I think the author is mostly right about the current state of AI, but his future predictions (or worries) are based on a false premise: that the massive LLMs will keep improving in the future.

As far as I have seen the improvements have clearly slowed down, while the energy consumption is rising linearly (or worse). It's like the energy (money) vs. performance graph is logarithmic, and the companies are expending double the energy to get a 10% improvement. Something like that is not sustainable, and the money seems to indicate so.

I really think that LLMs are a dead-end for AI. A really useful dead-end, once the bubble pops and with time, we get a useful working model for them, probably based mostly on local LLMs, maybe using specialized training data.

Give it time. My software career is also affected. At the rate they're spending money at an order of magnitude higher than they're making. They've also all borrowed money from each other. It's going to collapse in a big heap. Hopefully before it sucks in mum and dad investors.

Yeah, I find it difficult, too, especially since management hasn't caught onto this yet and still wants me to specialize.
And of course, the answer is that I should specialize in AI, because there's currently a lot of new development happening there. But that knowledge is also getting obsolete by the minute, with ever more tools coming out and then again other tools that operate those tools for you.

The one thing I hold onto, is that no matter how the situation evolves, the basic job requirements for software engineering, i.e. being smart and being able to learn quickly, will always be an advantage.
I don't think it's possible to hold onto the confort zone from before, even if the industry implodes from the AI costs becoming transparent. But yeah, I do think we'll land on our feet in one way or another.

The post misses a few things:

1. The ai bubble is currently being subsidized to an unimaginable degree. If you were to actually pay true cost for your token usage, you wouldn't be saving that much over an engineer's salary. Probably even worse once AI companies start to extract a real profit. 95% of companies diving into agentic labor will be in for a rude awakening when they balance next year's budget.
2. The cost to keep ai useful in its current form has a high floor. Unless you keep up with expensive training, your models will drift. You can only scale your model intelligence with more hardware (roughly). In two years, Claude opus 4.8 will still be bloating context to learn about the latest cloud platforms and libraries. A human engineer will get those passively at no cost to the company.
3. As the complexity of the task grows the complexity of the ai babysitter must match it. Even if Ai stays cost effective, companies can now save money by spinning up bespoke in-house software to cut out vendors (think observability platforms, task tracking, product design, marketing systems, etc...). No matter how many adversarial reviews and sub agents you spin up, an Ai can't grasp the full context of your company and it's shifting priorities. The software engineer role transitions to a pseudo-sysadmin + product architect.

C-suites don't want know about software and don't care about non functional requirements (security, availability, audit ability, etc...). They just want to wave a magic wand and have a product appear, which is what Ai provides the illusion of. That's why all current Ai software is garbage, but the smarter companies will catch on

Well frankly i think the field of eco terrorism is about to expand greatly so maybe start preparing for that.

I'm hoping the insanity calms down in the next year or so. I don't wanna have to give up weed to be able to start using my fallback option of my CDL but I'll be damned if I'm gonna take a job where I'm required to use LLM bullshit. not that I can even find that after 6 months of applying and not even a single recruiter followup, let alone any interviews.

But I'm just another off-the-shelf engineer now. I have no domain expertise that another Sr. engineer steering an LLM cannot match.

Funny because I think this is more important than ever. You can produce (ugly) code at remarkable pace now, being able to detect where it goes wrong is a very important skill nowadays.

I still can't let it loose too much because of all the garbage it still produces when given too much freedom. I'm honestly annoyed that my colleagues use it that much on production code, because it's so much meh code with hidden (on the surface) inefficiencies and a lot of avoidable boilerplate and duplication.

Then on the other hand I enjoy that it can create scripts (that are tedious to write) and prototypes at impressive speed. Generally prototyping and research has gotten quite enjoyable because you can stay most of the time on the theoretical side of things, quickly try things out or quickly get info to stuff you don't yet understand enough.

Not sure I agree with the "three pillar" framing in this one. I think it's more like he was able to use the LLMs effectively because he had already built those years of experience. Someone new trying to vibe code their way into designing robust distributed systems is going to need a similar amount of time to build the correct intuitions, because the LLM doesn't give a shit one way or the other, it'll just do what's in the prompt.

On the other hand, commodotizing domains that fit a repeatable pattern isn't so bad, is it? At work I sometimes have to remind people that we're generally just building CRUD apps to interact with data, which isn't exactly at the bleeding edge of software. I like the idea of software engineers getting to be more portable, but unfortunately the industry still thinks LLMs are going to replace most jobs, which is obviously not possible.

I'm still employed and I see myself employed (at least in that company) for a foreseeable future.

Hmm, well, ok. Moving on then

Try to argue with the LLM evangelists, the inevitable brain damage it causes will let you get on disability.

On one hand this is not the first hype cycle, on the other hand the other hype cycles didn't all fade. The inevitable bullshit brought on by vibecoding and shit like that is eventually going to be some kind of problem. People may or may not just ignore that problem, like most other issues in tech.

The world is changing. It happens from time to time. In this case the change is a particularly big one and it's still ongoing, so I can't make any predictions about where it's going to end. But I can be pretty confident that it's not going to magically change back. So my best advice is to try out the new tools, see whether you can adapt to them and use them to improve your own productivity in new ways, and if not then as a fallback start looking at other directions to take your career.

Harsh, perhaps, but the world does as the world does.

The only thing to do is ride the wave of the new technology. You either do that or drown. I've written a literal book's worth of text on how AI should think and behave when writing software and it straight up cannot write code as well as me when left to its own devices, even with all my guidance. LLMs, no matter the size, are structurally incapable of holding all the context needed to write good code in complex systems because they cannot compress complex topics into ideas they've never encountered and reason about the compressed state with their limited reasoning capacity (otherwise known as "learning").

It's a tool, for me atm it's like spellcheck and google had a baby

Time to learn a trade. Sometimes some industries just don't mesh with the changing world.

Am I the only one that has noticed the massive increase in buggy software across almost every domain? Like, EVERYTHING has so many more bugs now. Things just break constantly. AI isn’t one shotting fixing bugs, it’s one shotting making hundreds of new ones.

What he describes is exactly what I’m seeing as well. But I also think his last section emphasizes the importance of having a human steer the ship. If you think LLMs are sycophantic when they talk to you, the Code Review Agents / LLMs are doubly so for their own work “dev bot: you asked me to do x so that’s what I did.” QA bot “I can verify that you asked dev bot to do x and that’s exactly what it did” - Humans have to be in the mix to ensure that a) the code is organized and not recursive spaghetti and b) that the software actually does what you wanted it to do both from the front and backend.

The AI does homogenize some of the playing field, but experts are still needed who understand the intent of the software. I wouldn’t trust an AI dev to conform to all GAAP standards. They can help inform code and even provide the documentation, but ultimately a human needs to validate the deliverables.

And if you thought agents were bad at updating legacy code, just see what they can do with their own code when you ask it to add 50 + features that weren’t in the original request. 🤯

So, let me give some advice to others entering into this industry. The way I see it, there are going to be 2 distinct paths in software moving forward.

1. be customer facing - product management is going to be merged with implementation consulting. These will be the “private eyes” who understand how a customer is using the software, understands the enhancements, the bugs, etc and also trains the customer how to use the software. With how fast AI can now code, it will be interesting to see if we go back to a world where everything is bespoke and disparate (like those homegrown access dnd back in the day) - I certainly hope not or if we’ll stick with platform “COTS” software (think SAS, Intuit, workday, Kronos, Maximo, etc). I’m any case, customers will request more customization because they’ll start to realize it’s now much easier to customize. These “private eyes” will be who informs the developers and helps structure how the updates can work within the platform framework. I think the era of product management and implementation being separate careers is dead.

2. The development architects. These will be the folks who take in the inputs from the customer facing team and determine the best way to build it into the architecture. They’ll have a legion of agents that they control: research, development, QA, technical documentation. Each architect will be in charge of a district product line. You need people who are experts in their domain to do this work. And these people once again become those developer unicorns that were once (and in some cases still are) the backbone of software companies. Your developer will now be doing the work that a team of 5-10 people would have done in the past. But that should also mean that we should be able to employ more of these brilliant architect types.

——

What I struggle with long term, however, is why do we need system software at all? People notoriously hate entering info into systems. They may want to see the data in fancy graphics and infographics, but no one wakes up in the morning and goes “I’m really looking forward to doing all that data entry today.” Humans are also notoriously bad at updating where they are within workflows - we tend to work in the physical realm and update the software as time allows. Agents can and will do a lot of this work moving forward. And they’ll do it (if the systems are built with standards in mind) more consistently than any human would have - a person will just need to validate the data governance. This is how we get to the world of Star Trek, where the computer does the technological lifting and humans do the the human interaction bits. I think (but not sure - do we go back to mainframes???) that databases will still be needed. But software workflows can all be dialed in through training the LLMs. I’m already starting to see it with some of the tools available, but right now it’s very Wild West - at some point once we get the horse back under us, there will need to be standards built in, rather than each person doing it just a little bit differently and therefore getting slightly different results.