TL;DR

• AI’s everywhere, but it’s just a tool—dreams and goals matter more.

• AI Engineer = Software Engineer with AI tools, same as JavaScript Engineer. Don’t stress specializing.

• Front-end got boring for me, so I’m branching out. Open source still rocks (shoutout Tanner Linsley!).

• Job market’s tough—be a Swiss-knife coder, not just front-end.

• You’re as awesome as Kent C. Dodds. Build that thing you’ve been dreaming of! 🚀

Hello! Another week, another Friday!

If you’re following me on social media, then you might have noticed that I’ve been talking a lot about AI.

That is because I was assigned to a project where we’re integrating AI features into our existing platform. So I’ve been paying a lot of attention to AI blogs, tutorials, concepts, libs and frameworks… all AI things basically.

I can’t say I was an early adopter though. I only started to get interested in AI when ChatGPT was getting a lot of attention.

So this is when I started looking into what it takes to have AI as a part of a tech stack.

Everyone is trying to do AI, every platform is trying to sell us their product, promising us productivity, efficiency, and promising that we’ll achieve all our dreams. LIVES will be changed.

That’s a lot of promise, and yet…

You see, because it doesn’t matter what AI can do. If you have no dreams or goals, then AI is helpless.

Nonetheless, the AI craze is going on full steam. Companies are hiring AI Engineers. YouTube is full of LangChain, LlamaIndex, MCP, and other tutorials, which is great, don’t get me wrong, YouTube is how I became a software dev.

On the other hand, we’re still getting videos and courses such as this one. People are still interested in the basics, in software engineering.

Because an AI Engineer is just a Software Engineer who works with AI tools.

Same as a JavaScript Engineer is a Software Engineer who works with JavaScript.

So what does it mean to be an AI Engineer? That just means you’ll be working with a different set of frameworks and libraries.

Say, MCP, for example. Everyone’s talking about MCPs. Let’s talk about them, what are they?

Just a running server that an IDE or Claude Desktop launches instead of you doing python run main.py.

It’s harder to debug, but all in all, it’s not different from creating the FastAPI service or node.js express one.

Same .env variables, same if/else statements, same REST API calls and/or database calls.

Same shit, different library. You’re not missing out. Don’t be afraid to make a demo or do what I did, make a demo then repurpose it for something niche and what only you may need.

Should you specialise in something today to get a job, etc.?

It’s a mixed bag. Some say that only niche and specialised people get hired. Others - and this is what I’m seeing - say companies want “Swiss Army knife” types of people who can do a bit of everything.

I started my career on the front-end. At that time, I thought this is where I was gonna stay and would heavily specialise in that. I liked front-end… until I didn’t.

I mean, I still do, I just find it not as exciting as I found it before. I think I went through all the front-end till the end game on easy to moderate mode. That’s the mode that the majority of companies are operating on.

Right now, no matter what company I go to, the type of work I’m doing is always the same. Same React components, create a design system in Storybook, optimise Lighthouse scores, etc.

If you want to go for “beast”/”impossible” mode, then you get the kind of jobs that require 3D rendering with Three.js, maybe, or you go to FinTech, where they need to have real-time live updates of some prices, or you go to a data-focused product, where you’d need to deal with a bunch of optimisations, dealing with a constant stream of data, or some sort of data manipulations right on the screen.

In that case, you become a super front-end pro.

But then why would you? The number of jobs for this is minimal. Once this startup goes broke (and considering the market, chances are it will), you’re out with everyone else looking for a mediocre job where you need to paint pixels on the screen.

So I don’t know. I didn’t want to go full front-end pro and decided to branch out.

Small detour, there is a part of the front-end world that still gets me excited - open source and all kinds of projects it has: useful, essential, unruly, weird, all kinds!

I have tremendous respect for people like, and especially, Tanner Linsley.

I even considered sponsoring him before I had to leave my job and accept a pay cut. (Sorry, mate!)

He’s doing a great job at creating competition for Next.js with TanStack Start, I’ve been using React Query (now TanStack Query) on almost all of my projects, and I just discovered TanStack Pacer. Next time I need a rate-limiting util lib - I’m using it!

Tanner is great! I hope he gets to continue his open-source project for a very long time.

Back to our topic.

I am trying to stay up-to-date on the market and how much money I can get.

Recently I clicked on Gergely’s site, TechPays, when he mentioned it in one of his posts, where he was talking about AI crawlers that increase his hosting bill (it’s a thing!)

So I checked out his project, probably for the first time, and I found out that people with higher salaries are more likely to share their salary.

Or I’m just looking for the job in the wrong place.

Either way, if you sort by total amount, and scroll through the first couple of pages, you’ll see that front-end isn’t mentioned anywhere.

Yes, it’s not a complete picture, there definitely are people who make tons of money on the front-end, and specialise in something niche and specific.

But there’s still truth in that: if you’re planning on being employed, I’d say focusing on just front-end isn’t enough.

But hey, that’s just my opinion. 🤷

And also, seems to be the opinion of DHH and their Signal37 company that pays juniors $145,849. See, they are not hiring specialists. In fact, for their junior position, you MUST have less than 3 years of experience.

Why? DHH says it’s because those who have 5 or more years are already growing and developing in a “predictable trajectory”.

And those who have less than 3 years of experience may have a “promise”.

Does that mean if I didn’t make it to FAANG in my first 5 years, then I’m a lost cause? 🥲

Anyway, when deciding your “career trajectory”, you should know that you’re as awesome as Kent C. Dodds! Even if you have more than 5 years of experience and you’re not at FAANG.

Recently Kent replied the other day to one of the others’ posts where he said

He is doing a journaling app. How hard is it to make a journaling app? (Remember, we’ve established that MCP is not much different from any other tool/framework, it’s nothing you can’t learn!)

And Kent, i.e. one of the BEST developers out there, is doing such a trivial project like a journaling app.

He could’ve been debugging or migrating a gigantic legacy project for a REALLY hefty fee (does he ever do that, by the way? Does he do contracts like that?) and yet, considering all his knowledge (power!), he’s doing an app that you can do.

You’re as awesome as Kent. Go do/start/finish what you wanted to! You got this! 🚀

That’s it for this week. I feel like this newsletter was all over the place. I’ll try to get better next time. 😀 

Cheers!

🎓️ Top Learning Materials of the Week

📺️ Building Effective Agents with LangGraph - A solid YouTube tutorial for diving deeper into AI agents: Watch here

🔗 MCP - The official Model Context Protocol docs. What can be a better learning resource than official docs?!

⚒️ Tools of the Week

🧑‍💻 Trae - My now go-to code editor, Cursor competitor, who recently released new AI models, introduced MCP integrations and all of that for FREE (for now)!