In 2025, Coding is Changing Forever

🔥 STEF'S DEVELOPER BOOTCAMP AND MENTORING PROGRAM

➩ Private Consult with Stef for up to 20 minutes:

Ask me any questions about career, business, freelancing etc ...

🎤 Listen to my Uncle Stef podcasts:

🔥 JOIN STEF'S 'CODER'S CAREER PATHS' NEWSLETTER:

🔥 FREE: LIZARD WIZARD KOMODO - TRANSFORMATIONAL MIND TRAINING:

*************

POPULAR & EASY CODING COURSES:

*************

🦎 Lizard Wizard Course:

📚 BOOKS TO READ:
… Complements Studioweb courses on HTML5, CSS3 and JavaScript.

The Naked Ape:

✉️ STAY IN CONTACT:
Stef's social links:

Stef's business channel:

👉 GOOGLE REVIEW:
Leave a Google review about Stef.

MY MOUSE & KEYBOARD:

SUPPLEMENTS THAT WORK AMAZING FOR ME:

Protein Essentials Beef Gelatine Powder:

... Healed my very bad knee. If you have joint problems, this *could do miracles for you.

Webber Naturals 88862 Glucosamine Chondroitin

MY CAMERA GEAR:
Canon EOS C70 Cinema Camera

Thanks!

Stef

#mentoring #codecourses #unclestef #codingcoach

[agge]

I don’t really get this AI hype. To be fair, LLMs are really a better search engine. They haven’t brought anything to the table which you couldn’t do before. The only difference is that it’s way more efficient to use an LLM than to search.
I have been coding for over 15 years and the way I have been coding has always been kind of a copy paste from Google or stack overflow. You take some building block there, some other blocks here and then together with your deep understanding you build something. This is really the same way I use LLMs today. I ask it for building blocks. Together with my understanding I can then build something useful, and it’s way more productive than before, but only because it’s such a better search engine.
In general I disagree with you uncle Steff. I don’t think you need to start using these technologies in order to not get left behind. These technologies will change over the years. In 3 years there is a new way of doing things. What really matters is your skill of solving problems. Don’t believe the hype that LLMs are so much better than humans in coding. It’s just because they benchmark them in solving problems that kinda exists online. It’s like saying that Google is so much better than anyone because it knows everything. Understand what an LLM is, learn to think and use your brain, learn to code. You will be far ahead of any vibe coder in the future

[soufianedjeghim]

An engineer who has a good command of programming basics is the one who can take advantage of the potential of artificial intelligence.

[eduardobarron]

i'm already sending my cv to mcdonalds

[TrendingBoss_]

AI is making coding faster by helping developers with suggestions, generating boilerplate code, solving common bugs, and even building simple applications — but it cannot replace real developers who understand what’s happening under the hood.

If you don’t understand the code AI gives you, you’ll be stuck when:
• Something breaks.
• You need to customize things.
• You need to optimize for performance.
• You’re building something unique, not just another “basic website” or app.

Knowing how to code is still extremely important.
AI is more like a power tool — it can make your work faster and easier, but you still need the skills to use the tool properly. Otherwise, you’ll build a weak structure that collapses at the first problem.

[aleksandarstanisic]

Coding, HR, 5 rounds of tests everything is shit now.

[doriancerutti]

People overcomplicate this way too much. Yes, AI can help—but not in the ways those LinkedIn “AI gurus” claim.

It’s still incredibly important to build real, hard skills and understand what you’re doing. Not just because AI still makes huge mistakes, but because knowing your stuff actually makes the work more enjoyable.

AI is basically a glorified search engine. Instead of links, you get refined results. Generating code isn’t its strong suit unless you’re working with super popular frameworks or languages like React or TypeScript.

Where AI really shines is with words—brainstorming, direction, summarizing, and documentation. It handles that stuff really well.

Put that in the hands of someone who already knows how to code? Their productivity will skyrocket—and they’ll learn even faster.

[BillyRyan]

As a developer I'd still learn how to code and still write code but will consult AI for ideas and improvements just like we did with Google

[shrunkensimon]

And what happens when the LLM bubble pops, and they are forced to bump up the prices massively? They are operating at billions net loss already.

[muradcaucasus]

When all this AI hype will end. Learning AI over programming is wasting time.

[JohnSmith-gugl]

Learning AI is like learning Excel.
Learning how to code is like learning Math.
One is a soft-skill, the other is a hard-skill, but you need both! I would say 20/80.

[davidbasil]

Learning AI will lead to nowhere. One should study and do what is hard, complex and have high barrier of entry.

[ChicagoJ]

Low code and no code has its problems. It takes time to learn the platform. There is a learning curve to it. AI is going to replace most of this. Not right now, but eventually it will be easy to do the same with ai.

Precision coding won’t be replaced. When you are engineering software and you have specific requirements for everything from UI to how your applications are coded, coding will be required. It’s faster to code than to prompt in those situations. But AI can enhance your productivity. Combining coding skill with AI is the key.

[MaxGreen-gb]

I love to write code, but writing prompts is so boring...

[xqwerty]

Find a problem, check if it's something people would pay for to be solved easily, learn the tools, build it, learn marketing, sell it, grow it, praise God often.

[FreeWanderingThinker]

I've already started to farm. Potatoes and zucchini are in the soil. Tomatoes are going to join them soon.

[radupopescu]

AI is a marketing tool used for two things: 1)get more money from clients saying that 'oh, we have this AI tool, bla, bla';2) lower the IT industrry's salaries. You won't write an enterprise solution with AI.

[codeRS]

Job market is bad. But, not because of AI. Job market will soon improve, once economy gets stable. AI not replacing developers jobs any soon.

[Never-mind-k]

The debate between machine-generated code and human-written code is no longer theoretical—it’s a reality. I was skeptical until I witnessed the efficiency of self-driving cars compared to human drivers. If machines can outperform humans in complex tasks like driving, why should coding be any different?
😮

[nighteyes]

Learn how to leverage AI in the PROPER way is the most important skill you have to learn in 2025.
That means, knowing when to prompt and when NOT to. And when doing it, knowing HOW to accurately get it to get work done as quickly and efficiently as possible.
But you also need to learn WHEN prompting can actually slow you down because the AI can not handle it as good as you would and might start spitting out non-fonctional code / hallucinate (ie. very large context with intricate business side).
Althoughit's worth noting that this last part tends to disappear as these models improve.
Now, all things considered, you can decide to be a lazy engineer and never review the code the AI has produced to get a feature done or a bug fixed (=> vibe coding), launch it in production and never look back.
But not everyone is a unipersonal micro-saas CTO that pumps out a garbage website every month, some people have jobs where the company relies on their expertise and need them not to deliver a change in their software that will not compromise the company's reputation.
That means, reviewing the code, whether it be a junior's or an AI's code.

[aerocodes]

I think the only difference between the past new tech and AI, is that AI actually writes the code. I feel like there's a long way until we can connect the dots between outputting a blob of text that may or may not compile to an actual product with various pieces in it. At the same time it's worth noting that the barrier to entry will for sure be lower, and that's a great thing I believe. At the same time the people that actually studies CS, software architecture and so on will ultimately have an edge here. Think of it as asking to the 'genie in the bottle', if you don't know what and how to ask, then you will most inevitably fall in the traps that a CS course is trying to teach you not to. That is not to say that in 10, 20 years AI get's so good that it will handle those for you as a SWE engineer would. I feel they are certainly going in the right direction, but even then, who will be the ones using up this tech? It's all good "vibes" when you are prompting your way in, not so great when you realize you have created a system that simply doesn't hold it together.
🖖