The numbersBriefEpisode mapTakeawaysChartsQuotesClipsGlossaryGo nextTranscript
Episode 140 · Impact · 48 min

The fourth thing to flow through the wires

Food became abundant, then power, then data — and Saxena's claim is that intelligence is now the fourth utility on the wire, billed monthly like electricity. Every previous abundance multiplied jobs instead of erasing them: Indian IT went from about 500 people in 1985 to roughly five million. Scaler's bet is that the five million doing intelligence work today become fifty million — but only for the people somebody retrains in time.

Education & skillingAI & machine learningHiring & talent
AS
Abhimanyu Saxena
Co-founder, Scaler · with Vishal Krishna
The fourth thing to flow through the wires — episode thumbnail
48:20
Said in this episode
▶ 6:58
500 → 5M
Indian IT headcount, 1985 to now
Saxena cites an IT veteran from TCS's first hiring batch who put the whole Indian IT industry at a maximum of 500 people in 1985; the guest's figure for today is about five million, though the host uses two million in the same exchange.
▶ 7:42
5M → 50M
His forecast for intelligence jobs
Roughly five million people do intelligence work in India today by his estimate; he expects the AI shift to take that to fifty million, on the same pattern as the IT jump — explicitly a belief, not a study.
▶ 10:00
~30%
The demographic bet, as stated on air
Within ten years, he says, 'we are going to have almost 30% of world population which is between age of 15 to 35' — the caption leaves ambiguous whether that is India's share of the world's young or the world's own age mix.
▶ 41:35
450 / 12,000
Scaler School of Technology intake
About 12,000 sat the entrance test and 450 were admitted; he insists many of the rejected were good, and that taking 6,000 at the same price would have destroyed the outcome.
▶ 38:05
5–10M/mo
Direct users on Scaler's owned platforms
InterviewBit and Scaler Topics together draw five to ten million users a month in direct traffic, most of whom never become paying customers — the hedge against platform dependence.
▶ 25:48
₹10,000 cr
Annual public spend on the IITs, as cited
He cites roughly ₹10,000 crore a year going into the IITs and asks whether budgets of that size could buy more upskilling per rupee — a figure quoted from memory in conversation.
The brief

The argument in sixty seconds

Saxena's claim is that intelligence is simply the fourth thing to become abundant — after food, after power, after data — and that abundance has never once subtracted jobs. Between 1820 and 1920 the world's job count multiplied roughly a hundredfold; India's IT industry went from about 500 people when TCS hired its first batch in 1985 to some five million now, and he expects the same multiple in intelligence work: five million today, fifty million after. The catch is the transition, and that is where Scaler makes its money. He describes six companies bidding for the same engineer while each holds six people it is quietly preparing to release — people deep in the culture, missing one hard skill that three or four months of training would fix. Do the arithmetic he does: two months of severance plus one to two months' salary as recruiting cost is four months burnt on a swap, when a quarter to a half of that would retrain someone who already knows the system, and the outside hire needs three or four months just to settle. Inside Scaler the same logic already ran: an AI now audits every admissions call against a hundred rules and flags a breach within two minutes, and the twenty people who used to sample a tenth of the recordings were moved onto counselling students — work he argues pays better precisely because a machine cannot do it. He is selling that argument to CHROs, to GCCs, and to three or four governments — not as charity, he insists, but as value creation.

Worth your time if you are

CHROs holding a redundancy list and a hiring plan at once
Engineers afraid the model gets there before they retrain
GCC leaders told to staff 500 seats this quarter
Operators wiring AI into a live customer-facing team
Small-town founders with no obvious head start
Episode map

Where the conversation travels

Every block is a chapter, coloured by what it's about. Click any of it to jump straight to that minute on YouTube.

010:00–2:19 · Cold open: songs from a nervous planet022:19–5:09 · Four revolutions, three of them survived035:09–7:27 · Intelligence on a monthly bill047:27–9:27 · Five hundred to five million to fifty9:27–10:43 · Technology education, not education technology0610:43–14:07 · One auditor now covers every call0714:07–18:25 · Six offers out, six people out the door0818:25–21:30 · Two pains a CHRO thinks are separate0921:30–26:00 · Governments buying reskilling at scale1026:00–30:35 · A village with no telephone line1130:35–33:55 · Why they left Berlin and London1233:55–36:50 · The problem statement is employability1336:50–40:30 · Paying the Google and Facebook tax1440:30–44:40 · Saying no to twelve thousand applicants1544:40–48:20 · Six hundred books and an idiot hero0:0010:0020:0030:0040:0048:20
Education & skillingAI & machine learningHiring & talentGCCs & servicesIndia macroFounder journey
01Cold open: songs from a nervous planet 0:00 Vishal frames the episode around Ratan Tata's line that India must employ 150 million people by 2030, and asks the existential question — what happens to education and work on a planet this jumpy. 02Four revolutions, three of them survived 2:19 Saxena counts the moments that changed everything — food became abundant with agriculture, power with the steam engine, clerical processing with computers — and argues each one arrived with exactly the same panic. 03Intelligence on a monthly bill 5:09 Power flowed through wires, then data flowed through wires, and now intelligence does — priced like an electricity connection — while the historical record shows jobs multiplying about a hundredfold between 1820 and 1920. 04Five hundred to five million to fifty 7:27 The Indian IT industry went from roughly 500 people in TCS's first hiring year to some five million in thirty years, and Saxena expects intelligence work to make the same jump from five million to fifty — which is a story about upskilling, not about breaking the machines. 05Technology education, not education technology 9:27 He flips the edtech label deliberately, then puts India's demographic bulge on the table: a young population at peak capability is a superpower if it is skilled and a hazard if that energy has nowhere productive to go. 06One auditor now covers every call 10:43 Scaler's own admissions team is the worked example — a hundred rules running over every recorded call, flagged to a manager in two minutes, replacing a 10% random sample done by twenty people. 07Six offers out, six people out the door 14:07 The same candidate holds six offers while each of those six companies is preparing to release someone who only lacks one hard skill — and the swap costs about four months of salary that a quarter to a half would have spent on training. 08Two pains a CHRO thinks are separate 18:25 Saxena describes the enterprise conversation as backtracking from where a CHRO or CTO must be in six to twelve months, and the recurring surprise that the exit problem and the hiring problem are one problem with a bridge missing in the middle. 09Governments buying reskilling at scale 21:30 Three or four governments already run national reskilling projects with Scaler — for value creation, he stresses, not charity — while in India the intent is there but the fuel question stays open between CSR budgets and public education spend. 10A village with no telephone line 26:00 Growing up in an MP village with day-old newspapers and no petrol pump, he reached IIIT Hyderabad having only ever used a computer to check exam results — and the friend who read getchar and putchar as 'getcher' and 'putcher' now runs engineering at Google. 11Why they left Berlin and London 30:35 Two engineers quitting Facebook and a Berlin job in 2014 chose India on analysis rather than sentiment — highest GDP growth of any large country, unrealised resources, and his forecast of another 100 to 200 unicorns here in a decade. 12The problem statement is employability 33:55 Scaler is not an engineering company by doctrine — the mandate is whatever makes people valuable, which is why 19-year-olds running startups that make a few lakh a month count as a success case and management or biotech stay open options. 13Paying the Google and Facebook tax 36:50 Every business pays a discovery tax to the platforms, so Scaler built InterviewBit and Scaler Topics into properties pulling five to ten million monthly users — performance channels for growth, owned distribution to survive. 14Saying no to twelve thousand applicants 40:30 He turns away enterprise clients he cannot serve with excellence and admitted 450 of 12,000 test-takers into Scaler School of Technology, then admits the hardest iteration is balancing academic fundamentals against what recruiters actually ask. 15Six hundred books and an idiot hero 44:40 The close turns literary — a 600-book home library, Premchand's line that the worst fate is a dead dream, a friend's fear of feeling his peak is behind him, and Dostoevsky's The Idiot, where the protagonist is the fool and the hero at once.
Takeaways

Ideas to carry out of this hour

01

Every abundance so far multiplied the jobs

Saxena's frame is four moments in which something scarce became abundant — food with agriculture, power with the steam engine, data with computers, and now intelligence. Each arrived with the same panic and each ended with more work, not less: he puts the world's job count at roughly a hundred times higher in 1920 than in 1820. His point is not that nobody was displaced, but that the aggregate has never once shrunk, which makes the current fear a transition problem rather than an endgame.

▶ 2:49▶ 6:12▶ 6:26
02

Intelligence is now a metered utility

Electricity was the thing that started flowing through wires in the industrial era; data was the thing that started flowing in the computer era; intelligence is what flows now. The consequence he draws is a pricing one — in a few years you pay a monthly bill and take as much intelligence as you want, the way you take power. That is why he thinks the anxiety is real: most white-collar work is some intelligence applied to a problem, and a metered supply of it will reach a large part of that.

▶ 5:09▶ 5:23▶ 5:39
03

Five hundred became five million, and can again

When TCS hired its first batch in 1985, Saxena says the entire Indian IT industry held at most about 500 people; thirty years later it is somewhere around five million. That jump required massive upskilling — the nature of the job changed repeatedly and people had to learn new things to meet it. Applying the same multiple, he expects the five million or so doing intelligence work today to become fifty million, with the same condition attached.

▶ 6:58▶ 7:42▶ 7:58
04

AI took the audit, not the auditors

Scaler's admissions calls were being sampled at about ten percent by a team of twenty people, which caught problems late and incompletely. Every call now runs through an engine with roughly a hundred rules; a flagged breach reaches the manager within two minutes and the counsellor is corrected in real time, which he describes as ten times the coverage at twenty times the speed. The twenty people were moved to escalation management and career counselling — work that pays more, he argues, precisely because listening to recordings all day never could.

▶ 12:33▶ 13:04▶ 13:50
05

The person you fire and the one you hire are one problem

Saxena sees a candidate holding six offers while each of those six companies quietly carries six people it is about to let go — people with years of context on the culture and the systems, short one hard skill that three or four months would close. Firms treat exit and hiring as separate projects run by separate people, and the CHRO who calls him for an outplacement consultant calls again a fortnight later needing 500 smart engineers. His argument is that the bridge between the two is the cheapest talent strategy available, and that the pattern has only become frequent enough to notice in the last twelve to twenty-four months.

▶ 15:53▶ 16:23▶ 20:10
06

Swapping a person costs four months; training one costs less

The arithmetic he runs for enterprise buyers: letting someone go means at least two months of severance, and hiring the replacement means one to two months of salary as recruiting cost — about four months of an employee's cost burnt purely on the transition. For a quarter to a half of that you could have trained the person you already had. The new hire then takes three or four months to settle into systems the incumbent has known for years, so the retrained employee is likely to perform better as well.

▶ 16:54▶ 17:08▶ 17:24
07

Governments are buying reskilling as strategy, not charity

Scaler already works with three or four governments globally on national reskilling projects, and Saxena is emphatic that these are value-creation programmes rather than welfare — the buyer expects efficiency back from the system. In India he reports plenty of positive intent at both the centre and the states but little execution yet, blaming size more than will: a big elephant cannot run like a cat. The open question is fuel — whether the money comes from CSR budgets, education budgets, or the roughly ₹10,000 crore a year he says goes into the IITs.

▶ 22:00▶ 22:31▶ 25:48
08

Own the distribution; rent the platforms only for growth

Every business pays what he calls a tax to Google and Facebook for discovery, and he grants that they have earned the premium. But Scaler spent ten years building InterviewBit and Scaler Topics into properties with five to ten million direct monthly users, most of whom never pay. That is the strategic line he draws — performance channels are for growth, never for sustenance — and it pairs with a discipline of saying no: 450 admitted out of 12,000 who sat the entrance test, and enterprise clients turned away when he cannot serve them with excellence.

▶ 37:58▶ 38:29▶ 41:35
The numbers, drawn

What the episode measures

Every figure below was said on air — timestamps included, caveats kept.

Conversation share

portion of the hour spent on each theme
Education & skilling · 22%AI & machine learning · 18%Hiring & talent · 16%GCCs & services · 11%India macro · 11%Founder journey · 10%
Education & skilling22%
AI & machine learning18%
Hiring & talent16%
GCCs & services11%
India macro11%
Founder journey10%
Computed from the chapter map of this episode.

The jobs bet, in headcount

people employed
Intelligence work to5,000,000After the AI shift50,000,000
Saxena's own round numbers, offered as belief rather than research: about five million people in intelligence jobs today, possibly fifty million after the shift — the same 10x he attributes to Indian IT since 1985.▶ 7:42

Replacing a person versus retraining one

months of salary
Swap: severance plus4Retrain the person y2
As stated in conversation: at least two months of severance plus one to two months' salary as hiring cost, so roughly four months burnt on the transition, against 25–50% of that to train the incumbent (upper bound shown).▶ 16:54

Scaler School of Technology's funnel

students
Wrote the entrance t12,000Admitted450
Numbers as recalled on air for the latest intake — 'about 12,000' sat the test, 450 were taken; he says 6,000 could have been admitted at the same price and that doing so would have destroyed the outcome.▶ 41:35
Worth keeping

Lines that stay

Now the intelligence is flowing through the wires. Just the way you have an electricity connection — in a few years you can pay a monthly bill and have as much intelligence as you want.

— Abhimanyu Saxena ▶ 5:23

It's like an avalanche. You may try to stop it, but you can't. So if there's an avalanche coming, what do you do? You ride on it.

— Abhimanyu Saxena ▶ 9:12

They are being let go just because of one missing hard skill — one that can be filled in three or four months of training.

— Abhimanyu Saxena ▶ 16:38

There is a speed and there is an acceleration. What matters is acceleration. Where you are today just doesn't matter.

— Abhimanyu Saxena ▶ 28:47

In an ideal world one should be using Google and Facebook only to grow, not to sustain.

— Abhimanyu Saxena ▶ 38:29
Clips that travel

Short on time? Start here

Anyone anxious about AI arriving for their job

The fourth thing to flow through the wires

The whole thesis in one stretch — four revolutions, the broken steam engines and smashed computers that greeted them, and the hundredfold jobs multiple that followed.

2:19 → 7:27 · 5 min ▶ Watch clip
Operators wiring AI into a live customer-facing team

One person now audits every call

A concrete before-and-after inside Scaler: 10% random sampling by twenty people replaced by hundred-rule coverage with two-minute flags — and where the twenty went.

11:30 → 14:07 · 3 min ▶ Watch clip
CHROs holding a redundancy list and a hiring plan

Six offers, six people on the way out

The episode's sharpest business argument, with the arithmetic attached: four months of salary burnt on a swap against a quarter to a half of that to retrain.

14:07 → 18:25 · 4 min ▶ Watch clip
Anyone selling skilling into the public sector

Governments as reskilling customers

Three or four national governments already buying at scale, why he insists it is value creation and not charity, and the honest read on India's intent-versus-execution gap.

21:30 → 24:50 · 3 min ▶ Watch clip
Students and founders outside the metros

A village with no telephone line

Day-old newspapers, a cyber cafe that existed to show exam results, the getchar-and-putchar friend now running engineering at Google, and the line about acceleration beating position.

26:00 → 30:35 · 5 min ▶ Watch clip
Glossary

The jargon, unpacked

GCC
Global Capability Centre — the engineering or operations arm a multinational staffs and runs itself in India rather than outsourcing it, and the buyer under the most pressure in Saxena's account to fill hundreds of seats at short notice.
Reskilling
Training an employee into a different skill so they can hold a new role, as against upskilling deeper into the one they already have — the swap-versus-train arithmetic is an argument for doing it in-house.
Employability
Saxena's stated problem statement: whether a person's skills make them valuable to the industry, whether they end up taking a job, joining a startup or building one that hires interns.
InterviewBit
Scaler's free coding-practice platform, which together with Scaler Topics draws five to ten million direct users a month and functions as owned distribution rather than a paid channel.
Scaler School of Technology (SST)
Scaler's residential undergraduate programme for 18-year-olds, where 12,000 sat the entrance test and 450 were admitted in the intake he describes.
NEP
India's National Education Policy, which pushes reskilling and industry alignment down into schooling — the host's example of the state stepping into the skilling business.
CSR budget
Corporate social-responsibility spending Indian companies set aside for social causes; one Scaler client offered to point it at training people from disadvantaged communities.
Connections

If this resonated, go here next

EP 128 · Consumer

Who pays for the bridge to a job

Krishna Kumar (Simplilearn)
Education & skillingHiring & talent
Deep dive
EP 171 · Capital

The tag is worth ten per cent

Pradyumna (Prequate)
AI & machine learningEducation & skilling
Deep dive
EP 83 · Consumer

The whole family takes the exam

Ujjwal Singh (Infinity Learn)
Education & skillingAI & machine learning
Deep dive
Full transcript

The whole conversation, searchable

189 segments

Auto-generated captions, lightly cleaned. Click a timestamp to open that moment on YouTube.