I don't understand this line of reasoning. Like genuinely. So with AI coding (let's just limit ourselves to coding); are you saying that the Agent is going to prompt itself? Like it exists only to read your mind and create precisely the code you wanted or didn't even know you wanted? Or will you have to explain and verify that it did what you asked? At some point we run into magical thinking and absurdities.
Programming or math are not like Chess or Go. There is no endgame to win. And the human/input/judgement/whatever and where that begins or ends isn't a technical issue but a political one.
So my question: are you expecting that at some time N that models are so good that they can read your mind? Or are you saying that you will just be able to "speak" into existence any type of software? And how are you going to specify this if you can't already point to something similar?
Isn't that what a well run company does when creating a process? Bureaucracy and process, reduces the penalty of weak domain context and in fact is designed to obviate that need. It "diffuses" the domain knowledge to a set of specifications, documents, and processes. AI may be able to accelerate it, or subsume that bureaucracy. But since when has the limiting factor been "finding someone locally who knows the process?" Once you document a process, the power of computing means you can outsource any of that you want no? Again, AI may subsume, all the back office or bureaucratic office work. Perhaps it will totally restructure the way humans organize labor, run companies, and coordinate. But that system will have to select for a different set of skills than "filling out n forms quickly and accurately." The wage stagnation etc etc. predates AI and might be due to other structural factors.
I am not seeing any evidence of China "replacing teachers with AI" anywhere (did some googling/geminiing). Are there any sources on this? Seems like they are trying to introduce students to GenAI/ML principles and creating "AI literacy guidelines" without just "replacing teachers with AI". Their current guidelines outright prohibit the use of AI to replace teachers' responsibilities.
What is the point of asking it to teach you something to "understand it" if Claude can just do it for you? This is the real question everyone should be asking beyond just employment (employment will definitely change in the coming months, no doubt). I would pivot away from programming personally.
I don't disagree, writing code will probably no longer be a thing in the near future. This is probably also true for all knowledge work (math, design, etc etc.) which is literally anything that can be "reduced" to mechanical transformations on symbols. Including music gear design, design of plumbing fixtures, tooling jigs (CAD work) etc etc. It is all basically transforming a specific set of discrete symbols into other ones or stringing them together or re-combining them etc etc. I wouldn't call that a doomer take either. But yes, the "Claude 4.5 still makes mistakes" thing is played out "remainder humanism"/ "John Henry vs. the machine". I fully see the value of Software products to go to zero with a whole bunch of money being funneled into one of the AI companies. It is a scary time to be around. I would stop learning coding and/or any framework or specific technology.
Wait you used Claude Code to recreate patents and schematics? Are the schematics for this easily available somewhere? Was Claude just able to one-shot this?
more pointedly: the commenter presumes that the friends are unhappy with their lives. Also that some them would be better served performing back-breaking menial low-wage labor while otherwise being illiterate. Any PhD (even one in specializing in Plankton and especially nuclear physics/engineering) would equip you with a bunch transferable skills that normally would be valued in a modern society ... 1) public speaking 2) initiative 3) resourcefulness 4) analysis and communication etc etc. If I was being uncharitable I would say finance and law are actually worse for society: at least the subset of those that get paid the highest with respect to their impact on the broader society (but that is debatable).
Would you think it would be better for "society" to allow more people to go into finance and law? Or that advanced knowledge should be gate kept by only the select cognitive elite that are most adept at playing the "glass bead game" by age 18? Would you change any of your opinions if AI renders most High IQ practical/technical tracts obsolete? Perhaps, a more sane society would be one where curious people could develop themselves in whichever way they so choose: if they want to study the mating habits of marmots in the Central Asian steppe then so be it.
No one really knows. But a few things I think about myself.
1). There are many many people there couldn't probably already write more lines of code than me and work for much much cheaper (in India or wherever). Same is true for you (probably). Yet you still have a job.
2). I have a friend that works as a Software and System Engineer for a complicated product that interfaces with the real world. He has to use Natural Language to create requirements that gets turns into code by natural agents down in the "Supply Chain". There are also integration engineers that work with the naturally intelligent agents that create the prompts/requirements to make sure things don't fail (then triage and root cause when they do)
3. Why not diversify your skills beyond code but also hardware, systems, soft skills, business etc etc.
I don't think there is anything more than the standard advice. Just stay curious, make friends/build a community, keep learning, stay healthy. Why not get the AoE? you can also, check out "Practical Electronics for Inventors": AoE assumes you have some Electronics background imo. But seriously, I don't get the doom/gloom: things are going to be rough ... but maybe they won't? Many things I learned I did for their own sake! Things have always been uncertain and absurd I guess we might as well embrace it!
Programming or math are not like Chess or Go. There is no endgame to win. And the human/input/judgement/whatever and where that begins or ends isn't a technical issue but a political one.
So my question: are you expecting that at some time N that models are so good that they can read your mind? Or are you saying that you will just be able to "speak" into existence any type of software? And how are you going to specify this if you can't already point to something similar?