That feels like not in our life times to me. Possibly not our children's either. I don't know...how long away do you think that will be? Decades?
The AI is taking our jerbs thread
- Thread starter mid_gen
- Start date
That feels like not in our life times to me. Possibly not our children's either. I don't know...how long away do you think that will be? Decades?
Well, as mentioned in my other thread about finding work, I'm considering jumping with a new fintech startup that is making extensive use of cutting edge AI systems. I've spent a couple of days looking into it, and frankly, there a massive warning klaxons going off about what is coming in terms of employment. Don't think I can pass up the opportunity to skill up on it, because the things that are in the pipeline are going to be a massive paradigm shift, particularly in services.
Just reading up around AI agents and seeing what is being done with it, it's just like, yikes. So many jobs are gonna be disappearing.
Just reading up around AI agents and seeing what is being done with it, it's just like, yikes. So many jobs are gonna be disappearing.
Last edited:
Robots are used for surgery all the time, in fact it is the preferred approach. My wife had a major operation mast year done entirely by a robot.
Technically, it would be trivial for robot arms to do hairdressing today. The only reasons this isn't the norm is the cost, and social aspect because people like to chat to their hairdresser etc. I suspect automated hair saloons to be a thing within 5-10 years starting in China or Japan.
Cost will inevitably always be a factor. The only reason far more manual work is not done by robots is that paying people minimum wage is cheaper. The moment that equation changes then anything from chefs, hairdresser, electricians, gardeners etc. will become automated
Realistically that honestly is not happening for a long, long time. Interfacing with a robot that turns up at your door; "Hi I'm here to rewire your house".
Sure you can pick and choose some extremely complex cases like that are going to take a while.....but it will get there. Lots of simpler problems to solve that will get automated in the near future.
Caporegime
- Joined
- 22 Oct 2002
- Posts
- 28,664
- Location
- Boston, Lincolnshire
We already have robots that are capable of replacing a human body. It is only the brain that is required.
With Ai it is just a matter of time but we do not know how long that will be.
Although look how far Ai image generating has come along in the space of 12 to 18 months.
Currently the limitation is in having a suitably dextrous robot that is also mobile. We have super dextrous and precise arms/hands but not attached to anything that can move. The next biggest hurdle is simply the cost - as long as labour id cheap then it likely doesn't make sense to maintain a complex robot.
What is highly possible in a few years would be to replace electricians and plumbers etc. with a very lightly trained technician wearing AR glasses that does all analysis and gives directions - driving the human. Humans make cheap mobile robot arms.
I see this kind of battle of high cost high tech vs low cost labour played out in the vineyards where i live. Some vignerons have super expensive automatic harvesters, trimming machines and even drones for spraying. Others get in labour from poland and Romania. Both approaches are viable but if labour costs increase then they will all just invest in automation.
Car mechanics could be something to be automated soon. Labour costs are already high, but there are not that many differences between cars and the robot wouldn't have to be fully mobile, only needing to circle the car on some rack system. Costs of the machine would be high, but it could churn through cars 24/7 for years and years. Ot wouldn't even have to do everything , being programmed to solve the most common problems/services would probably cover 90% of all requests
Generative AI is not replacing jobs or hurting wages at all
: 'When we look at the economic outcomes, it really has not moved the needle'
www.theregister.com
I don't think that anyone is saying that its replaced lots of people already however anyone using these and agents can see the potential that very shortly (years) it will be able to. Payroll decisions normally are longer term so people are cautious especially following the Covid period of over hiring that the same mistake isn't made again.
At the moment it's India who is taking all the jobs and their currency is on the decline so it makes the ROI even better for anyone moving functions there.
At the moment it's India who is taking all the jobs and their currency is on the decline so it makes the ROI even better for anyone moving functions there.
'so far' is doing the heavy lifting here.Generative AI is not replacing jobs or hurting wages at all
: 'When we look at the economic outcomes, it really has not moved the needle'
Doing a job that involves sitting in front of a computer? You should probably understand what is coming. Well, not coming, it's already here.
That's pretty broad though. It's good for relatively simple and well understood tasks, with some level of human oversight.
Hopefully there's equal effort getting it into the medical and legal space. Could massively reduce costs and wait times; whether that gets passed on to the general public is another story.
That's kinda the point. You can just put the agent in front any web app and it can figure out what to do.
There are a LOT of people working in jobs that involve mainly simple and well understood tasks, that require some level of additional human oversight.
Oh I'm sure it is. The economics of having an agent do a job unsupervised 24/7 versus a human that wants holidays, sick days and pensions and all that....is going to be irresistable.
I was including more highly skilled roles, like software development where they're more of an aid than a replacement, but can absolutely replace/streamline a lot of script-style work.
It's unfortunately going to spike unemployment because companies only care about the bottom line.
A lot of these AI companies are haemorrhaging cash so will be interesting to see what happens when there is a decent level of adoption/reliance.
I'm thinking of longer term learning to be a diving instructor. Obviously the pay will be small, and ill need to leave the UK. But it seems like something I can do, I'd love doing and it will take me out of the UK and it's a long way off AI-able.
Or something along those lines. Something in tourism.
Or something along those lines. Something in tourism.
Last edited:
Interesting idea. Though, thinking about it, it feels like a Tesla or similar could be quite effective at teaching somebody to drive. We've already seen how effective the LLMs can be at teaching. In the case of driving, unlimited patience, can stop the learner from doing something unsafe, doesn't need payment!
I feel the jobs that require high levels of dexterity such as builder, electrician, plumber, etc, are going to be safer bets, at least until a highly dexterous and affordable robot becomes available.
Last edited:
Diving is too niche. For starters, none of the short horizon robots would be anywhere near able to swim. I could imagine dive assist underwater AVs eventually coming to market but doubt they would be used in a teaching manner.
The economics just wouldn't make sense. Compare that to an autonomous truck where tye benefits of 24/7 driving and minimal mechanical complexity makes it a no-brainer. And even jobs like an electrician has a mechanical complexity, but there is a large demand and on build sites could work 24/7.
I was at AWS Summit London yesterday, seems that AWS is continuing to push with new features etc but also push the line 'AI as a judge' etc, ie get used to handing over control for both building and verifying (including security) to ai. A nice level 400 (ie most technical) example of AI use in cybersecurity incident detection and management, but also resolution (with caveats about being experimental).
Nope.. I suspect we're at least 5 years off having a genAI detect and then modify the firewall settings etc. The example had a human in the loop 'do you wish to modify' but as part of a Jupyter notebook and so with a massive 10,000 application estate, that's not going to wash.
Nope.. I suspect we're at least 5 years off having a genAI detect and then modify the firewall settings etc. The example had a human in the loop 'do you wish to modify' but as part of a Jupyter notebook and so with a massive 10,000 application estate, that's not going to wash.
Associate
- Joined
- 20 Jun 2023
- Posts
- 773
- Location
- uk
It can ******* have mine. It'll soon regret it.
Lol I read it as driving. My bad. Yes - diving probably a good bet!