That is to say never.
Our jobs will change. We might look more like product managers than code artisans. Or like lead engineers guiding mentoring and nudging robots to build the right thing.
But I suspect AI will take about as many engineering jobs as spreadsheets did. More people will be writing code and automating their tasks. When the spaghetti starts to break and becomes a bottleneck, we can come take over and turn it into a scalable engineering product. Perhaps with the help of our own robots doing the boring stuff.
This is why I repeatedly tell software engineers not to be “ticket takers”. If your job is just pulling well defined stories off the board where the business requirement is already known - which is considered mid level behavior at every tech company where I have seen leveling guidelines - your job may be in danger by AI. But it is definitely in danger of being outsourced to cheaper labor and you’re just commodity and will find it hard to stand out against thousands of other commodities.
There's an old joke.
There's two types of 10x developers
1) those like Knuth
2) those that complete 10 the number of jira tickets
(Yeah jokes aren't perfect. You don't need to nitpick and but actually. You can interpret )
A theoretical “AI dev” will be able to gather these requirements through conversation no problem. The same way a human dev would. In fact it will probably do a better job. I really don’t think this is some special skill that protects developers from automation.
The industry is about to change drastically. I don’t foresee software engineers going extinct, but there will most definitely be a lot fewer of them.
Maybe! So far what I’ve seen are AI devs that are too afraid to say ”No you’re wrong this will never work” and can’t ask those awkward questions like ”Ok A sounds great, but you already promised to customer B that not-A will always be true. What do we do about that?”
It is very important to have a working mental model of the system to do requirements gathering effectively. AI struggles mightily in that area.
Even without AI the competition from fellow job applicants, overseas talent, etc., would mean we all have to get more productive and more skilled.
But I’m looking to get off the rat race / wheel of futility and looking at things like consulting to augment and maybe replace my job. The autonomy and the flexibility and being my own boss so to speak is what I envy.
(11) The term "supervisor" means any individual having authority, in the interest of the employer, to hire, transfer, suspend, lay off, recall, promote, discharge, assign, reward, or discipline other employees, or responsibly to direct them, or to adjust their grievances, or effectively to recommend such action, if in connection with the foregoing the exercise of such authority is not of a merely routine or clerical nature, but requires the use of independent judgment.
When you think about the power structures of the modern world, there are virtually no examples where the powerful have gained power by swearing off modern technology - that would be a political non-starter. But we do see countless examples of powerful technologists simply crushing the cultures that stand in their way. Rational people see this and act according to their self-interested desire to not get crushed.
All that is to say, I don't think a strike based on AI protection will succeed. A strike based on more fundamental workplace and human rights, possibly. But a reactionary anti-technology strike ain't going to happen.
It's his job to keep the stock price up and people are keeping their ears tuned into "AI" as if it's some magical productivity panacea.
Of course one should always consider what motivations lie behind a persons statement. And be distrustful of people trying to make a gigabuck.
I’m skeptical given how common of a sentiment it is that Facebook and Instagram are dead.
Over the last few years they’ve been adding more and more “suggested” content in the feeds and then there’s talk of adding AI-generated content.
Why would a site flush with activity need to generate it’s own content?
I’d expect AI-generated content from a site getting off the ground to try attracting users, I wouldn’t expect it as a potential mainstay for a site that’s confident of their relevancy.
That said, Meta deserves credit for an impressive launch of Threads and may very well be able to pull it off with another type of social networking site (YouTube competitor anyone?), but the Meta brand is feeling dated.
Additionally, more states are looking into clamping down privacy laws and social media use by <18yr olds.
1. You can take some extremely expensive or geographically irreplaceable assets out of commission while doing it. See port workers.
2. You realize that one day it will all come crashing down, and very suddenly and that you must save very aggressively for that day. Forklift drivers and manufacturing workers making 90K a year until their employers collapsed in 2008 and that freed them of union contracts.
This is exacerbated by tech being a high cycle industry, with a lot of new entrants and large firms (beyond the largest) often dying or collapsing, so chaining a few major players with contracts is unlikely to protect the industry as a whole.
There isn't a scenario where AI engineers exist and we all keep our nice tech jobs long term as they are (handwriting code).
It doesn't help that in a large factory, having 70% onboard is enough to stop a plant. In tech, if 30% break rank, the thing is still getting built.
Need to figure out where the ball is going.
My belief is that we are over optimized for a particular situation that is ultimately unsustainable. You cannot make money if no one can pay you. That is the most serious issue corporations face right now. If you want some sense of that, ask yourself how the stock market is doing versus income. Then ask if it is sustainable for the stock market to do well people do not have money to spend.
There was an interesting, to me, discussion about cargo cultism recently. You have seen "cargo" appear and so you think it is caused by the observable factors. And perhaps we can generalize the term:
Cargo cultism: The belief that you know how to achieve a result when you do not understand the process that produces the result.
We do not understand the effects of AI at this point. It is likely that AI will help people become more productive programmers. But this is a two edged sword for Mark. What happens if I say "AI: write a version of FaceBook in Rust". Or "AI: write a new a library in Rust that reads, writes and displays all files supported by Microsoft Word"?
Hey for now I would settle for an AI that converted all my JavaScript files to TypeScript and back again so I get the benefit of types without the bother.
> AI: write a version of FaceBook in Rust"
If you had all of the source code to Facebook today and could set up the servers, you still couldn’t create a better Facebook.
Microsoft has enough money to compete and create something as good as Google. But still was never able to beat Google Chrome or Google search.
> AI: write a new a library in Rust that reads, writes and displays all files supported by Microsoft Word
Today there is plenty of open source code that does well enough in reading and writing Word files there has been third party code to read and write Word since the 1990s. Yet Microsoft has maintained its dominance.
But of course people just want to move ahead. I know I can't fight the future. I'm fine with that. I'm an old man over 40. I'm going to be obsolete in about 10 years. I'll be debt free and some half million in cash and small passive income too. And you can do whatever you want. Apres moi...
If you are interested in organizing, I recommend reaching out to CWA: https://cwa-union.org/
If AI can truly replace me at scale, I'll direct it to make billions of dollars. Why should we be afraid of AI taking our jobs, our companies should be afraid of what we can now quickly create to compete against them.
Why is engineering work any different?
It is feasible, but industry wide strikes don't just happened. Companies can move software jobs overseas easier than factories, so international solidarity is important - something US labor has notoriously been poor at.
It can happen, but a lot of work and time and effort and defeats would happen first. It starts on a small scale, with an enormous amount of effort to build up. The momentum of the past 75 years have been things going in the other direction in the US.
If AI is going to succeed to that degree (I have doubts), then it only has to succeed in that degree anywhere. Are you going to successfully organize such a strike in India? China? North Korea?
No, you aren't. New technology is here. Those who refuse to use it are going to get run over by those who use it well (if it's all that useful).
If you really think that you will be replaced, you already are valueless to your company.
(Pro tip: If you are valueless, striking is the surest way to get fired)