Last edited:
The AI is taking our jerbs thread
- Thread starter mid_gen
- Start date
Forever job's haven't been a thing since the 80's
I actually have a forever job. But I meant most industries have cyclical downturns and upturns. I don't think AI is going to make everyone unemployed.
I think what's being discussed goes beyond cyclical, as cyclical processes happen in the same order repeatedly.
At the moment, we're in limbo over the extent of AI's impact and I don't think anyone has suggested that it will make everyone unemployed. The key question is how quickly it will become proficient at performing low risk, menial tasks, allowing you to move beyond the stage where a human must vet everything.
The mass displacement of people in certain sectors could be the issue, as well as the types of jobs that become/are available and how oversaturated those areas then become.
For example, experienced software developers are okay because their knowledge and experience are highly valued, even if AI can make certain parts of their role more efficient. However, based on my experience over the last year, more and more employers in this sector are becoming reluctant to take on apprentices, often citing AI as a reason (rightly or wrongly). We have also put ourselves in a difficult position (not my doing), as we cannot really challenge them on this, given how much AI is mentioned in our materials and is now embedded in the curriculum.
At the moment, we're in limbo over the extent of AI's impact and I don't think anyone has suggested that it will make everyone unemployed. The key question is how quickly it will become proficient at performing low risk, menial tasks, allowing you to move beyond the stage where a human must vet everything.
The mass displacement of people in certain sectors could be the issue, as well as the types of jobs that become/are available and how oversaturated those areas then become.
For example, experienced software developers are okay because their knowledge and experience are highly valued, even if AI can make certain parts of their role more efficient. However, based on my experience over the last year, more and more employers in this sector are becoming reluctant to take on apprentices, often citing AI as a reason (rightly or wrongly). We have also put ourselves in a difficult position (not my doing), as we cannot really challenge them on this, given how much AI is mentioned in our materials and is now embedded in the curriculum.
Last edited:
I was replying to the comment about tradesman becoming over saturated.
I've been here before with building booms and collapses, recessions, introduction of computers, the Internet, mobile phones, robots. Dot com bubble.
I've lost count the number of times I was going to be out of work replaced by technology or outsourcing. I'm more worried about ageism or office politics.
I've been here before with building booms and collapses, recessions, introduction of computers, the Internet, mobile phones, robots. Dot com bubble.
I've lost count the number of times I was going to be out of work replaced by technology or outsourcing. I'm more worried about ageism or office politics.
Last edited:
Many of the things you've mentioned aren't cyclical and are nothing like the way AI is being discussed. They never had the potential to replace a person entirely on their own. However, robots have certainly replaced low level manual labour jobs in a few sectors or reduced the number of people required to do that role. Fortunately, the demand for things like car manufacturing has increased steadily, so despite robots being able to perform things like welding or spraying tasks, there is still a need for a similar level of low level manual workers. A similar argument could be made for AI, but I don't believe many businesses will be able to grow exponentially to the point where they require a similar number of low level workers.
A building boom and collapse is normal, as are recessions, both are recoverable. I'm not sure who suggested that mobile phones were going to put people out of work.
These public LLMs are a few years old and most people's experience of using them involves asking dumb questions, such as who the current US president is, despite knowing there is a cutoff date in their training data or easily being able to ask a follow up question as to why it "thinks" that.
It isn't just about your job or sector, it's mainly being discussed in terms of replacing lower level desk jobs of which there are millions in this country alone, where the daily routine involves repetitive tasks and similar outputs. Based on what even the latest models can do when used alongside other tools, we aren't quite there yet, but considering the progress already made, it's not a massive reach to see it being possible within the next few years.
A building boom and collapse is normal, as are recessions, both are recoverable. I'm not sure who suggested that mobile phones were going to put people out of work.
These public LLMs are a few years old and most people's experience of using them involves asking dumb questions, such as who the current US president is, despite knowing there is a cutoff date in their training data or easily being able to ask a follow up question as to why it "thinks" that.
It isn't just about your job or sector, it's mainly being discussed in terms of replacing lower level desk jobs of which there are millions in this country alone, where the daily routine involves repetitive tasks and similar outputs. Based on what even the latest models can do when used alongside other tools, we aren't quite there yet, but considering the progress already made, it's not a massive reach to see it being possible within the next few years.
Last edited:
Mobile phones replaced public phones, and phone exchanges and all the switching equipment used in them. Had a contract replacing computers in a telecoms all these exchanges all stripped out. But everyone just got different jobs.
People don't see it because they mostly exist in the now. I used to work on content that was sold on floppy disks, then on cdrom then internet. People had jobs doing all that, now all gone. I was trained in technical drawing using pen and ink. Also now all gone. Printing using bromides all gone. Developing film for photos all gone.
I use AI now a bit at work. Building agents etc. I think it's meh. It will be another tool to be mastered. Anyone displaced will get a job somewhere else.
People don't see it because they mostly exist in the now. I used to work on content that was sold on floppy disks, then on cdrom then internet. People had jobs doing all that, now all gone. I was trained in technical drawing using pen and ink. Also now all gone. Printing using bromides all gone. Developing film for photos all gone.
I use AI now a bit at work. Building agents etc. I think it's meh. It will be another tool to be mastered. Anyone displaced will get a job somewhere else.
Last edited:
Right, but that would probably have involved a few thousand jobs spread throughout the country and the phone exchanges and switching equipment would still have been needed, since not everyone got rid of their landline as soon as mobile phones became available, and the exchanges were still required for broadband.
The hope with AI is that it will be a gradual process like that and that's what's needed, because I don't see how an economy can function if demand for low level office workers plummets in a short space of time, when most of those people can't just move into a similar low level role.
When talking about AI "replacing" workers, there's a crucial distinction here. Will businesses choose AI over human employees to cut costs? Absolutely. Will AI actually perform those jobs as well as humans do? That's a different story entirely, at least for now. The real threat isn't AI's competence, it's how well AI vendors can sell the dream to decision makers. Your job security depends less on whether AI can truly match your skills and more on whether someone can persuade your boss that it can. What worries me about the future is that we're heading toward a genuine skills shortage, not the fake "skills gap" companies complain about when they don't want to pay decent wages. As AI gobbles up entry level positions, particularly in low level tech, we'll stop developing new human talent. Soon we could find ourselves completely dependent on systems that almost nobody knows how to build, fix, or even understand properly. This same pattern has/will hit customer service too, though apparently society has already written off those workers as expendable.
In the short term, I don't think AI alone will truly replace a job (although it may be used as an excuse), but people who use it efficiently may see others made redundant or fewer job opportunities. At the moment it's just a tool that can allow you to complete certain tasks more quickly and efficiently than you could without it. Those who learn to use it effectively will excel, which becomes immediately obvious when using one of the latest models, such as Claude 4 Opus, alongside a tool like Cursor. Within a year or two, we'll probably look back on that model as being rudimentary, as we've only just begun to scratch the surface of using them in combination. Many businesses are also developing SLMs with domain specific knowledge.
The hope with AI is that it will be a gradual process like that and that's what's needed, because I don't see how an economy can function if demand for low level office workers plummets in a short space of time, when most of those people can't just move into a similar low level role.
When talking about AI "replacing" workers, there's a crucial distinction here. Will businesses choose AI over human employees to cut costs? Absolutely. Will AI actually perform those jobs as well as humans do? That's a different story entirely, at least for now. The real threat isn't AI's competence, it's how well AI vendors can sell the dream to decision makers. Your job security depends less on whether AI can truly match your skills and more on whether someone can persuade your boss that it can. What worries me about the future is that we're heading toward a genuine skills shortage, not the fake "skills gap" companies complain about when they don't want to pay decent wages. As AI gobbles up entry level positions, particularly in low level tech, we'll stop developing new human talent. Soon we could find ourselves completely dependent on systems that almost nobody knows how to build, fix, or even understand properly. This same pattern has/will hit customer service too, though apparently society has already written off those workers as expendable.
In the short term, I don't think AI alone will truly replace a job (although it may be used as an excuse), but people who use it efficiently may see others made redundant or fewer job opportunities. At the moment it's just a tool that can allow you to complete certain tasks more quickly and efficiently than you could without it. Those who learn to use it effectively will excel, which becomes immediately obvious when using one of the latest models, such as Claude 4 Opus, alongside a tool like Cursor. Within a year or two, we'll probably look back on that model as being rudimentary, as we've only just begun to scratch the surface of using them in combination. Many businesses are also developing SLMs with domain specific knowledge.
Last edited:
That's why you have to be early in.
I'm thinking about changing out of IT into something safer. Yeah it could become swamped. But earlier you are in the better imo.
I'm less concerned about the current AI. It's that big AGI leap. If that does come around though I guess all bets are off.
Seems like it could be 1 year or 10 years away. But when it does come, we will know about it quickly!
I raised this earlier. The problem is also there is not really any growth in trades, they scale linearly with the population.
Wages will drop and there will be lots of competition.
In the energy industry we're looking at how to use AI to do more with the same resource TBF.
Meter readers are (mostly) not trained in much beyond being able to do just that.
What we can do with AI though, is use photos taken of the meter by a reader to validate their readings to do additional proactive safety checks.
Even early tests are showing that it can do maybe 5 people's work for a day in a few minutes with very impressive accuracy.
I dont think anybody involved is suggesting, yet, that we'd leave gas and electric meter safety down to AI alone though.
Meter readers are (mostly) not trained in much beyond being able to do just that.
What we can do with AI though, is use photos taken of the meter by a reader to validate their readings to do additional proactive safety checks.
Even early tests are showing that it can do maybe 5 people's work for a day in a few minutes with very impressive accuracy.
I dont think anybody involved is suggesting, yet, that we'd leave gas and electric meter safety down to AI alone though.
There are not that many layoffs currently and unemployment rate for software engineers and related tech workers is low, lower than national average unemployment. Covid was just an anomaly when tech companies made a weird land grab for any developer or tech worker and paid new recruits 200k a year because they once read a blog post about AI.
A lot of the news about tech company layoffs is not even layoffs of the tech workers , but HR, CX, PM, paralegal, executive assistants etc.
e.g. with IBM
Why the Viral “IBM Fired for AI, Then Rehired” Posts Are Misleading—and What They Really Reveal
Over the past weeks, our LinkedIn feeds have been flooded with posts celebrating a dramatic “victory of human judgment over AI.” Headlines like “IBM fired 8,000, had to rehire them when their AI flopped” are getting massive engagement.
www.linkedin.com
They reduced HR staff and hired as many people but in higher paying roles that add more value. The AI layoffs resulted in more engineers and tech workers being hired.
A lot of the agentic systems being developed right now will mature in short order, it's going to take a while for trust to materialise. Developing LLM application is also so fundamentally different to any kind of technical advancement previously it's taking a while for people to learn how to use them effectively.
I've been building agents in LangChain/LangGraph for a few weeks now and, and the possibilities for these systems to phase out a large number of roles is obvious. Just the coding tools alone can make an experienced programmer vastly more productive. I'm being approached to advise on AI assisted coding, and even had chat with a publisher about writing a book. Demand right now is enormous.
I've been building agents in LangChain/LangGraph for a few weeks now and, and the possibilities for these systems to phase out a large number of roles is obvious. Just the coding tools alone can make an experienced programmer vastly more productive. I'm being approached to advise on AI assisted coding, and even had chat with a publisher about writing a book. Demand right now is enormous.
Caporegime
- Joined
- 7 Apr 2008
- Posts
- 25,733
- Location
- Lorville - Hurston
Aren't you not using python to write llm?
You will never be able to fully trust those systems, and the degree to which it matters will vary across industries. Obviously there are efficiencies to be had, which will likely lead to lay-offs, but there should also be hiring to take advantage of said efficiencies.
Most of the hype is from people who have very little understanding of the technology, or just want a badge on their LinkedIn profile.
Just the same as you never fully trust a human agent in any critical process. There's nothing fundamentally different about having an AI agent do a task versus a human. You still need all the same checks and processes in place. AI agents are actually rather good at checking each other's work....and even then, any decent LLM application will have human-in-the-loop processes to handle aberrations.
100% is going to be a major net reduction in headcount across a huge range of industries.
Most of the dismissal of AI is from people who have very little understanding of how AI systems are developing imo, the pace of advancement is insane.
LangChain is the main framework for building agents. LangGraph is what you use to orchestrate them together into multi-agent workflows. Using Python for both.
Not training the LLMs themselves, no point, just plug into whatever model you want to use.
When the topic comes up, it becomes immediately apparent when you ask as little as one or two probing questions about how they have used it so far.
Issue comes down to three things:
* data privacy/sovereignty (recorded sessions and changing vendor T&Cs are concerning legal teams)
* shared responsibility of continuously keeping up with the next version
* ensuring differentiation as a business
First cloud and now AI is carving away the business model from the inside out.
Last edited:
You're not bound to a single vendor and it's quite straightforward to switch to locally-hosted models, if you can get the GPUs
No different to any other asset, software or hardware.
The big differentiator is that it's actually quite easy to just start testing the water with AI systems by integrating it into one tiny bit of your business, with a human supervisor. Once the tools are proved, the ROI is an absolute no-brainer. Why have meatsacks that sleep and expect things like pensions and sick days, when you can just have an agent do the job for a tiny percentage of the cost?
Yes AI won't replace ALL jobs. But it is more than GOOD ENOUGH to make the commercial decision a no-brainer for a significant number of roles.
It's going to be a lot more transformative than cloud imo.
You're not bound to a single vendor and it's quite straightforward to switch to locally-hosted models, if you can get the GPUs
No different to any other asset, software or hardware.
The big differentiator is that it's actually quite easy to just start testing the water with AI systems by integrating it into one tiny bit of your business, with a human supervisor. Once the tools are proved, the ROI is an absolute no-brainer. Why have meatsacks that sleep and expect things like pensions and sick days, when you can just have an agent do the job for a tiny percentage of the cost?
Yes AI won't replace ALL jobs. But it is more than GOOD ENOUGH to make the commercial decision a no-brainer for a significant number of roles.
Issues are that large organisations don't have the skillsets for cloud and modern data privacy, let alone AI!
I kid you not - I've had to raise an objection at a security review that someone wanted to pull production PII data onto their laptop to 'analyse' the data. When I veto'd it, the fall out of we always allowed this in the past etc etc.. Another veto was a project that had a third party vendor who, as part of the integration to provide container controls, had to have full AWS access at org level and their code then managed the AWS instance from outside the organisation.. I kid you not on that either.
I get AI.. I get its pros and cons. Having been the guy accountable for things (including AI offered through the platform), it needs skills before you can run with it.
Just as tangent - the US DoD already have accelerated and now provide IL5, IL6 and Top Secret AI instances. So that means (within their compartmentalisation) that they have AIs. The US and UK clearance levels map which means it's like the UK having TOP SECRET classified information being processed by AIs.