[SnaazTee]

So, last week, The-Ken had an insightful podcast and newsletter (Check it out here on The-Ken.com).

Software engineering careers used to be a ladder. You studied for four years, got a job as a fresher, and could virtually take for granted a steady career filled with learning opportunities, salary hikes, and role promotions.

Image Source

The-Ken has also quoted a New York Times article

I have a pretty good sense how fast the progress that students should make in a semester should be,” he said. “In 14 years, I’ve never seen students make the kind of progress that they made this year.”
And he knew exactly why that was the case. For the first time, Mr. Ammirati had encouraged his students to use generative artificial intelligence as part of their process — “think of generative A.I as your co-founder,” he recalled telling them.

Many AI chatbots are fully capable of writing code now. So your technical co-founder could be an AI?
Meanwhile, AI tools are making some of the best engineers and programmers even better at what they do. Companies and investors are realising that perhaps they don’t need as many people to write or maintain programs as they did earlier. The real reason hundreds of billions of dollars are being spent on AI is because everyone is assuming a disruptive productivity jump at some point. Allowing one person to do the work, of say, four, or 10.
Where does that leave engineers? Are we staring at the end of the golden era for engineers?

Image Source

From the podcast excerpt, I found this particular exchange very interesting:

Rohin: We’ve been reading again and again from journalists, CEOs, venture capital firms, etc., that the reason why there’s so much money going into AI is because of the payoff that essentially comes from potentially dramatic shifts in productivity. Every CEO who’s investing heavily into AI is hoping that what they can achieve with, let’s say 100 engineers, can be achieved with 20 engineers.

Now, it means that those 20 engineers are becoming more productive, like with their workflows, etc. But it also means that 80 engineers…

Amod: See, companies that have 100 engineers, most of them should have been running with at most 30 engineers.

*Laughter*

Amod: Honest truth. The engineering productivity of pre-2010 is completely different from the engineering productivity of today.

Image Source

Kailash: I’m just calling out how things actually happen in practice. And something that I strongly disagree with is where people are hired not for their intelligence, but for their ability to translate.

In this kind of an environment, most of the technical people you are hiring are your “programming language feeders”.

These are people who take the logic from you. They do not determine the logic.

And they will translate it into a language that the computer understands. AI can naturally do that, which is why non-technical people who get exposed to the coding abilities of an LLM get so super excited. You know, like, “oh, I have found my technical co-founder”. If your technical co-founder was doing this, you were anyway screwed.

[GTO]

Thread moved from the Assembly Line to Shifting Gears. Thanks for sharing!

[madhukarmohan]

It is not just software engineers; many professions will be severely affected by AI. I often get asked what to do after matriculation or graduation, or which courses to pursue. Unfortunately, I find it increasingly difficult to provide a clear answer. The reality is what's in demand today may become obsolete in just a few years. Technology is evolving at an unprecedented pace, and it is challenging to predict where the future will take us.

[Jeroen]

I am probably going to get some stick over this, but I never understood this software engineer hype in in the first place.

I think many software engineering roles were overhyped and grossly over paid to start with. If anything the retention rates of your typical software house are abysmal. If you can afford to shed that many people you are providing a pretty common type of service I would think?

Another factor is what kind of people are attracted by fast career path, promotions and salary hikes? I don’t want to put it down, but there are plenty of people who choose their career based on very different parameters. E.g. my wife is a Speech and Language therapist. Took a five year Master at Edinburgh Univeristy. The pay is decent but nothing like a software engineer. But she absolutely loves her job, would not do anything else. She will do this job if she had to pay herself. Similar with for instance teachers, nurses and so on. People who make have a direct and substantial impact on individuals lives.

I for one, think it is high time for those software engineers to descend to planet earth.

In my world we had software engineers. It used to take a master from a technical university and at least 5 years of supervised experience before you would be allowed to touch code yourself.

A lot of what is called programming and coding today, is no more then relative simple configuration. Yes, it does take skill and some expertise, but it is a available in abundance! No need to pay through the nose for it.

Jeroen

[HillMan]

It is a long way before software engineers will be impacted. Yes, no doubt Generative Ai does help improve efficiency but it is no where close to build a full-fledged application on its own.

It will help junior developers write better code faster, write test scripts and documentation.

We will see ground breaking impact on migration and modernisation projects. I can say this because myself lead such initiatives building GenAi based tools for any to any code migration for enterprises. Monolith to micro service and language to language too.

So Gen Ai has its strength in summarising content and research. Some models can also interact in human like language response.
We have crunched millions of pages of data for medical research reducing years of manual effort to months. So such jobs and L1, L2 support work are majority at stake.

So what is the future, that's what will be a good topic to read here. AI is here to stay just like Google disrupted search space. But there are ways to Google and get the right stuff in a single query. Gen ai is also a tool, those who know how to use it right have great opportunity.

[SmartCat]

One needs to keep an eye on employee count trends of major Indian software companies. Historically, the number of employees has been growing every year along with revenues. Even 2008 Great Recession or 2020 Covid Pandemic did not stop the trend.



But TCS saw a dip for the first time (FY24). We need to see if it is a one-off dip or flat trend/ continuous downtrend.

[standardherald]

Been in the industry for 23 years mostly in product development and while generative AI is over hyped and many tall claims are being made, it is very good for certain types of software engineering tasks. These tasks are at the lower end of the value chain and were done by new hires from college or by not so great performers. Due to this I can see two trends already playing out

#1 A very noticeable reduction in hiring of fresh grads. What they do can be done by the gen ai.

#2 Mediocre performers are now actively weeded out and there is no replacement, instead teams are asked to increase their productivity.

However I doubt the profession will disappear because gen AI doesn’t cover all the aspects of software engineering and also generates buggy and suboptimal code which needs to be checked and modified. Ingenious engineers will multiply their productivity using the new tools and become even more valuable.

[Samurai]

Originally Posted by Jeroen

I think many software engineering roles were overhyped and grossly over paid to start with.
.
.
A lot of what is called programming and coding today, is no more then relative simple configuration. Yes, it does take skill and some expertise, but it is a available in abundance! No need to pay through the nose for it.

This has been true since the gold rush years of the 90s. I have mentioned it many times over the years. I'll just quote them instead of re-wording them like ChatGPT. All AI is doing is bringing some correction, some extra efficiency.

Here is one from 2008:

Originally Posted by Samurai

Generally most IT guys are overpaid, actually very few are underpaid, and it is them who finance the rest.

Ok, hold on to your horses, before exploding on me, hear me out. As somebody who runs a software company, I can explain the dynamics.

You may have heard this statement, 80% of the world is owned by 20% of the people and the remaining 80% people own the remaining 20%. These percentages can vary, but the idea is clear.

The IT companies are heavily depended on the skill level of their employees, more so than any other industry. The same 80:20 (it can vary, even 90:10 in product comapnies) idea works in IT companies. When a company hires 100 people, it is the top 10 performers from those 100 who really make the difference and bring success to the company. The remaining 90 people matter, but their contribution is supplementary and they can be easily replaced. I am not talking senior management vs junior staff here. Consider all 100 to be of same seniority. If one goes purely by contribution, the top 10 deserves 10 times or more compensation when compared to the the rest. But that never happens. At the most the top most performer may get twice as the worst performer at the same seniority level. If you average it out, most people among this 100 are actually overpaid compared to their contribution, and the few top performers are underpaid compared to their contribution.

No company would really mind paying 2X to a top performer, the problem starts when an employee thinks he is part of the top few and the company doesn't. Who is right, it depends on individual cases. Sometimes, the company is underestimating the employee's real value and sometimes the employee is overestimating himself/herself. But in the current IT job scene I can definitely say that most employees overestimate their worth. The companies still pay because it is the industry standards. All the while they are secretly hoping that a few may turn out to be severly underpaid, in other words earn 10-100 times of what they are paid.

From 2013:

Originally Posted by Samurai

IT is one field where gauging a employee's worth is extremely difficult. In this, it is very different from other fields.

"A great lathe operator commands several times the wage of an average lathe operator, but a great writer of software code is worth 10,000 times the price of an average software writer." - Bill Gates

For better understanding of this, check out these links:

http://programmers.stackexchange.com...old-difference

Originally Posted by Samurai

The only people who can gauge the worth of an employee are the ones who know technology and business side of that company. Such people can be found only in small companies. In big companies, technology people never get exposed to business side, and business people never get exposed to technology. No matter how much the former can convince the latter about a candidates worth, the latter will be skeptical. That is because the latter (some MBA type) will value business knowledge over technical knowledge. Now bring in the HR guys who lack both business/technical knowledge, and put them in charge of hiring and fixing remuneration. They will take input from both technical and business side, but will listen more to the latter. And they will bring in their own set of filters like academic marks, age, college, industry standards and what not. Ultimately, nobody really gets what they deserve.

However, highly skilled software engineers (top 20%) don't need to be afraid. An average engineer armed with AI tools won't be able to outperform them, since AI can't innovate unlike good human engineers.

[itisravi]

We must also keep in perspective the cost involved in running a generative AI. The amount of CapEx and OpEx for maintaining these systems is insanely high and right now, big tech is absorbing it and giving us access to these tools. Unless the technology advances significantly to make it cheaper than hiring engineers, I don't think a replacement is happening. It is a business after all and no AI infra provider will operate in loss making mode forever.

[mack.mehul]

I don't think so. AI can never replace human intelligence. It can only supplement human labor. If AI is left to its own to do the work, results can be disastrous which these big corporations will realize soon enough.
For example ADAS systems on car these days is also a sort of AI. However one can't close his eyes and let ADAS driver the car because invariably it been shown in many road tests THAT ADAS is not foolproof, its real use is in conjunction with human mind.

[ritedhawan]

I hope AI doesnt replace software developers , coders, et al. Otherwise, my efforts in becoming a (frontend+backend developer) will be undone. For someone who has experience in IT only in a specific L2 role(Service Desk) and is in bad financial condition--I am totally dependent upon my father's monthly grants to run my house. For someone like me who made a stupid mistake two years ago, leaving my job. Since, that time have not been able to garner even a single job, even after filling innumerable applications, only to get an impression from interviewers that my age doesnt correspond to my job profile; itself forced me to switch profile and head for software development, and aim for frontend+backend developer by end 2024.

[Rajeevraj]

The thread seems to have deviated into AI eating up jobs. Keeping that aside, I think from the perspective of the point mentioned below, the golden age is over.

Software engineering careers used to be a ladder. You studied for four years, got a job as a fresher, and could virtually take for granted a steady career filled with learning opportunities, salary hikes, and role promotions.

The golden age was the period where you could do any degree in any stream, land yourself a good IT job. Progress to an extent depended on your skills and performance, but even a very average performer could work their way up to a point, get yearly hikes and maybe go a little further with some lateral jumps.

I think this phase is done, it is still a lucrative field with plenty of opportunities, but I see it has plateaued due to various reasons and I expect it to remain so. A person will not be able to float along easily. A simple example is how CTS is able to open fresher hiring at the same packages they were offering 20 years ago. There was massive outrage around it, but I am pretty sure there are 1000s of freshers ready to take it up.

[Samurai]

I would like quote something VN pointed out some time ago:

Originally Posted by V.Narayan

Some years ago we decided to develop our own proprietary software to sell to our aviation world customers as a subscription. Now better known as SaaS (Software as a Service). We gave up after some painful false starts and coping with attrition amongst the vendors team and the nightmare of trying to get the s/w youngsters to learn enough about a very niche area to write the code. We finally set up a our own small s/w development unit in Ohio, hired folks who knew s/w and enough of aviation to understand the customer (us). All of them are in the 35 to 55 age bracket, attrition has been almost zero over 6 years and they own the product as their own and only 1 of the 9 is a graduate engineer.

This is what AI will have hard time replacing. I know this because I run a setup like this, creating products/service in a niche area where domain knowledge is very key. It takes many years in the domain to develop the deep knowledge and skills to work in this mode, but most Indian techies rarely stay in one place to gain this knowledge. If one only has shallow knowledge in 10 different areas, AI can easily replace such a person.

[PB65]

You can have AI write a code for you. But what we need to understand is that it's an LLM. It's got no brain or thought process of its own, and it misses out the intricacies that only a human can touch and pick.

Remember Swiggy and Zomato in their early days? Same features, same functionalities. But they co-existed on application stores because the code base was different. To escape the plagiarism concerns you need to have a certain degree of variance. Same in case of websites. You cannot just flood the search engines and application stores with AI generated code. It's straightaway going to be rejected.

AI will enhance productivity by giving you the basic framework which will have to be customised as per requirements. But yes, all jobs with repetitive patterns will be replaced by trainable AI models. Knowing humans, we'll find a way around that too.

[androdev]

There will be disruption for sure. I believe the good engineers will thrive and the not so good ones will really struggle.

1. Decision making or design choices: AI will make it very easy to present you with options. You are saved from the donkey work of doing the research but the critical decision making has to be done by you. As with a senior doctor or a CEO, one needs to take the right decision based on inadequate and ambiguous information. You can engage your staff (AI in this case) to do the low level work but you still need to take the important call.

2. Implementation and deployment: AI can again do the grunt work but tying up the loose ends, integration touch points, troubleshooting and debugging, etc. needs to be done by you. If you are not a good engineer, you will waste more time doing this final 5-10% than actually doing the whole 100% by yourself. A good engineer would know which parts are better generated by AI and which parts need to be hand-crafted to optimise the overall productivity. Remember, the difference between useless pile of garbage vs working software could be just one wrong instruction to the computer. This is very different from a computer generated art work vs a human created one.

3. Impact on employment: Employment will shrink if economic activity doesn't increase after introduction of superior technology. I don't see this happening in software industry. It's an industry that doesn't need physical resources to produce and there is no limit on how much software the world can consume. So I expect the economic activity will continue to increase but the engineers are likely to be of a different skills set (lot more analytical than garden-variety code writer). However all this massive software creation drives automation in other industries which will eliminate a LOT of jobs. Say 1 software engineer position is created to eliminate 100 "TV commercial video creation" jobs. Even though there is a lot of unrealistic hype around AI taking control of humanity and being dangerous like Matrix movie - the immediate danger or reward depending on whether you are an employee or investor, is pervasive automation and elimination of lot of jobs. Hopefully the new economy of automation itself will produce new types of jobs - difficult to predict.