Stop Paying for Permission

That conversation stuck with me. I kept thinking about how, in the debate over skills vs credentials for engineers—in engineering, AI, software, all of it—it’s easy to confuse learning with credentialing or access. I used to think structure was the thing holding me back. If I’m honest, I’ve spent too much on programs just for the structure and the signal, not for the knowledge itself. It’s a pattern that wastes months, drains funds, and leaves portfolios empty while chasing permission instead of skill. Knowledge is everywhere and cheap. What costs is the feeling that you need to buy credibility.

Skills vs Credentials for Engineers: What Are You Actually Buying?

Here’s the question I wish more people asked themselves before clicking “Enroll.” Are you paying for knowledge, or are you paying for something else entirely? It’s easy to tell yourself you need to “learn more.” But the real decision—the one that decides whether your money and time turn into real progress—is about what, exactly, you think you’re buying. Let’s get direct. Most education pricing isn’t about the facts or the technical skills. It’s about everything wrapped around them.

Look at the landscape for a second. In the skills vs credentials for engineers debate, degrees and courses feel similar, but they deliver different results. A degree still commands respect—59% of employers give more weight to degree requirements now, and 70% haven’t dropped them at all. That isn’t true for most one-off courses. A shiny certificate alone isn’t opening doors; recognition comes from the signal a credential sends, and not every program carries the same weight. If you need the label, sometimes you do have to pay for it. But if you’re after actual skill-building, sometimes the price tag is a distraction.

Honestly, the cost of pure information has crashed. You can build skills without a degree because nearly everything can be learned for free or cheap. These days you pay a premium for curation, accountability, and the keys to gated rooms.

It’s completely normal to worry that taking a do-it-yourself route will leave you short on credibility. Maybe you don’t trust your own discipline, or you’re convinced that only formal paths get you seen. I’ve had those doubts. Most ambitious people do.

But here’s the kicker. Overpaying for structure usually just delays your first real project and chews up precious cash. Take project-based skill building—its impact on outcomes sits at 0.44 effect size, which is meaningful. No class beats actually starting something. Action creates real learning.

The 90-Day Plan: Spend for Impact, Not Permission

Start with a pragmatic decision framework. Before you drop money or hours anywhere, nail down what delivers the progress you actually want. Is it skill, a credential, structure, or a network? That’s your filter. Force yourself to pick just one main value for this sprint—then go hunt the smallest, least expensive option that provides it. Don’t buy the whole kitchen when you just need a spatula.

First up, lock in what you want to prove by day 90. Is it a working prototype, a demo, a small contract—or just getting shortlisted for a role? Pin down the “evidence.” What would you show a colleague or post online that makes the skill real or qualifies you for the next step? Make it as clear as sending a link to your repo or dropping a two-minute video.

Next, map your chosen value type to its cleanest options. If you need knowledge, get scrappy: pick a hands-on project, grab open docs, grab a focused YouTube series, even do a self-guided Udemy course if you must. Skill comes from making, not just watching.

If what you need is structure, join a cohort with weekly touch-ins, build a calendar with small deliverables, or use a peer accountability tool—don’t pay more for a syllabus than for a challenge. If network is the missing gear, drop into active Discords, sign up for an indie hackathon, or ship something and ask for feedback—the kind of lessons you can’t Google. For a credential, research who’s hiring and what signals actually get filtered. Chase just enough formality to unlock the gate—but always pick the cheapest route the job deems credible. You’re not obligated to care about the prestige, just the bar you need to clear.

Now put it on the calendar. Break your 90 days into weekly goals. Plan for two build cycles (say, one rough draft, one real attempt), a public demo that tells a story towards the end (to friends, peers, or a hiring manager), and at least two mentor or expert check-ins—where you actually share work, not just talk theory. Cap your budget so that every dollar is attached to getting closer to the outcome. If spending $99 gets you the proof you need, don’t pay $999 for fluff.

Last step: tie in some failsafes. Write out an accountability contract—send it to a friend or post it somewhere you’ll be embarrassed to back out of. Use a public progress log (Twitter threads, a Notion database, whatever fits). Schedule a mid-point review around day 45. Ask yourself if the plan’s working. If it’s off, pivot. Take one night this week to audit whether you’re chasing the right target, before you’re halfway through the budget and still nowhere close.

Not gonna lie, sometimes my own plans feel solid on paper, but half the battle is sticking with them once the first bump hits. I haven’t fully solved for that yet. Sometimes I double back and tweak goals after week three, and maybe that’s alright—progress rarely moves in straight lines.

This is the part almost nobody teaches. If you figure out what’s really required—and you stick to proving it, building it, or showing it—you’ll move ten times faster. Not because you’re smarter, but because you stopped paying for permission. That’s how progress actually happens.

Portfolio over Permission: How to Use the Framework for Real Progress

Let’s make this concrete. Say you’re a software engineer itching to break into ML. Here’s what the default advice tells you: shell out for a certificate-heavy course, dutifully plod through the lectures, hope it lands with recruiters. But if your goal is skill, you’re better off getting your hands dirty. Build a tiny classifier—predict dog vs. cat images, or whether a tweet’s friendly. Use a resource like fast.ai for guided, practical learning, or dive straight into the PyTorch docs if you prefer to puzzle it out yourself—and practice using AI for coding efficiency as you build. Find a mentor or a peer willing to review your code; that feedback loop is worth more than any Udemy badge.

Then, write up your process: what worked, what stumped you, where you hacked through messes. Publish it—on your blog, as a GitHub README, or even a LinkedIn post. Only consider paying for a micro-credential if you know credential signaling in tech means your target employer auto-filters resumes for it. This flips the script: don’t spend unless you have a specific gate to clear, and always tie the investment to the narrowest requirement.

Here’s another common scenario—a platform engineer gunning for Staff. At this stage, the real measure is skills vs certifications—not another AWS cert. It’s real systems design and the ability to rally other teams. Instead of dropping cash on more courses, draft an internal RFC for a hairy problem, pilot a performance improvement, then present those findings to your org. What moves you up is network—real cross-team credibility—not one more certificate on your resume.

Now let’s talk about founders. When you’re dreaming of launching a product, you don’t need thousands for that, despite what accelerator websites whisper. Your first move isn’t to buy access—it’s to interview five customers (cold emails work just fine), ship an MVP inside two sprints, and join a local founder circle for rough feedback. Pay for an accelerator slot or legal advice only if it unlocks customers or capital you can’t get solo.

But yes, sometimes you do need to pay—and pay wisely. Some domains are locked down—only licensed engineers can sign and submit official drawings, according to NCEES. If your outcome demands a regulated credential, an immigration visa, or an executive badge, pick the smallest, shortest program that legally unlocks what you need. Buy the key, not the castle.

Make the Decision: Audit, Checklist, Action

Take sixty seconds—right now, not tomorrow. Write down what you want to have or to do in ninety days. Don’t dress it up for anyone else. Then mark what you actually need to get there: is it more knowledge, a credential, a repeatable structure, or access to a network? Be brutally honest. “Are You Paying for Knowledge, or Are You Paying for Something Else?”

Here’s your checklist. For the next three months, design your plan around these moves: one solid portfolio artifact (real work you can show), one mentor or guide (even if just reviewing your code once), one community (a Slack, a Discord, anything with feedback cycles), one public demo (doesn’t matter how rough), a fixed budget cap (money you won’t go beyond), and a mid-point review at day 45 to check if you’re actually on course or lost in the weeds. Run that list for every outcome you chase.

Still on the fence? I get it. You worry your work won’t look credible, or that you’ll slip without some official guardrails. But here’s the deal—credibility comes from clear outcomes and references, not just certificates. Discipline isn’t magic. It grows from constraints you set and commitments you make visible. Recognition flows from what real users say and what you actually ship.

So make your next bet count. Invest in formal signals only when they unlock gates you can’t force open yourself. For everything else, the fastest path is this—build, share, and iterate. A degree can open doors—but in the skills vs degrees trade-off, are a couple of expensive courses really worth the cost?

That conversation from Tuesday keeps echoing in my head. I don’t have a perfect answer for everyone. Maybe I never will. But if you keep your eye on what moves you forward instead of what looks impressive, the rest usually sorts itself out.