2026 predictions?

Questions
1

Just curious about the community’s thoughts about 2026… seems that all companies want to automate everything with AI now…

Personally, I expect the current idea that AI can fully replace software engineers to start opening CEOs’ eyes—it’s not that simple. AI is already a part of our lives, and I personally use it a lot. For focused, well-defined tasks, it’s amazing. But software development is far more complex than that, and there’s an intuitive, human aspect to it that can’t be easily replicated. AI will continue to improve, but it will also create serious problems in environments where it’s embraced blindly, like full vine coding without much skill, and other scenarios. This is already happening, and until the consequences become more obvious, many CEOs won’t learn that lesson.

On the business side, I have been selling WordPress plugins since about 2005. I have about 50k commercial licenses sold and 100k + free versions all over the world, but this year was bad; sales overall fall 90% even though demand is still there. Everybody is confused about this.

I believe CodeCanyon and ThemeForest—the marketplaces where I sell my plugins—will shut down sometime around 2026. I can’t say exactly when, but it feels like they’re reaching the end of their lifecycle.

Since Hichame Assi became CEO, coming from a very different background, the marketplace has deteriorated rapidly. In just a few years, a series of poor decisions and a complete refusal to engage with authors severely damaged an ecosystem that once worked like a charm!

As a result, many developers are now actively searching for alternatives. One example is wpbay.com, a new marketplace created by two former Envato Elite authors who simply had enough.

About Three.js, it will get better and better, nothing to be worried about this :slight_smile:

On a personal note, health-wise, I really hope I won’t get COVID again. It has literally destroyed me—I’ve had around 32 infections in four years, and I’ve had enough.

2 Likes
2

I think the need for software engineers is gone. With LLMs you can just tell them what you want and get results. Three is also getting incredibly simple to use. Few people used to touch shaders before. Now it’s been democratized and everyone can do it.

1 Like
3

I totally agree. With LLMs and Three, things that used to take so much skill, like writing shaders, are now accessible to anyone. But I also think we as engineers need to evolve and learn how to guide these tools and work at a higher level if we want to stay significant in this new landscape

1 Like
4

Interesting. Personally, the more I use it, the more I realize that strong skills don’t become less important—they make you even better.

When it comes to shaders, good luck explaining what you need and implementing it if you don’t already understand how they work and have a good foundation.

For me, the idea of creating something without understanding it still makes no sense. And this applies to everything, not just software development.

Would you trust your money to someone who has no clue what they’re doing but just “vibe-codes” something you depend on?

Or put differently: would you buy an app—or anything else—made by someone who only knows that “it works” because a prompt said so?

I respect your opinion but I formed mine at this point, lets see if I am right next time at this time :slight_smile:

5

And even if it fails, you can just come here and ask someone with decades of experience. They will just fix the AI slop for you, for free.

I don’t think anyone will need to hire an engineer.

6

My boss got 7 million dollars to build a startup and vibecoded most of it. They’re about to fail, but man, was it a good run. We rented an office for $15.000 a month where we sat for a month. I got pissed because it was in a dirty dangerous neighborhood in SF and just decided to stay home. We were still paying the 15k.
We send an intern to buy stuff at Costco, he forgets milk, so he pays a courier $40 to deliver one carton of milk to us.

Some of the best steaks I had were when I worked there.

Compare this to the pre AI boom startups, frugal, efficient, tons of work…

There’s still a lot of money around, it’s just not being used on engineers.

7

Maybe you right… but you still have to sell that product…

8

As I said, AI is amazing for a skilled developer—it gives you superpowers. But vibe-coding a commercial project without understanding what you’re doing is not the way, even understanding if you let it take controll it will create such a mess in the code that it will be unfixable by any dev or by the agent itself. The more complex a system becomes, the more mistakes AI will make, not to mention the subtle bugs that you need experience and intuition to even notice, let alone fix.

Don’t forget we are engeneers not coders! At least I am.

I honestly expect some really bad things to happen in the near future because of this mindset: “do everything with AI and don’t learn anything anymore.”, by the time this big corporation will wake up the demand for software devs will be so big but it will be a bit to late since nobody is learning anymore… also poeple forgets that coding is not easy is one of the hardest things you can learn and it take years to master it.

I believe things will fall into place as time passes; this is my overall prediction.

9

2026 prediction

the moon becomes the earth and then the earth crashes into itself