[padrino]

Writing this as a part of the [Weekend Entry] competition the Legends group is running as I've been meaning to write it for a while, and have not made the time for it. It won't win, but it's an excuse to get my ass into gear and write up this little essay on my thoughts on the subject.


Introduction
So, in my industry there has been a lot of uncertainty about the use of AI - specifically around the job security of software engineers and the concerns regarding potentially going into a career in software development when engineers are seeming ripe for replacement. Well, in this post I'm gong to exlpain why I think AI will not fully replace flesh and blood engineers, and more over, how we as engineers can use this technology to our advantage.

Why we're not getting replaced... yet.
It's true, AI and LLMs have made coding and building software more accessible. For the first time ever, it is actually feasable for anyone to have an idea for a script, or a small program/project and make it a reality with relatively little technical savvy. You want an app that can work out your tax based on your pay? You can do it! you want to generate a forula for a spreadsheet in work that's been pissing you off? you want to write JS script for filling out long forms in the browser to stop you having to do it yourself?.... it's actually possible! But here's the thing, using AI in this scenario is a great way to get the job done, but if you're not aware of what the code is actually doing - there's no way of verifying what is generated is actually correct. And if we scale that out to large enterprise applications, you are going to need trained and capable software engineers to review code. for one thing, to ensure it's actually doing what we want and will not producee any odd edge acase scanarios, adn two to ensurew it's not doing anything else... you know.... like contacting an external server, or copying local files...

How can we as engineers use this tool?
You need to be using these tool. There's no 2 ways about it. If you don't you're being left behind. And there's no reason not to - we can utilise this amazing technology to take the grunt work out of a lot of the daily coding tasks. Need a script that takes a CSV and formats it into JSON? Get AI to do it. Then run it, and check the results. You'll save yourself a bunch of time. Need a method that takes a JSON object and commits it to a database? Get the AI to write it . Neeed to google something or check strack oevrflow? AI will do it, and quicker. Have a SQL script that's just not working and you can't figure out why? Give it to the AI and get it to tell you. Need to document a bunch of legacy code? Give it to the LLM and get it to tell you what the code does in paymans terms. You don't need to be getting it to do everything for you - but utilise it to make your life easier. Every other engineer in your company will be doing the same. It's worth mentioning though, that if you're working for an organisation and are using AI on your product work, you need explicit permission from engineering management, and a corporate license. and for the love of god you better not be putting PII and customer data in there - you'll get canned real soon if you do.

What does the future of software engineering look like in the age of AI?
I had a chat to my CTO about this and this is what he said: AI won't replace a skilled engineer - but it might replace a traditional Junior Engineer. And I think he's right on the money. With a Junior, you set them a task, they do it, and then you go over their work, point out improvements to be made, and review it thoroughly before you put it live. This is where we are with AI. Does that mean that we no longer need to talent entering the profession? Absolutely not, we will still neeed human engineers to learn and utilise these tools - but it does mean that the job market will be smaller and more competative so that is something to bare in mind.

Will we ever be able to truly trust an AI without a human check
For a blog? yeah, sure. For the finance, aviation, and medical fields? I wouldn't count on it. Is that becuase the AI models we're using aren't as good as a human? Well, no explicitly. But they hallucinate, and are vulnerable to bullshit training data. And, if I may quote IBM: "a computer should not be making bsiness decisions, becasue it can not be held responsible" (or soemthing like that...). Anyway, I think that goes for the use of AI too. You may disagree, but that's my thoughts.

Let me hear your point of view!
The above is just my opinion as someone who has been in the industry a while and has seen the change to AI as a tool in the industry first hand. Maybe I'm just hopeful that the job I love won't be lost to an AI agent? I'd love to know what your thoughs are. Let's start a conversation!

[Lowkey]

Anyone who owns a SaaS company is and has already replaced the programmers / coders with the exception of human oversight monitoring + general maintenance.

It doesn't make any sense to absorb the cost of an employee from a business perspective when I can pair a self learning / diagnosing LLM that is constantly active + re-writing and adapting with a power-house gpt 4 or turbo model to automate and centralize workflow as well as increase speed and productivity for literally %90 or less of the cost of a traditional employee.

The other big reason coders + programmers will be replaced is because AI wont go on a forum and talk about being replaced. They'll find a way to adapt and survive or work with the next best thing to maintain relevancy and continue to stay optimized for future use.

Think about it from a business perspective where the goal is to make money.

Should my business:

A) Hire 2 employees with benefits, $80,000 salary each, limited time frame for project completion, data collection for the onboarding process, paperwork to file and tax codes + payroll and more to adhere to.

or

B) Pay $2,400 a year for the leading public power-house AI globally to instantly create and automate any idea or back-end portion of my business I can come up with a prompt for and let it sit and run with 0 paperwork aside from the expense of this service itself which literally can written off as a write off.

It just does not make sense for a company short or long term to not make the switch.

On the flip side the only thing I'll say that would still carry a demand is private / government grade projects that require niche and closed environment development teams. In which case this would be an exception. Even then... think 50 - 100 years down the road and be realistic.

[padrino]

Anyone who owns a SaaS company is and has already replaced the programmers / coders with the exception of human oversight monitoring + general maintenance.

It doesn't make any sense to absorb the cost of an employee from a business perspective when I can pair a self learning / diagnosing LLM that is constantly active + re-writing and adapting with a power-house gpt 4 or turbo model to automate and centralize workflow as well as increase speed and productivity for literally %90 or less of the cost of a traditional employee.

The other big reason coders + programmers will be replaced is because AI wont go on a forum and talk about being replaced. They'll find a way to adapt and survive or work with the next best thing to maintain relevancy and continue to stay optimized for future use.

Think about it from a business perspective where the goal is to make money.

Should my business:

A) Hire 2 employees with benefits, $80,000 salary each, limited time frame for project completion, data collection for the onboarding process, paperwork to file and tax codes + payroll and more to adhere to.

or

B) Pay $2,400 a year for the leading public power-house AI globally to instantly create and automate any idea or back-end portion of my business I can come up with a prompt for and let it sit and run with 0 paperwork aside from the expense of this service itself which literally can written off as a write off.

It just does not make sense for a company short or long term to not make the switch.

On the flip side the only thing I'll say that would still carry a demand is private / government grade projects that require niche and closed environment development teams. In which case this would be an exception. Even then... think 50 - 100 years down the road and be realistic.

Totally respect you input my dude - and you make some super valid points. I'm not surprised to see SaaS companies making the switch, but as you say that's with the exception of human oversight monitoring + general maintenance - and when it comes to reviewing code, maintining legacy systems, and conducting security audits on AI written code, that something that can only be achieved reliably (right now, anyway) by a human at the keys. Finance, Aviation, Medical, they're all going to be using AI but I don't see them being implemented without human checks and reviews - and you need trained engineers for that (even ones that go on forums talking about being replaced).

I stand by my belief that it will be a tool to be used and not a full replacement. I think it will reduce the need, but not eliminate human coders. I think there is the tendency to believe that the current LLMs we havea are never wrong, and we are too readily accepting what the output is as factual without checking it. I'm also keen to see what effect AI generated content on the web will have when it is scraped and ingested as training data. Seeing claims of AI eventually essentially cannibalising itself in that way but I've not researched it to make any conclusion on that.

[7]

I think its a crutial time for many people to realise they have relied on the comforts of technology as a whole, for far too long and the human brain hasnt been evolving on a global scale, the way it could have. People have become weak, complacent and have dopemined their heads with useless shit. But soon they will need to evolve or fail.

1) If your the type asking the questions and assigning workloads.. your in the right place

2) If you have been working with the right mentallity of optimising your workload using the tools evolution have provided. Your evolving with it... less concern

3) If you get given a task to do, take your time completing it, occassionally take some extra breaks and sickness days on mondays.... be concerned and maybe start looking for new ways to do nothing and earn money.

Its the cold reality of evolution. Sometimes it takes a hard fall to realise that.

[Online Sir]

I think its a crutial time for many people to realise they have relied on the comforts of technology as a whole, for far too long and the human brain hasnt been evolving on a global scale, the way it could have. People have become weak, complacent and have dopemined their heads with useless shit. But soon they will need to evolve or fail.

1) If your the type asking the questions and assigning workloads.. your in the right place

2) If you have been working with the right mentallity of optimising your workload using the tools evolution have provided. Your evolving with it... less concern

3) If you get given a task to do, take your time completing it, occassionally take some extra breaks and sickness days on mondays.... be concerned and maybe start looking for new ways to do nothing and earn money.

Its the cold reality of evolution. Sometimes it takes a hard fall to realise that.

Can 100% agree with this statement.

[padrino]

I think its a crutial time for many people to realise they have relied on the comforts of technology as a whole, for far too long and the human brain hasnt been evolving on a global scale, the way it could have. People have become weak, complacent and have dopemined their heads with useless shit. But soon they will need to evolve or fail.

1) If your the type asking the questions and assigning workloads.. your in the right place

2) If you have been working with the right mentallity of optimising your workload using the tools evolution have provided. Your evolving with it... less concern

3) If you get given a task to do, take your time completing it, occassionally take some extra breaks and sickness days on mondays.... be concerned and maybe start looking for new ways to do nothing and earn money.

Its the cold reality of evolution. Sometimes it takes a hard fall to realise that.

Totally agree with this point of view - it’s an evolution and we need to move with it. I don’t think engineers will be totally replaced - at least not yet. Like I said in the post I see it as a tool to work with and if engineers don’t take advantage of it - they’ll be left behind

[Oh_Well]

Engineers will never be "replaced" in the sense that they will be out of job.

You might had 50 companies with 10 developers each, and now with AI you might have 500 companies with 1 dev each.

What AI can do with a little help is magnificent.

Eventually everyone will turn into a Lead Dev and their team will be AI agents.

[padrino]

Engineers will never be "replaced" in the sense that they will be out of job.

You might had 50 companies with 10 developers each, and now with AI you might have 500 companies with 1 dev each.

What AI can do with a little help is magnificent.

Eventually everyone will turn into a Lead Dev and their team will be AI agents.

This is my thoughts exactly - we will still need engineers to lead, and will still need new talent entering the space. However, I suspect the competition for the jobs will be fierce as the supply will likely out-pace the demand.

[A_P_T]

This is a very timely article as I was just wondering about this. I’ve been specifically wondering how I can differentiate myself amongst all the coders/coder wannabes using generative AI. What value can I add? What can I do differently?

One of my closest friends had 10,000 + laid off at her large software organization. Essentially whole IT teams gutted for AI, including high performers according to my friend.

However, I agree with you - I think mostly lower level people will be replaced. The future will belong to the creative, I believe - the ones who can think outside the box an AI would.