Will AI take your job?

BrokenArrows

As a software engineer im certain that within 20 years the majority of engineers will be out of a job. There is already almost no need for junior engineers.
It will take much longer to get rid of highly specialised and senior roles but stuff like website design, standard app development for xyz will be easily automated.

As software becomes easier to develop and deploy via AI there however will be more jobs in the design side of things. Its easily for AI to generate code, but you still need to be able to accurately know what to ask for.

There are also likely to be more roles in physical manufacturing of products suited to AI. Robots, cars, VR experiences.

poop emoji

https://www.boards.ie/discussion/comment/123536891#Comment_123536891

Have you tried Cursor with its new Claude 4 model? Seems to be way ahead of current Q output

Mr. CooL ICE

https://www.boards.ie/discussion/comment/123536946#Comment_123536946

I haven't, but I've heard good things.

Been mainly sticking to AmazonQ for now as most of my current work involves migrating existing services to AWS, so it fits my needs. It was dodgy and slow up until a month or so ago but has massively improved.

techman1

follow the AI industry very closely and watch videos each day on the advances and trust me…..

60-80% of manual jobs are gone by 2028-2030.

Explain your reasoning the conventional wisdom is that AI is the biggest threat to desk based jobs not manual? Yes robotics got rid of alot of repetitive manual tasks in factories decades ago but how is AI such a threat to the tradesman having to climb onto a wet roof with tools, how does AI replace him?

Pa ElGrande

https://www.boards.ie/discussion/comment/123536873#Comment_123536873

Not disagreeing with you, that is possible when you have the right use case and dataset, even with that, the "hallucination" problem in output has not gone away. When you know what you are doing and are reasonably competent and knowledgeable on the topic you are querying, it really does improve your productivity (at a cost).

My experience is company CEOs are trying to sell they have an AI story to shareholders, that its going to deliver companies cost savings through better productivity and reduced overhead, they are prepared at present to throw money at it today. However, the people under them are then kludging the technology into their existing business processes and in a lot of cases making a b@lls of it. As far as I'm concerned when AI can be used to generate output that is accurate with an acceptable margin of error, grand, I'm all on board. Right now, I'm dealing with the fallout from the mis-application of the technology.

purplefields

https://www.boards.ie/discussion/comment/123536911#Comment_123536911

Also as a software engineer, I'd largely agree with this.

However, I would say much sooner that 20 years. Billions are being pumped into AI and it will get better and better very quickly.

Anyone that writes any kind of code that says AI hasn't already replaced large chunks of their work load are simply doing it wrong.

To answer the main question of the thread, if most or all of your day is sitting at a desk, then yes, AI will eventually replace you. This is a good thing because humans are not meant to sit at desks for years on end. I know this from all the health problems I'm now getting. Good riddance to deskwork!

Top professions to be replaced:

- Legal people of any description, including Judges

- Software developers (including me!)

- Office Admin

- Politicians

- Artists/graphic designers (Strangely enough. Would never have thought that 10 years go)

Expect friction when politicians start getting replaced.

halkar

https://www.boards.ie/discussion/comment/123536283#Comment_123536283

Governments will find a way to tax AI bots 😄

Apiarist

https://www.boards.ie/discussion/comment/123533606#Comment_123533606

Same here. My "software development" job is safe from the AI threat, as most of my time is spent not in writing software, but talking to customers, marketing and management, trying to understand and to define what needs to be done!

https://www.boards.ie/discussion/comment/123536253#Comment_123536253

Self driving cars are a step too far!

He could own 100 companies that failed at 100 products/services each….the only stat that matters as regards a measurement of his success….let me repeat….World's Richest!

https://www.boards.ie/discussion/comment/123536911#Comment_123536911

it will be a lot sooner than that….cannot keep up with all the advances each day.

The biggest challenge for AI isn't the technology itself but energy….and cheap energy.

All the top AI services are SAAS but who is going to solve cheap energy for their Server Farms and reduce price or even free unlimited image generation for instance.

Tony EH

https://www.boards.ie/discussion/comment/123534299#Comment_123534299

Exactly.

On a long enough timeline, AI will take everyone's job, including the guys who "climb a ladder carrying 30kg and stay on a roof in the pissing rain".

At the moment AI is more than just a little stupid and still just a novelity. In fact, there's no such thing as AI. There's just machine learning and that's not necessarily "intelligence", artificial or otherwise. It's not intelligent, it just collates data into readable chunks for its own use and for anyone who wants to use it. It doesn't understand itself or its place in time or space, which are true marks of intelligence.

And it's often very, very, wrong.

However, in about 10 or 20 years time, it will be a vastly different thing altogether. And in 50 years god knows what it will be capable of doing. Climbing a ladder with 30kg and staying up there in the rain will be the least of its tasks.

silliussoddius

My job of drawing hands with six fingers was gone overnight because of ai.

Tony EH

https://www.boards.ie/discussion/comment/123536261#Comment_123536261

That will only give governments the excuse to eliminate social safety nets. UBI will end up being useless for people.

Tony EH

https://www.boards.ie/discussion/comment/123536260#Comment_123536260

Which is an extremely bad development. It was bad enough when some HR blob was looking through CV's without a single clue as to whom they were supposed to be hiring or what the actual role entailed. But there's just Mr. Robot looking for buzzwords.

🙄

ancapailldorcha

https://www.boards.ie/discussion/comment/123537992#Comment_123537992

I don't see how. Surely, it'd be easier to have one universal benefit that several which each require means testing, adjustments and so on.

Tony EH

https://www.boards.ie/discussion/comment/123537505#Comment_123537505

So, in other words you're fine until a sophisticated chat bot is developed than can do all of that.

Tony EH

https://www.boards.ie/discussion/comment/123538010#Comment_123538010

UBI will be the end of social welfare. That's rather obvious I would have thought.

LOL, even AI sort of agrees…

UBI will give governments, especially those on the right, all the ammo it needs to gradually whittle away at social structures like the benefit system that exists in the UK. In addition, you can bet that that will be expanded to include cuts in universal healthcare too.

Government is not going to pay out UBI and keep current social welfare payments and benefits intact.

In theory a UBI sounds good. But in the wrong hands, it could end up meaning more poverty for more people.

littlefeet

This might interest some, to sum up large reasoning models have hit a cul-de-sac

https://www.theguardian.com/technology/2025/jun/09/apple-artificial-intelligence-ai-study-collapse.

Ekerot

As a romance/fantasy writer I want to believe my job will be safe but eventually I know that an entire book might be created for someone to read with just the press of a button.

I think decades from now Art will be divided as before AI and Post AI - there will always be a question mark on how much was really the input of the artist's work with the latter.

poop emoji

https://www.boards.ie/discussion/comment/123537875#Comment_123537875

Nope just thinking of my day today

• several hours long meetings brainstorming trying figure out what exactly customer wants and how it can be delivered and when only to conclude need to talk more 😂
• Helping new team mate understand obscure business logic and features spread across dozen services in different languages and years and of course badly documented with actual implementation long diverged from any documents or code ai might consume
• Spend few hours debugging line by line only to to reach an external library written by Indian team in some prehistoric time, of course error messages and stack traces pointed at completely different area

I’ve seen plenty of code where the AI goes “uh hell no **** this, best of luck” 🤣

In short modern software development is not a technical problem, it’s a peoples problem

People are the worst

ancapailldorcha

https://www.boards.ie/discussion/comment/123538061#Comment_123538061

Yeah, I know. I'm not seeing what the problem is. If governments want to slash welfare, there's nothing stopping them. All I see here is an opportunity for a massive efficiency gain at no cost.

Tony EH

https://www.boards.ie/discussion/comment/123538295#Comment_123538295

On the contrary, there's a lot that stops them. A political party wants to remain in power above all else. There's a reason why the Tories have never attempted to completely do away with the NHS. It would mean their complete demise. So historically they picked at it piece by piece and Labour would have to undo their damage when they got back in. Unfortunately since Blair they haven't been as vigilant with regards to that.

But even so, Conservative governments know that outright dismantling of something like the NHS would not go down well at all with voters, so they tend to negotiate around issues like that and benefits, which they'd love to do away with in the morning if they could.

UBI could be a vehicle for that, if it wasn't tightly regulated and if safeguards weren't put in place with regards to welfare before UBI became a reality.

Look, I was once all in on UBI ever since I first came across the concept from Robert Heinlein many years ago. But over the years I've come to realise that bringing something like that in may make things worse if it's in the wrong hands, and there are just too many wrong hands in politics these days.

Anyway, this is an entirely different discussion, so I'll leave it there.

Pa ElGrande

https://www.boards.ie/discussion/comment/123536911#Comment_123536911

Not really. LLMs allow for an explosion in code, not all of it good quality. All that has to be debugged and maintained sometimes over decades. If anything there will be even more code to fix and maintain, requiring more engineers to keep critical business processes humming along.

https://www.boards.ie/discussion/comment/123538215#Comment_123538215

I am not doubting your skills or problem solving….but the ultimate aim is AI getting to a point where it is smarter than a human. Musk says that is less than 10 years away. I ignore his personal preferances politically and all the other stuff and only listen to his business facts.

If I wanted to be super intelligent, most people would judge that by results and ability to recall facts on most subjects and form coherant facts and thought processes on each.

LLM is like having the knowledge of several blocks of libraries of information and being able to spit them out withing 10 seconds to 10 minutes.

It is not a matter of if but when you get replaced.

Flinty997

If AI is that good why does it get so much stuff wrong, or invent facts that don't exist.

Flinty997

https://www.boards.ie/discussion/comment/123536307#Comment_123536307

My problem within that is I'm often asking it things that should be easy to know and it's unreliable. I get that it will incrementally improve. But it seems a lot of it is poorly constructed if it can't admit it doesn't know something.

poop emoji

https://www.boards.ie/discussion/comment/123539239#Comment_123539239

Musk says a lot of things, most are 🐂 💩

eightieschewbaccy

https://www.boards.ie/discussion/comment/123539239#Comment_123539239

Musk claimed full self drive would be good to go by the end of 2020. He also promised over a million robotaxis by 2020. The man is notoriously wrong on his predictions.

Flinty997

https://www.boards.ie/discussion/comment/123536873#Comment_123536873

I think that's actually a fundamental flaw with it.

AI seems to create answers it thinks you want, but actually isn't what you asked. Unless you fact check it, people don't realize that.

I'm actually finding it to be a really black art to manipulate it to get close to what you want. I'm looking at how I can data cleanse our data and structure that data so its useful in a model. Because a lot of our information and data repositories have a lot of junk and errors in it.