Ask HN: Why won't you be replaced by AI?
Posted by atleastoptimal 1 day ago
AI models are rapidly getting better. The general public still hasn't seen the capabilities of Anthropic's Mythos model, which is already 4 months old at this point.
I've seen many arguments about why certain jobs will always need a "human in the loop", or that certain skills aren't replaceable by LLM's, but I am skeptical of this notion. It seems that if the general intelligence of these models continues to increase, then every job is just a matter of feeding in the right context, designing the right structures for agents to collaborate, having the right verification loops, etc, all of which are difficult to create, but not impossible.
If the improvements in models continue at the pace they have over the next 3 years (reminder, 3 years ago the best model was GPT-4), do you really believe that what you do now will not be done better by a system of LLM-driven agentic harnesses?
Comments
Comment by mrothroc 6 hours ago
I'm fully into agentic coding, and I'm actively studying agent reliability. I can make all kinds of deterministic guarantees about agentically produced code, but no, I would not accept either of these, or many other examples.
We write code (or we used to, before AI), so we naturally value that. But the code is one small part of a deployed system, and this has always been the case. Numerous studies have shown that writing the code is actually the cheapest part.
All the most important things that you need to know about what makes the code correct, both before it is written and after it is deployed, are not in the specs. They come from walking around and talking to people. Looking them in the eye, sussing out what their real requirements are, and figuring out how to address their concerns with empathy.
Until LLMs can do that, I'm not worried. Let them write the code, that's the least important part.
Comment by atleastoptimal 4 hours ago
This claim depends on humans remaining the bottleneck to high-leverage information, necessitating a human<->human interfacing role that solicits requirements, ascertains intent, etc.
I don't deny that that is very important and cannot be done by AI now. However, my concern is that AI will be much better at any domain of information processing, and organizations that gate important decisions being made by a network of barriers and information silos dependent on "talking to people" will be outcompeted by largely autonomously run AI agent organizations, which have, by their very design, far higher throughput, auditing, memory, parallelization, etc.
It's kind of like saying that machines could never make fabric because it is impossible for a robot to replicate the complicated motion of a human threading a needle. The industrial revolution was prompted by creating machines which redesigned the entire process to account for machine limitations, and allowed the superior speed and scale of machines to drive higher productivity, delegating humans to a role of maintenance and simply feeding the machines their needed input.
Comment by mrothroc 2 hours ago
The output of agents has to have economic value, and for the foreseeable future this means someone is going to have to buy something.
Right now, it is humans who ultimately make the economic decision. Even if you have fully autonomous agentic organizations selling to each other, there are one or more humans at the end of the of the chain who agree to exchange money for value.
Science Fiction writers have envisioned futures where some currency other than money is used to track value, but so far as I can tell we are nowhere near moving to anything like that.
Comment by andrei_says_ 1 day ago
LLMs do not replace jobs, managers constantly try to increase profit. Managers and business owners try to replace people’s jobs via automation.
LLM companies actively try to convince managers and business owners that LLMs are capable of replacing humans in white collar jobs.
An LLM/AI will not replace you, or me.
A business owner or manager might. They also might use an LLM to make decisions.
But let’s not use language which anctively obscures the decision makers in this process.
Comment by nicbou 16 hours ago
Comment by taurath 22 hours ago
Comment by torben-friis 1 day ago
My job is safe.
Comment by atleastoptimal 1 day ago
My question is, is that thing which you are doing, ascertaining the subtle concerns, soliciting requirements, etc. truly out of the range of what an LLM or LLM-guided system could do?
Comment by nicbou 16 hours ago
The stuff I document is happening offline behind closed doors. I create tools and leverage my contacts and audience to surface that information.
Then LLMs suck up that information and serves it to users without attribution or consent, and that is destroying the economics of doing my job.
But the job still needs to be done.
Comment by byoung2 1 day ago
Comment by atleastoptimal 1 day ago
For cashiers, beyond simply ringing up customers, they serve the function of:
1. Validating IDs
2. Preventing theft
3. Creating a positive atmosphere
4. Helping customers bag groceries
5. Resolving issues/questions about products/the store
For waiters, likewise they have the job of
1. Creating a positive atmosphere
2. Physically bringing food to the table and setting the meal
3. Upselling items, providing recommendations, catering to specific unusual guest needs
etc. Basically all these jobs have a huge soft-skills dependent interface which no technology currently can replicate what humans can do.
I don't think that every job can be trivially automated by a large language model, but any job where the inputs/outputs are entirely via a computer, LLM's are approaching the point where they are equivalently enabled to a human, and there is no "real human body in-person" soft-skills interface.
Comment by byoung2 1 day ago
Comment by adampunk 22 hours ago
Comment by byoung2 8 hours ago
Comment by jlengrand 15 hours ago
Comment by ex-aws-dude 1 day ago
Who is the best possible person you could hire to operate the LLM?
Who has a good mental model of what its doing underneath and has the best expertise to direct/guide it?
IMO no one is better positioned to use these tools than software engineers
Comment by atleastoptimal 1 day ago
Comment by ex-aws-dude 1 day ago
At minimum you at least need CEO -> product person -> LLM
There can any number of agents/loops after that but someone has to translate the requirements to the LLM and monitor/verify the results.
What I'm saying is in the extreme case the SE evolves to become the product person.
If you want to argue for fully autonomous companies then thats drifting into sci-fi
Comment by atleastoptimal 1 day ago
> If you want to argue for fully autonomous companies then thats drifting into sci-fi
Why couldn't a software company be completely automated with sufficiently powerful LLM-run agents? What fundamentally is the barrier, if the models are intelligent enough?
Comment by mejutoco 4 hours ago
Comment by bdangubic 4 hours ago
Comment by ex-aws-dude 1 day ago
Long term maybe but its not a very interesting conversation to talk about things that far out.
Comment by atleastoptimal 1 day ago
Comment by Copemaxxing 20 hours ago
Yes.1 where there used to be 1000.And they are the cream of the crop.Their time is precious because their skills are equally so.
Are you out coding the LLM's even now, or do they already do what took you weeks, in hours?
While you are scoping out the scene and hesitating whether to tell the boss you will have it in two or three weeks, LLM got it done 5 minutes ago.
Now off to the other massive LLM Super Code Reviewer.
The off to the Mega Brain Ultra talented human coding team of one or a handful for the human touch.
Are you the cream of the crop?
If not, start making contingencies ...or cope harder.
Wake up!
Comment by ex-aws-dude 17 minutes ago
But seriously my whole team is using LLM tools and not a single person has been layed off.
Productivity doesn't even feel that much higher in terms of number of features shipped.
I think the reason is Amdahl's Law
Probably 40% of my job is actually writing code, so even if you speed that up by double its still a fraction of a fraction.
Comment by 8b16380d 11 hours ago
Comment by comparedge 1 day ago
Comment by andsoitis 1 day ago
Comment by atleastoptimal 1 day ago
Comment by blinkbat 1 day ago
Comment by erminpour 1 day ago
I'm so tired of hearing about AI.
Comment by atleastoptimal 1 day ago
Until it's no longer true
Comment by balefulboy 1 day ago
Comment by andrei_says_ 1 day ago
Comment by comparedge 1 day ago
Comment by cyanydeez 22 hours ago
Comment by volume_tech 11 hours ago