Hacker News Reader

Comment by ericbarrett | Hacker News Reader

AI is a ratchet. It doesn't die, or retire, or change careers. Once it's good at something, it stays good. Once it gets better at something, it stays better. (If not, the model is disposed of without pity or remorse.) Right now, it may make you better at your job. But your very use of it is training it to do your job.

It seems to me that if AI is currently helpful to you, it is likely to eventually replace you. And me, and perhaps most skilled jobs as they exist today. It won't be instant, but if you are early- or mid-career, take it very seriously.

View on HN

What does "take it seriously" mean? For me, taking it seriously means that I am trying to use it as multiplier to be better at my own job. Is there some better course of action you think we should be taking? I personally do not think it is going anywhere, so I figure I had better be the expert.

“Take it seriously” to me means that I’m using AI to be more productive while simultaneously preparing for a future where I can house and feed my family without relying on being employed as a knowledge worker. What exactly that looks like, I’m still not certain.

> so I figure I had better be the expert

Be an expert at what? Prompting AI to do things? There will be little to no expertise needed in the future as these tools evolve. That’s precisely what makes AI so disruptive. A basic set of instructions can be used to create something that previously would have took an entire team of skilled specialists to produce.

Sure, there will always be some experts, but these jobs will be few and far between, and they’ll be much more academic in nature. Having a job like this will be a luxury only a tiny percentage of the earths population will enjoy.

The only real question at this point is how fast will AI move to automate all of these jobs away. It could be 5 years it could be 20.

The jobs left in the wake of AI won’t require much education, skill, or talent, meaning they’ll be extremely low paying because almost anyone will be able to do them.

My problem is that without some legislation in place, there is no stoping this

In 10-20 years it will be robots with guns patrolling food production and distribution so you won’t even be able to steal a loaf of bread for your family.

I’m actually really concerned that if a more equal system doesn’t come soon. Many, many people will suffer very badly.

IDK saving money? Make your life more about just your career like some people do? Train mental resilience? Learn a new skill? Vote better politicians who actually talk about this as a real problem?

We can all do a lot about this. Not saying it's easy though.

Treat your job like a major league sports career. You have 10 more years at most. Earn and save what you can.

I'm pretty much responding to this by trying to maximize my earning potential immediately (TC is the only king. No more gambits where I might take a lower-paying job to "learn more" or have better WLB) and planning on limiting my expenses.

I think the question is, how fast can it get better? You could apply the same reasoning that you did to compilers, and some people probably lost their jobs, while others grew along with them. Compilers have been improving for a while. If AI grows as the same rate as compilers, I won't worry much. It's if/when it starts growing 10/100 times faster that things will get very serious.

Speaking about that, do you have any suggestions? I'm very early in my career, and have never experienced big transformations, huge layoff, large crisis. I guess one of the ways is regulation/unions/stuff like that, but I also don't know much about it. I'd be happy to hear how other people prepare themselves for possible industry disruption.

Any tool that is helpful to me is likely to eventually replace me?

And if I'm eventually replaced, so be it. I'll go do something else.

If there's nothing else to do I'll starve with the rest of the peasantry.

And remember, the last to starve will be the first to suffocate.

Every 10 years the software paradigms change so much that you practically have to start over. Some skills carry over, but many don't. LLMs will do the same thing - they will force us to change our approaches.

That 10 years is not a constant, it used to be 30 years then it was 20 and now it seems to be 10 or it may already be less than 10.

That's a great comment thank you.

Reply on news.ycombinator.com