (Arun's Framework)
Arun argues the new builders will combine first principles thinking with a deep sense of the problem they are solving. First principles means starting from the fundamentals rather than from the rules a company or industry has accumulated. His example was Elon Musk questioning why rockets had to cost what NASA said they cost, and whether they could be recovered and reused. The syntax, in that case, was whatever NASA would say. Arun's point is that building is now cheap, so the scarce skill is knowing what is worth building and why.
3. Doing Better Things, Not Doing Things Better (Arun's Framework)
Most leaders ask how AI can help them do the same things faster or cheaper. Arun warns that measuring AI by hours saved is the trap, because it replaces thinking with cost cutting. The real prize is doing better things.
He pointed to the long shadow of big software vendors. Many corporate clients have felt their pace of innovation was dictated by their suppliers, waiting quarters for a single capability. With the cost of building collapsing, building your own technology becomes a viable option and a way to control your own destiny. And there are plenty of unsolved problems worth the effort, from clean energy to basic water, food, and hygiene for large parts of the world.
The question is not how to speed up what you already do. It is what high-impact systems you can build now that the cost of building has collapsed.
4. Guardrails and Mandates (Arun's Framework)
Arun shared a lesson from his own company. About three years ago he mandated prescriptive ways of working on AI. It made him feel good, but nobody did anything about it. When he let the pressure off and invited engineers to build their own agents, they began solving problems nobody had even identified. His conclusion: you cannot mandate a generation of builders, but you can stop getting in their way.
He also argues that governance and guardrails belong inside the system, not in documents managed by people. Documents are not followed every time. Guardrails built into the system are. He was clear that regulators matter and should not be ignored. The point is to embed compliance so it stops slowing builders down.
5. Three Predictions for 2029 (And Your Monday Morning Action)
I asked Arun to put three predictions on the record and, for each, to tell me what a leader should do on a Monday morning. He was upfront that these are predictions, and that only time will be the judge.
Prediction 1: Core enterprise systems come under pressure
Within about two years, Arun expects companies like Salesforce and SAP to be in trouble. Customers will ask what differentiation they receive when they could build software that runs their payroll or accounting themselves. Systems people believed they would never replace may change.
- Your Monday Morning Action: List every constraint you use to run and operate your business. Re-examine why you need each one. Over the next six months, aim to remove every constraint that cannot justify itself.
Prediction 2: The tiny giant
Arun expects between fifty and one hundred companies valued within the global top thousand to have no more than ten employees. He gave a range and was open that it was his best estimate.
- Your Monday Morning Action: If the organisation you work for is slowing you down, Arun's view is that it may not be the place for you, and that there could not be a better time to start something new. Think about asymmetric value. If customers pay a big vendor a notional one thousand dollars a month, build something that delivers the same value for a dollar. That means a new construct with minimal bureaucracy, where everything is system-driven and AI is used to the maximum.
Prediction 3: The shelf life of software falls from years to weeks
In Arun's words, it will be a very interesting world of builders and consumers, and consumers and builders.
Why the Future Belongs to the Builders
If you are an engineer or technical leader feeling the weight of industry shifts, here is the takeaway from Arun: the cost of building has collapsed, so the scarce skills are now domain knowledge, first principles thinking, and the courage to build something new. Do not wait for your organisation to tell you where you stand. Run the audit below on your own role this week.
The 4-Step Career Barbell Audit (My Framework, Built on Arun's Barbell)
Step One: Build your professional dossier. Gather your professional history: core competencies, daily responsibilities, technical proficiencies, and the specific problems you solve. Exclude any proprietary or confidential corporate data.
Step Two: Identify the mechanical overhead. Submit your dossier to a frontier reasoning model and enter this prompt:
Analyse my current day-to-day responsibilities against what autonomous AI agents and workflow automation are likely to execute within the next twenty-four to thirty-six months. Estimate what proportion of my current role represents mechanical execution, syntax generation, or routine administration, and explain your reasoning.
Step Three: Map your transition. Once the model has separated out the routine mechanics, enter this follow-up prompt:
Based on my specialised background, what high-value bottlenecks are likely to emerge in my sector that demand deep domain problem-solving, systems architecture, human governance, and trusted stakeholder alignment? Detail the transition path required to position me at either end of the talent barbell.
Step Four: Build an independent capability. Build something you completely control. Map your personal workflow logic, automate your routine administration, and stop relying solely on someone else's corporate platform for your leverage.
Connect and Join the Discussion
- Connect with Arun Varadarajan on LinkedIn: Reach out directly at linkedin.com/in/arun-varadarajan-0a26393/. Send him a message letting him know which of his predictions you want to test first, and mention that you heard him on Influential Visions.
- Listen to the Full Interview: Tune in to the complete podcast episode on the Influential Visions podcast feed or watch the full video discussion on YouTube.
If the cost and time required to build custom software dropped to near zero tomorrow morning, what is the single tool, platform, or workflow you would build for your business first? Tell me in the comments or reach out to me directly.
100% Ad-Free Strategic Intelligence
Join MONDAY INFLUENCER® for direct access to weekly actionable briefings, prompt engineering frameworks, and private advisory sessions.
