On Episode 913 of The Core Report, financial journalist Govindraj Ethiraj talks to Neil Shah, Vice President-Research & Co- Founder at Counterpoint Research as well as Jayanth Neelakanta, Founder and CEO of Equip.

SHOW NOTES

(00:00) The Take

(04:41) Indian Markets Will Stay Cautious On Oil Price Movements

(06:46) A US Judge ​Ordered The Justice Department On Friday To Justify Its Decision To Drop Criminal Charges Against The Adani Group’s Gautam Adani.

(07:56) Chipflation Is Here As Companies Like Apple Announce Sharp Hikes, How Worse Can It Get?

(19:53) How AI Is Helping Hiring Decisions And What Are The Job Market Trends That It Is Throwing Up

(30:21) Australia Is Doubling Down On 6-Month Ban On Child Access To Social Platforms

—

NOTE: This transcript contains the host's monologue and includes interview transcripts by a machine. Human eyes have gone through the script but there might still be errors in some of the text, so please refer to the audio in case you need to clarify any part. If you want to get in touch regarding any feedback, you can drop us a message on feedback@thecore.in.

Good morning, it's Monday the 29th of June and this is Govindraj Ethiraj broadcasting and streaming weekdays from Mumbai, India's financial capital.

The Take

The US missile strikes against Iran over the weekend, a direct response to an attack on an oil tanker in the state of Hormuz have once again tested the fragile ceasefire.

Yet in a testament to the resilience of the global economy, oil prices have remained muted, actually pulling back below $73, a level not seen before the four-month conflict began. This resilience invites a necessary audit of the last six years. We have navigated the body blow of COVID-19, which effectively rewired global supply chains and permanently altered the nature of work.

While the Zoom era brought us the convenience of remote work, it also underscored the decline of productivity without the social cohesion of an office. Looking back, one wonders, had the pandemic lasted four months instead of three years, would we have permanently abandoned the office or would we have recognised the screen as a poor substitute for real-world collaboration? It's not like the latest crisis has not helped, following as it did from the tariff war crisis, which hit economies around the world and caused more damage to businesses with uncertainty rather than the actual flow or lack of it of goods. But the net result was that supply chains have become even smarter and the US economy, fuelled by a massive AI investment boom, has proved remarkably robust.

Even with layered supply shocks, the S&P Global says from a report last week that the US will maintain trend-like economic growth of 2.1% through 2026. The resilience, it says, is no surprise. It's the natural outcome of substantial domestic energy production and lower energy intensity.

In India, the story is similar in vigour but different in risks. The economy grew a robust 7.8% in the Jan to March quarter, driven by private investment and construction. Though the government's response to the West Asia conflict has been a study in the dangers of the populist energy model, we've not achieved much in terms of moderating energy consumption because price signals were suppressed for months.

When the government delays price hikes, it does not prevent inflation, it often creates a black market, and the LPG shortage or liquid petroleum gas cylinder shortage is the most visible by-product. By keeping prices of 14.2kg cylinders artificially low while market costs rose, the government incentivised the diversion of supplies, in this case to commercial use. Government officials, including elected representatives, have offered half-hearted promises to reduce the size of ministerial convoys.

But like most such displays, these gestures are being forgotten as quickly as they were announced. But the most pressing challenge, however, now is not the price of oil. As many have now told us, it's the availability of water.

The India Meteorological Department says cumulative rainfall for the country for the month of June till 28th, that is yesterday, stood at about a 45% deficit. Mumbai, for example, is now looking at 7% water levels in primary lakes. The rains, which made a brief appearance after a two-week delay last week, are awaited once again.

Unlike oil, which can be secured through global markets, water is a local necessity with no substitutes. We may find ways to manage oil and gas supplies, but we cannot engineer our way out of a drought, or only partly so. The lessons of the last four months are clear.

Whether it's energy or water, we are entering an era of greater scarcity. The scarcity mindset must now shift from a crisis response to a baseline for future sustenance. Individuals and organisations should stop waiting for state-directed solutions, which are often delayed and distortive, and start managing their or our own consumption with greater discipline.

The last four months may not have been as instructive as the previous crises of this decade, but they have provided a sound warning. The year ahead promises to be the true test of whether we are capable of adapting to a world where resources are no longer guaranteed.

And that brings us to our top stories and themes...

Indian markets will stay cautious on oil price movements.

A US judge ordered the Justice Department on Friday to justify its decision to drop criminal charges against the Adani Group's Gautam Adani.

Chipflation is here to stay as companies like Apple announce sharp hikes. How worse can it get?

How AI is helping hiring decisions and what are the job market trends telling us?

Australia is doubling down on a six-month ban on child access to social platforms.

Markets, The War and a US Adani Case Update

Indian markets are responding to the oil price fall somewhat cautiously. One is for obvious reasons, which is that the ceasefire is fragile.

But the second is that the fall in oil prices is but not a necessary trigger for all the pre-war concerns to be wished away. Economists told us over the weekend, for instance, that oil prices may rise once again because there could be a rush to replenish depleted oil reserves, including in countries like India. Now, while a rash of incentives have brought in foreign portfolio investments, particularly into debt, there is not much movement on equities.

The rupee is thus in the same holding pattern, though drifting downwards and likely to remain there for some time. In the markets on Thursday, the benchmarks were up for the third straight week and that made it their longest winning streak in seven months after crude oil prices fell to pre-war levels. And of course, those measures to bring in more dollars.

Markets were closed on Friday for a national holiday. On Thursday, the Nifty 50 was up 34 points to 24,056 and the Sensex was up 109 points to 77,100. The broader markets were down slightly.

The Nifty mid cap and small cap were down 0.5 and 0.4. Each over the weekend, Brent crude prices were down to about $72.70 a barrel or just under $73 a barrel as more tankers exited the state of Hormuz following that initial US-Iran peace deal, according to Reuters. The rupee also strengthened this time for the fourth consecutive session on Thursday. Even as government bond yields declined, the rupee closed at 94 rupees 40 paise per dollar against the previous lows of 94 rupees 66 paise per dollar.

Incidentally, most Asian currencies have weakened against the dollar while the rupee as we can see is holding strong. Foreign investor interest in domestic debt has remained strong in June. Net foreign portfolio investment in debt was at about 8,100 crores on the 24th of June and the highest single-day inflow so far in 2026 and the second largest since 24, according to National Securities Depository data quoted by Business Standard.

Elsewhere, a US judge has ordered the Justice Department on Friday to justify its decision to drop criminal charges against the Adani group's Gautam Adani, declining to rule immediately on Adani's lawyer's request to dismiss the case, Reuters reported, adding that a Brooklyn-based US district judge said federal prosecutors' May 18 announcement that they would no longer pursue the case which charged Adani with securities fraud and wire fraud stemming from an alleged bribery scheme did not sufficiently explain their decision. Back home, in a further sign of the shift towards electric and away from fossil fuels, penetration of electric two-wheelers has crossed the 10% mark for the first time. Electric two-wheeler registrations hit about 162,000 already in June, which is 10.3% of the overall 1.56 million or close to 1.6 million registrations of all two-wheelers, according to a Business Standard report quoting official data.

Now, that penetration was about 7.3% a year ago and a little over 9% in May. Meanwhile, just to return to monsoons, the India Meteorological Department is saying that monsoons are advancing in India, though of course not in Mumbai, where it is presently absconding.

Why are Laptop and Phone prices going up?

Fortune magazine is calling it the Ramageddon or the unprecedented rise in RAM or memory chip prices, which in turn has sent prices of various products soaring.

Last week, Apple raised prices of almost all its product lineup excepting mobile phones by over 20%. Microsoft said it would increase the price of its Xbox game console by $100-$150 and also warned of some configuration changes which is happening elsewhere too. So, manufacturers are either raising prices or reducing the amount of memory in their products including laptops.

Apple's price rise, as Fortune points out, is instructive because the company was forced to raise prices despite its famous supply chain prowess and massive purchasing power. Apple's move has sent shockwaves and the Nasdaq fell about 1.4% the day Apple announced those price rises, that was Thursday, as investors worried about the implications of these rising prices. Fortune also quoted an IDC analyst saying that the storm isn't over yet and this is just the beginning, referring to the memory crunch.

So, what lies ahead? I reached out to Neil Shah, Vice President at Counterpoint Research, who a few weeks ago forecasted some of these price rises and provided the backdrop on why this was happening, which is really the insane race to set up data centres, which is hogging all the memory chip production in the world. I began by asking him how long did the big companies hold on to prices, which also suggests or would suggest how prices could go in coming days.

INTERVIEW TRANSCRIPT

Neil Shah: So if you look at the consumer side of things, as we discussed last time, it's all because of the AI structure built out. And if you look DRAM particularly, it has been growing since almost, I would say, Q4 last year. So Q4 last year, it doubled compared to the Q1 2025.

Since then, it has grown almost 10x compared to the Q1 2025. So in just 18 months, the price of DRAM has gone 10x. Let's say for hypothetically, Apple bought DRAM for $8.

Now it is for $80. Since Apple held it for last couple of quarters, I would say, I think there's a mix of strategy over here. One is it gave them a six-month window to capture as much market share because the competitors like Dell, HP, and all, they started increasing the price since February, March.

So Apple got that three, four months window to actually capture the market really fast. And the timing of it was really good because they launched MacBook Neo at a very aggressive price point. And actually everyone was going in a different direction where a $500 laptop became an $800 laptop.

And Apple came up with first time a $500 laptop. So it held it on for a couple of quarters. And now I think it is passing on to consumers because the memory pricing is increasing every quarter now, quarter and quarter.

Actually, it's double digit increasing.

Govindraj Ethiraj: Right. So if let's say prices of some of the MacBooks have gone up by let's say $100. Does that mean that is the actual cost of the memory chip that has or chip or chips that has gone up in that laptop or product?

Or is Apple also using this to maybe raise prices and absorb some of the losses? That's a good question.

Neil Shah: So traditionally, Apple has monetised its block materials with a specific gross margins and operating margins targets, right? So if you look at the gross margins, so even if they hypothetically bought a DRAM configuration for $8, it used to still sell the incremental ones for a $100 arbitrage, $100 gap. If you bought a 8GB to 56GB configuration and you go for 8GB and 512GB, then it used to monitor the NAND upgrade, but the DRAM still remained the same.

So it's still got a $100 gross margin benefit. So what Apple has done is slightly absorbed the gross margins for the last couple of quarters, and it had luxury to do so because their services business has become the second largest business, right? And it is very, very high gross margins.

So it had some cushion to eat up that. So the corporate gross margin was still at the overall high level. But now I think it is going beyond what it can do and to maintain that level.

Otherwise, investors will start punishing. And already you saw stocks slightly tumble, but actually it might again come up because Apple will maintain the gross margins for the next earning calls.

Govindraj Ethiraj: Right. What else do we use DRAMs for which might see prices being hit and including in a country like India or in India? And the supplement to that is where could prices go given the rate at which prices of the final products go, given the rate at which they're rising as you pointed out?

Neil Shah: Yeah. So DRAM is fundamental of wherever particular electronic device has a processor, a high level processor, not an MCU, but more of a Cortex ARM class CPU or SoC where it requires graphics or some kind of processing, AI processing and so forth. So all those products has a DRAM, only the configurations are different.

So I would say the lowest would be the smartwatch. The smartwatch has a DRAM, but it is maybe 1GB DRAM. But as you go up to tablets, you go to even the cameras, security cameras, right?

Those have lower configurations, but tablets, phones, PCs, all have more than 4GB these days DRAM.

Govindraj Ethiraj: And you're saying CCTV cameras also have some memory and therefore that's another product that we could see prices go up. Absolutely.

Neil Shah: So CCTV routers, even nowadays refrigerators have DRAM because now they have big screens and it has because many of these refrigerators or like smart TVs, they have advanced graphics now. They're doing OTT. It's no longer just a display.

All of these require different level of control and everyone will face the bill of materials crunch, or everyone is facing rather.

Govindraj Ethiraj: And that includes TVs as well. So phones are the one area that seem to have been relatively untouched because I'm assuming it's a larger market and the scale is much bigger. What's your sense there?

When could we see or might we see prices go up for Apple or other big companies like Samsung and some of the Chinese majors?

Neil Shah: So if you look at most of the Android vendors, non-Apple vendors have started increasing the price by 8 to 10% starting January actually. So January, February onwards they had inventory and after that the inventory, the old inventory was almost gone. So after that they have started increasing 8 to 10%, but they have not increased at the level of what I would say the overall growth has been for the memory DRAMs pricing because they can be creative with the bill of materials.

They can go down from a Sony camera to maybe a Taiwanese or Chinese version of the camera, right? So they can save like that $15 or something like that. So they have been very creative with the bill of materials.

I would say it is still increasing and what is happening with smartphone is the market is moving so fast. Every nine, 10 months you have a new phone. So when they launch a new phone, they will launch it with a higher pricing.

Only issue is with the older inventory. How much do you raise in terms of pricing? So for Apple also, I think when they launch the new iPhones, potentially looking at the trend, they'll have to increase the base pricing of the models to absorb the cost.

Govindraj Ethiraj: Right. And in past situations, and we touched upon this in the last time we spoke as well, when there is so much demand pressure, supply tends to catch up or there is obviously investment in more capacity and so on. It appears that the gap right now is so huge that even if there are big investments in supply, it's not going to catch up.

What's your sense on some balance or rationality returning to the market?

Neil Shah: The supply crunch is going to remain at least 2030. At least what we can see from the build out and committed hundreds of billions of dollars or $1 trillion infrastructure which is being rolled out. And every day we are starting to use more and more AI.

And now we're moving to agentic AI, which will have so many agents consuming tokens. So the compute demand is insatiable and it's short in supply, even memory along with that. So these top three vendors, Micron, SKINX, Samsung, they are building out capacity, but the overall capacity industrial level, it's going to increase just by 20% year on year in terms of DRAM bit produced.

But the demand is like 40 to 50% or even more. I don't think we'll ever reach a level, at least in near to midterm, where supply will be higher than demand. And I think this will be a new normal, the what they call a chip inflation, it will remain.

You look at the inflation, when you increase the price of a burger post pandemic, they have not scaled it back. I'm in US and Chipotle, I use that Mexican chain because I'm vegetarian. So that used to cost before pandemic, like $10, one.

And now it's like $25. It has gone up, it is not going to go back to $10 or $15 ever. So I think this is a new normal, because AI growth is going to happen.

Once we start using AI, once you use StartChat, GPT or Cloud for coding, you don't go back and do manual coding.

Govindraj Ethiraj: Right. I know you're joining us from New York, you're attending an analyst conference, or you've been attending an analyst conference for a chip company. So what's the mood like?

And what are some of the takeaways that you can share with us?

Neil Shah: Yeah, so I was here to attend Qualcomm's investor day. Qualcomm made a big splash by entering into data centre market. So they're normally known for chips and components for smartphones, PCs, wearables, IoT, and so forth, that was more of edge play.

Now they're going into data centre because they're quite like three or four different companies over the years to build that entire solution for data centre, where they can build inference racks to compete with Nvidia, AMD, and so forth. So the mood was pretty good in a sense where now the executive team and everyone says, okay, now investors will give us some thought because most of the investment and investors are bullish about all those companies which are present in AI infrastructure race, right? You see Nvidia, everyone's at a trillion dollar company, even all the three memory chip guys are trillion dollar in market cap.

So if you have that data centre play, then the investors reward you. And that is what Qualcomm is expecting as they start seeing the revenue growth, which they're seeing from almost $300 million this year to almost $15 billion annual run rate by 2029, which puts up their earnings overall revenue to almost $60 to $80 billion a year. So that is what they're bullish about.

And that was what the strategy was they're unveiling. And so the big question for them, everyone was asking is, okay, you have all these compute solution for AI rack, but are you going to procure enough memory? Because Nvidia is the biggest one who is bundling memory with their GPUs, right?

With HBM and similarly AMD. And so you have to secure memory, you have to secure supply TSMC to produce those chips as well, which is TSMC also been raising prices for chip. It's not about memory pricing, but TSMC is also raising the price by 5 to 10%, almost every six months now.

It used to do every year. So everything is getting expensive. Essentially, the biggest brunt is happening for non data centre industry segments.

Govindraj Ethiraj: Right. And I think we have to now get ready for a prolonged period of chipflation as you called it on that note. Neal, thank you so much and have a good rest of the trip.

Neil Shah: Thank you very much. Nice talking to you.

How can AI help with Hiring?

With all that data centre expansion, there is obviously a lot of artificial intelligence utility, including helping companies hire more effectively, particularly given the large volume of applications for jobs in countries like India.

But how effective is it and what are the underlying trends telling us, not just about the role of AI, but the nature of job applications and more importantly, the job market itself. I spoke with Jayant Neelakanta, founder of Equip, a Bangalore-based AI hiring platform that works with some 500 companies. Neelakanta has a PhD in theoretical physics from Syracuse University and returned to India a few years ago to build this company out and I began by asking him how he was seeing the latest hiring trends.

INTERVIEW TRANSCRIPT

Jayanth Neelakanta: So at Equip, we help companies shortlist candidates faster. So typically when you publish a job, we get a lot of applicants, and a lot of the work right now is fairly manual. So you go through the CVs manually, you reach out to the candidates manually, you probably have to do a lot of interviews, four, five rounds of interviews, right?

And at the end, the candidate may not even join. So there's a lot of steps, several stages, and we're automating each of the stages. So we have an AI resume screener.

So the AI is going to look at the CVs and then judge, at least rank the CVs. And then we have AI interviews. So instead of a human talking to the candidate, we have the AI do that for them.

We have skill assessments. So the AI is going to generate questions and check how skilled the candidate is at certain aspects, right? The idea is a lot of this early stage training can be automated away so that companies spend most of their time on the candidates who are the best fit.

And then there is that human element that will convince them to join the company. So we are automating a lot of the early stage screening stages.

Govindraj Ethiraj: And give us a sense of scale for, let's say, any company that you work with. The name does not matter, but what's the kind of numbers that you're looking at and what gets eliminated, what gets retained?

Jayanth Neelakanta: Let's say the company has a thousand members, strong workforce. It posts a job, it'll typically get about a thousand applicants. And these applicants, you will then screen their resumes.

The AI will screen the resumes, bring that number down to say 200. These 200 will then go via a skill assessment test. Could be a coding test, communication skills test.

Bring that number down to say 50. 50 will then take the AI interview. That number reduces to say 20, 25.

And these candidates are then, you know, human interviewed. Maybe the top five get the offer letter and a couple of them may join. So starting from a thousand, you come down to about two people joining the company.

Govindraj Ethiraj: And you're saying the AI will really bring it down to 20 to 25 from thousand.

Jayanth Neelakanta: Yes, absolutely. Think of it as almost like a sales pitch, right? You have a lot of people reaching out to multiple companies.

And then at each stage, you know whom to bring in, right? So you can't start with the AI interview, which is the most effective tool right at the top, because then every candidate will not be willing to give the entire AI interview because they're applying to multiple roles. So you also want to bring in the right tool at the right time.

And it increasingly becomes more of an involvement from the candidate. At the top of the funnel, you have something that's very easy for the candidate. And as and when they realise that, okay, the company is actually shortlisting me, you can get them to do more and more from their end.

And that's why it's structured that way.

Govindraj Ethiraj: Right. Now I'll come to the use of AI even on the other side. But what is this telling you about the job market itself?

I mean, is thousand people applying for one job par for the course? Is it unusual? Or what are the changes that are happening in that side?

Jayanth Neelakanta: No, we're actually seeing an increase in the number of applicants. I think in general, I would say the trend is companies are hiring lesser. You can attribute it to AI, you can attribute it to the market, like macroeconomy.

But in general, we see companies hiring lesser, especially for sort of early stage careers, right? Which is what Equip helps with. Zero to seven years of experience, that kind of a range.

So there's less of a demand, which means obviously you're seeing a spike in the number of applicants. Also, it is these job platforms that make it very easy for you to apply. So LinkedIn, for example, has an option of easy apply, where it already has all your data.

You can just like three seconds, you can finish applying to one company. Right. So they're making it so easy.

It's very indiscriminate, right? So can they just apply to multiple jobs? So tools like these are relevant for that.

And that's why I think you're seeing a lot of applicants.

Govindraj Ethiraj: And people are also using AI to spruce up their resumes. And therefore, I guess it becomes tougher and tougher to distinguish. So what are you doing on the other side to break through?

Jayanth Neelakanta: So you're absolutely right. So a lot of tools just look at the job description and then create the relevant CV for you. What we help with is the company or the recruiter tells us specifically what they're looking for in the CV.

And that is not published. So there's a job description, which talks about the role. But for example, recruiter may say, this candidate should not have switched too many roles, right?

They want, they look at loyalty. This is not public to the candidate. They may say, I prefer skills over education.

So the pedigree is not that important. Or maybe they'll say, I'm very particular. They must have worked in the same industry.

So we let them customise this. And this is what a human recruiter anyway does, right? The hiring manager will tell the recruiter, look out for these attributes.

So those attributes, the recruiter shares, then they create the job post on our platform. And like I said, it's not visible to the candidate. So they apply, and then we filter out based on these objective sort of parameters.

And so we are able to achieve, I think, a higher job fit score.

Govindraj Ethiraj: And if you were to, let's say, look at a median of the kind of asks your top clients have today, what would it be? I mean, in terms of going beyond the resume.

Jayanth Neelakanta: So we've seen, interestingly, if it's sort of a tech kind of a role, it's much more skill and pedigree based. So they want them to have come from a good college. They want them to have those skills.

And when it comes to non-tech, so let's say looking at sales, they're looking at support, those kinds of roles is very industry focused. So they don't care as much about the pedigree or what skills they have. If they've already worked in real estate, if they've already worked in fintech, we see that they care a lot more about that.

Because if you're writing code, it doesn't matter whether you're doing for a fintech or a real estate firm. But if you're selling, obviously it does, right? That is one distinction we see.

But one good thing we noticed is pedigree is not as important now. What we also do is we let them pick the three most important factors. So otherwise they will check everything.

So they only pick the three and they prioritise typically skills or industry, work experience, those kinds of things, not as much pedigree, which is sort of a welcome change.

Govindraj Ethiraj: And how does the AI interview work and what kind of training or customisation goes into that?

Jayanth Neelakanta: From the recruiter's perspective, the input is a job description. So they share the job description, which should have things like what skills should be assessed, what is the difficulty level. Typically the experience level looking for maps to the difficulty level.

And the AI passes the job description and then it will create a set of topics on which to test the candidate. And for each topic, the recruiter can decide how many follow-up questions need to be asked, right? So that's the configuration from the recruiter's perspective.

A lot of the work is actually done by the model itself. I want to be honest, it's not like you've trained too much, right? What is important is prompting and a lot of these guardrails that you need to have.

For example, we need to be very clear that the interview cannot go off script, right? The candidate can't start talking about their girlfriend or whatever, right? So you need to keep that within the framework.

And also we allow the recruiter to customise their instructions, right? We have customers across. In India, for example, they'll say, okay, this should be in an accent.

Speak slowly, right? So they allow to customise that and that gets transformed into the user experience for the candidate. But to be very honest, off the shelf models do very, very well.

As long as you have the right kind of prompting. For example, if the recruiter says it needs to be at three years of experience, this is the kind of question you need to ask. So if you give it those sort of guidelines, it does a fabulous job of sticking to it.

Govindraj Ethiraj: Right, and this is an evolving space. So what are the kind of changes that you're making all the time? So let's say if you were to look at the last three months, what kind of changes you would have made to your models at any of these stages that you just described in order to keep pace with either the technology or what you've detected in terms of how people are responding to it?

Jayanth Neelakanta: One big change we did is, earlier it was just a conversation. Right now we allow them to whiteboard. So the AI could earlier only listen to the voice, right?

Now they can type, they can write code, they can draw. So these also become inputs. Now think of a human interviewing, right?

They will say, okay, open up a Google Doc and type something or draw something, maybe a diagram. So we are now allowing, it's called a multimodal kind of application, right? So that's one thing that people have asked and we've done.

The other is what I just mentioned. The off-the-shelf model was, it was a US female voice, right? They want to be able to customise that because we saw a lot of people start using this for sales, like feed sales jobs, right?

We charge very low. We charge a couple of dollars a candidate. And so even for these kinds of roles, it's quite relevant.

So to be able to customise the voice and the rate of delivery of the speech, that is something that people wanted. Basically we've scaffolded the prompting and we've made it much more specific to the difficulty level they wanted. Otherwise the AI would not really know, right?

So it would not be consistent. So if I say medium level of difficulty, two different candidates may get slightly different levels of difficulty. We've added enough prompting now that it will not be the same question, but the difficulty level is similar.

And so it becomes fairly standardised. Those are some of the changes we've implemented.

Govindraj Ethiraj: Right, you mentioned real estate and of course IT and so on. So are there industries where let's say AI or AI-led tools like these will still find it difficult to assess candidates or do you feel that it's now achieving a high level of competence across?

Jayanth Neelakanta: I still think that this is much more for the initial levels of screening. So it's more for objective skills, right? So let's pick sales.

A lot of, especially the most senior sales force, these people are doing relationship building. We can help you evaluate industry knowledge, domain expertise, communication skills, but a lot of it is how empathetic are they? And that is really difficult to measure by AI, right?

So those things which are very human traits, anything that's very creative as well. So let's say you're a designer or you're writing marketing copy. I mean, one thing with AI is it praises you a lot, right?

It's always saying, whatever you say, this is the best thing in the world. So if I'm going to judge marketing copy, I don't think it'd be very easy for the AI to judge. And anything that is also very local, right?

Let's say, especially Karnataka, right? Maybe there's some topical thing that's come up. I need some skills that leverage that.

I think those kinds of roles, I would say, you know, AI can't do, but anything very objective, even if it's delivered in a subjective format, right? When you're speaking, that I think it'll excel at, but I'm sure there's some skills that it'll find it difficult to do.

Govindraj Ethiraj: All right, Jayanth, thank you so much for joining me.

Jayanth Neelakanta: Yeah, thank you very much.

Has the Australian Social Media Ban for Teens Been Effective So far?

Australia said on Saturday it would double the maximum penalty it can impose on tech firms found to have failed to uphold its six-month-old social media ban for children, as evidence mounts that the ban has not had much effect on teen use, according to a CNBC report. Under the changes, the maximum penalty for systematic failures to uphold the ban goes up to 68 million dollars, which is about 99 million dollars in Australian dollars from about 49.5 Australian million dollars. The government also reiterated that it's actively investigating possible non-compliance by five platforms, that's Meta's Instagram and Facebook, Google's YouTube, Snap, Snapchat and TikTok.

Britain this month said it planned restrictions that go further with gaming and live streaming platforms also affected. Australia said that since the ban has been put in place, more than 5 million under-16 accounts have been deactivated or restricted, though the CNBC report added that numerous studies have shown that age assurance mechanisms like selfies put in place by tech companies are being easily circumvented by children and that in many cases children have never been asked to prove their age.