🦥What's a good project in 2026?

Hello nerd

Welcome to another goofy Sloth Bytes. I hope you’re having a fun week.

That investor was wrong. Gamma is now worth $2B, with 50M users and more than half their growth driven by word of mouth.

They're one of 6 AI-native startups in HubSpot for Startups' free Bold Bets Playbook. Replit grew revenue 50x after half the team pushed back on the strategy. Ramp generated 100M+ views from a single stunt. Clay's co-founder wouldn't hang up a sales call until the prospect DMed him in Slack.

Each one took a GTM risk most founders would never greenlight. Each one paid off.

Get the free Bold Bets playbook

Speak your prompts into ChatGPT or Claude and get detailed, paste-ready input that actually gives you useful output. Wispr Flow captures what you'd cut when typing. Free on Mac, Windows, and iPhone.

Try Wispr Flow free

I remember the days when building a basic to-do list, tic tac toe, and a weather app got you an interview with Google.

Like genuinely, if you showed up to an interview in 2016 -2020 with a working chat app that had JWT auth, a PostgreSQL database, and real time features with WebSockets, people thought you were the greatest programmer of all time.

What a time right?

But nowadays it takes a whole company to get a recruiter to spend 3 seconds on your resume.

I mean to be fair, you can basically vibe-code entire projects in an afternoon.

Not even a focused afternoon too. You could be brain rotting on TikTok and it’ll be decent. AI coding tools have gotten so good that these projects just aren’t impressive anymore.

I mean, imagine how many to do apps and weather apps these models are trained on? It has to be thousands of them. Those are like the classic default resume projects.

So I started thinking. If a to-do app won't cut it, and a full-stack app with auth and real-time features won't cut it either, what does a "good" project even look like now? Do you have to build an entire startup just to have a shot?

Almost a year ago, Gergely Orosz from The Pragmatic Engineer spoke with 30 hiring managers and 3 recruiters about the job market. This is the numbers and info they shared (kinda crazy):

What’s interesting is that even though companies get a lot of applicants, they’re are struggling to find good engineers.

Everyone is applying, yet almost nobody is standing out.

This new standard for projects DOES NOT apply to everyone.

It’s hitting certain areas much harder than others.

This is where it’s been affected the most.

These are the areas where AI assistants are strongest, where boilerplate is basically free, and where a finished product looks the same whether you wrote every line yourself or had Claude Code do 70% of it.

These areas aren’t affected as much. AI is genuinely weaker there. The knowledge you need is more specialized so the barrier hasn't collapsed the same way.

It’s still not an area where people can “vibe.”

Unfortunately, most people reading this want to be a software engineer.

And software engineers are in that affected group….

If you search LinkedIn right now for "software engineer." It's overwhelmingly web services, APIs, cloud infrastructure, mobile apps, data pipelines.

The exact bucket AI hit hardest. So yeah, this sucks.

Companies are starting to include AI tools in the interview. They want to see how candidates use these tools and not whether they could write boilerplate from memory.

The ironic part is that once you're inside a company, slop is kind of fine. Ship fast, move fast, iterate fast, but the second you try to do that during the hiring process?

They hate it.

AI-tailored resumes are being rejected on sight and candidates are getting caught using AI tools during live coding interviews.

The thing they encourage on the inside is the thing getting people filtered out on the outside. It’s definitely unfair. Your best best is to use AI to build things, not to fake things.

If you have strong feelings about using AI to build your projects or for writing because it feels like cheating, I respect it. I really do, but at least understand you're competing against people who don't share those feelings.

They will do whatever it takes to get a job.

Create a landfill of code slop, work for less money, or even get on their knees and give the CEO slop.

Trust me, when you’re desperate, you’ll do anything.

Don’t make it a fake problem and don’t create a problem to justify the project.

If you do that, then you actually have some talent in business.

The smaller and more specific the problem, the better.

When someone says "tell me about this project," you have a real story. It’s not something that you would find on youtube.

Now of course the complexity would have to be on the higher end. So something like:

A running music player that auto-adjusts playlist tempo to match your cadence using your phone's accelerometer. Spotify exists, but Spotify doesn't do this. You wish they did, so instead of waiting, you built it.

A dependencies scanner that crawls your package.json or requirements.txt, tells you what each package actually does in plain English, flags ones that haven't been updated in over a year, and shows you which ones could be replaced with native browser or language features.

You’re welcome for those ideas btw.

The thing these two projects have in common:

One is a code exercise. The other proves problem-solving, product thinking, and real-world impact. Same technical complexity. Completely different story.

This is something most people are probably not doing.

Now, companies hate AI-tailored resumes, but what they're really rejecting is people who use AI to fake qualifications they don't have.

That's what gets flagged and filtered. That's not what this is.

You actually have the skills. Which means tailoring your projects to match a company's stack/requirements isn't dishonest. It's just smart.

Building a project foundation used to take weeks. Now you can get a “working” version in a day. Which means you can now be be more strategic.

Then do the work AI can't: build the unique angle, make the real decisions, own the architecture.

Then ask it: "here's my project and here's the job description, what's missing, what matches, what would make this stronger?" That analysis used to require years of experience. Now it takes thirty seconds.

The result is a project built specifically for that company, in their language, using tools they already use. And you can genuinely talk about every part of it.

Technically, nobody's verifying your projects before the interview. Recruiters aren't cloning repos during resume screening (if they even know what a repo is).

Nobody is looking at your projects before they email you or call you.

So technically… you can lie about the projects.

People have done this and even gotten interviews, but uh it’s a pretty risky move:

If these situations happen and you lied about your projects:

I promise you, making up technical decisions is NOT easy. I don’t care how good of a liar you are, you can’t make that up like you can a normal story. Experienced people will know you’re making stuff up or have no idea what you’re talking about real quick.

If you think about it, you could lie on your resume, get the interview, AND THEN start building the project…

Think of it like a loan or a payment plan (I’m still NOT recommending this, but hear me out)

You now have days, sometimes weeks, before that interview.

You can use AI to spin up the project fast and then refine it by making real decisions on top of it. You’ll probably hit some bugs and now you have a story on how you struggled and fixed them

TECHNICALLY, you're not lying anymore. You actually built the project even if the timeline is a little messy.

Once again. NOT RECOMMENDING THIS.

Let’s be honest. It’s always been who you know and not what you know.

So remember to do these things:

Here’s one strategy you can do to make your project helpful for networking:

Now you're not begging for a job. You're a curious developer who clearly did the work and wants to learn from the person who originally made it. A great conversation starter and that person will definitely help since they probably spent days or weeks on that problem.

Projects are going to matter less as proof you can code, and more as proof you can think like a developer.

The question that's going to become more important is not "what did you build" but "why did you build it this way."

And with over 36 million new developers joining GitHub in 2025 alone which by the way, is the fastest growth in the platform's history.

There's going to be a lot of people coding, but the ones that get hired will be the ones that can explain it

That’s all from me!

Have a great week, be safe, make good choices, and have fun coding.

If I made a mistake or you have any questions, feel free to comment below or reply to the email!

See you all next week.

If your company is interested in reaching an audience of developers and programming enthusiasts, you may want to advertise with us here.