The career advice going around right now feels out of touch. That’s the plainest way I can put it.

I spent a weekend a few months back doing exactly what every AI career video tells you to do.

I rebuilt a workflow I’d been running manually, wrote about what I learned, posted about it on LinkedIn, and updated my portfolio with the finished project.

I’ve actually been doing versions of this loop for years now.

I was rebuilding things, documenting the process, and sharing it long before “build in public” was a phrase people use to sell a course. (It’s how I moved into engineering to begin with.)

You know what happened next?

Nothing 😑

Or, if I got “lucky”, a generic auto-rejection 2+ months later for a role I was well-qualified for.

Sadly, this isn’t a one-time story I’m telling for effect. I’ve had versions of this happen across different applications, projects, and months.

It’s mostly silence with the occasional automated rejection that clearly never touched a human.

And I know I’m not alone here because this exact complaint is all over LinkedIn and online forums, almost like a slap in the face to specific advice that keeps circulating from founders, CEOs, and career-advice creators.

Very much like that from a recent YouTube video I watched where seven people running companies with 30,000+ employees answered the same question: who are you hiring right now?

Their answers were consistent, specific, and reasonable on paper. They want someone who can:

• build public demos
• adopt an “explorer mindset”
• master specific AI tools
• become a collaborative generalist
• walk into interviews with an upskilling pitch ready to go

But I’ve done a version of nearly everything they described, and I still hit silence, automated rejections, or nothing at all.

So either I’m the exception, or something about this advice doesn’t survive contact with the actual hiring pipeline.

Why “Post Your Work on LinkedIn” Assumes Someone’s Actually Reading It

In that same episode, Ryan Roslansky, who ran LinkedIn for six years, talked about the platform’s most-requested feature: “show me what a typical career path looks like.”

Only, there’s no longer a path, according to him. That’s why his advice is to post publicly.

That’s supposedly where employers go looking for evidence of what you can actually do, not just what your resume claims.

Now, the theory is sound. What breaks down is the gap between “employers are looking” and “employers ever see it.”

Explore: This Is The Free Browser Extension Your Job Search Needs

The Resume Gate Nobody Talks About

Nobody’s LinkedIn post gets fed through an applicant tracking system. Your resume does.

Before a hiring manager ever gets curious enough to click through to your blog or portfolio, that resume has to clear a keyword filter that has no idea your LinkedIn presence exists.

Note: 65% of employers say AI rejects applicants before a person ever reviews them, often over formatting or missing keywords, per a 2026 survey.

Picture it like this: you write a solid blog post, one that took you a full weekend, and you link it right there in your resume header.

Doesn’t matter. If your resume gets flagged for missing one keyword the posting wanted, that link never gets opened.

Nobody’s reading the write-up you spent a weekend on if the document that was supposed to earn you that click got filtered out first.

Tip 🫣
I don’t understand why most of this advice somehow completely either ignores the resume submission step or doesn’t even address it. Your resume needs to mirror the language of the job posting, keyword for keyword, before anyone gets a chance to notice the interesting stuff you built.

Related: 10-Minute Easy LinkedIn Fixes To Get You Seen By Recruiters

The ATS Myth Isn’t the Whole Story Either

Turns out the algorithm doesn’t really eat everything, since 92% of recruiters actually say ATS platforms don’t auto-reject candidates the way most job seekers assume.

The real bottleneck seems to be volume. High-demand roles are pulling anywhere from 400 to over 2,000 applicants within days.

This means that a recruiter with a stack of 1,500 applications isn’t opening each one and reading it carefully. They’re skimming, rather fast, and most resumes never get more than a few seconds of attention, which is a different problem than the robot eating your resume.

Unfortunately, the outcome ends up being the same for you because, at the end of the day, a human was never going to get to your application, regardless of how good your public portfolio looks.

Silence Is Now the Default

Either way, whether it’s a keyword filter or an avalanche of applicants, the result is identical.

53% of job seekers reported being ghosted by an employer in the past year (as of 2026), the highest rate measured in three years. That’s over half of everyone applying right now, hearing absolutely nothing back.

Not a rejection or a “we went another direction.”

Just silence.

You can build the most honest, detailed write-up of your work in the world, but if it never gets opened, it might as well not exist.

The “Explorer Mindset” Sounds Great Until You Read the Actual Job Posting

Getting past that first filter is only half the fight. The other half is whether the advice even matches what the filter is looking for.

Yamini Rangan, who runs HubSpot, said he’s hiring “explorers,” not “map readers.”

She wants people comfortable running an experiment, forming a hypothesis, and figuring it out without a playbook, because AI has removed the map entirely.

A scientist’s mindset, staying close to the actual workflow, and genuine curiosity about the problem rather than the tool are all good pieces of advice.

If this were the whole story, I would take it at face value.

Explore: 9 Hidden Signals Killing Your Job Search Right Now

The Frankenstein Job Description Problem

However, that solid advice skips the job posting itself, as my most recent encounter with a listing that read like three separate roles got zipped into one proves.

I’m talking engineering responsibilities, product management duties, and people-management expectations, all under a single title with a single salary band. It was enough to make me do a double-take.

Mind you, this isn’t rare, either. It’s common enough that people have started calling these “Frankenstein job descriptions.”

Note 👇
Job postings that ask for deep hands-on engineering skill alongside product strategy ownership alongside team leadership, aren’t one job. That’s a company trying to hire three specialists at the price of one generalist, and an ATS filtering hard on all three keyword sets at once.

An applicant tracking system doesn’t read for curiosity or hypothesis-testing. It reads for exact keyword matches against a rigid list.

It doesn’t care how brilliant your explorer mindset is if your resume doesn’t already contain the six specific tools the posting demands.

The most extreme version of this that went viral this year was the Yahoo job listing that required 10 years of experience with Claude Code. At the time the listing went up, that tool hadn’t existed anywhere near that long.

Ten years, for a tool that’s been around for a fraction of that?

That’s not a hiring team looking for explorers but one that wrote a wish list and let an ATS literally enforce it 🙈

To Be Fair, Not Every Posting Is a Frankenstein

I’m getting a little heated here, so I’ll keep myself in check by reminding us all that the market isn’t static.

Some research is showing that the average tech job posting has actually gotten more specific. The average number of skills listed per posting fell from 30 to 21 between 2024 and 2026, while required years of experience per skill went up.

So, instead of a posting asking for nine different frameworks and a vague “AI familiarity,” you’re seeing postings ask for three specific skills, at real depth.

That’s employers increasingly wanting deep expertise in fewer things, not the sprawling everything-under-the-sun list (which, you’ll see, runs counter to some of the expert advice).

Why Betting Your Career on One AI Tool Is a Bad Trade

Depth in one specific area is one thing. Depth in one specific tool is a different bet entirely, and it’s a riskier one.

Hot take? I think not.

Aaron Levie, who’s run Box for two decades, gave specific tool advice, encouraging people to download Codex or Claude and go deep enough to understand how the agent, MCP connections, and CLI are working underneath, not just how to prompt them.

He called it a genuine leg up for the next three to five years.

I agree that understanding how an agent actually calls tools, reads context, and chains steps together is worth your time no matter what happens to any specific product. It’s why I’m constantly exploring and experimenting.

What I’d push back on is getting hyper-specific about one tool’s prompt syntax, configuration files, or exact quirks because that’s something you don’t control the outcome of.

Explore: It’s Remarkable That We Rely On Models We Don’t Own

Your Employer Picks the Tool, Not You

I don’t pick which AI tool I use day to day for work. My employer does.

That decision comes down to budget, existing contracts, whatever the company decided to standardize on that quarter.

None of that is something I have a vote in.

Say you spend six months becoming the person who knows every configuration trick for one specific tool. But your next job runs on a completely different one.

In that case, the specialized knowledge doesn’t transfer at all. The syntax does not survive a change of employer.

What Actually Survives a Tool Switch

What does survive is understanding why an agent breaks a task into steps, why it needs specific context to avoid hallucinating a wrong answer, and why certain workflows fail regardless of which model is running them underneath.

Understanding why is the actual skill 💁‍♀️

It’s the same distinction as learning programming fundamentals versus memorizing one framework’s exact syntax.

While the fundamentals move with you, the framework-specific muscle memory doesn’t.

Hey! This is largely part of the reason why some of the most recent AI related topic posts on THT are more fundamentals-based rather than strict one-for-one examples of prompt wording or specific structure. I’m trying to practice what I preach in this because a foundational understanding of what’s going on underneath, and what can be implemented across, these shifting AI models, tools and services is more important than telling you what works today for one specific use-case.

The Generalist Advice Contradicts Itself in Real Time

I’ve heard both sides of this advice delivered with equal confidence: on one hand, they tell you to be a generalist, while on the other it’s to avoid having that label associated with you.

For the pro-generalist camp, it’s usually because roles are flattening and companies want people who can move fluidly across functions.

The anti-generalists push back with the argument that it reads as surface-level knowledge of everything and mastery of nothing.

Grant Lee, co-founder of Gamma, described his team’s ideal hire in exactly the first framing. He talked about a designer on his team who also codes, ships end-to-end, and moves across domains without needing a handoff to an engineer.

He called it a “superpower” that lets the whole team run leaner, and you know what? That’s a real and reasonable thing to value in a small, fast-moving team.

But walk that same resume into a corporate hiring process at a larger, more traditionally structured company, and the instinct flips.

A candidate who lists five different skill areas without one obvious specialization often gets screened out. The assumption also flips since they know a little about a lot, and haven’t gone deep on anything.

Note: The generalist who looks like a superpower at a 170-person startup can look like a red flag to an ATS trained to match against one narrow, specific title. Same person, same resume, completely different read.

Though I don’t think that either side is lying about what works for them, I still find the dual consideration confusing.

Because there are no strict or particular guidelines as to which framing applies, job seekers are really the ones absorbing most of the whiplash.

The scarier part is that it seems the higher up you go in internal organizations, the more generalist-favored the advice becomes, whereas the actual roles in the real teams within those organizations are looking for far stricter personas.

You Can’t Pitch “Upskilling the Team” From Outside the Building

Conor Grennan’s “holy grail” of a 2026 interview is to walk in with a specific plan to show how you’d use AI to reinvent a workflow.

Then go a step further and show how the whole team around you could adopt that same change.

In his view, that’s what actually separates candidates more than knowing how to use any specific tool.

Now, this pitch works once you’re in the room.

The problem is the assumption baked into that advice that you get a room at all 😑

No Interview Means No Room to Pitch Anything

You can’t pitch a team-wide upskilling plan to a hiring manager you’ve never spoken to.

You can’t demonstrate a reinvented workflow to a human who never opened your application.

Every piece of advice in that episode assumed the interview stage as a given starting point.

But for a huge share of applicants right now, the interview stage is the thing that never arrives.

The RTO Mandate Makes It Worse

Even when candidates do everything right up to that point, the environment they’re walking into often doesn’t match what they were promised.

Return-to-office (RTO) mandates are accelerating at exactly the moment flexibility has become non-negotiable for a large share of the workforce (including me).

Note: Strict RTO mandates led to a 13 to 14% increase in abnormal employee turnover at the companies that enforced them, per research out of the University of Pittsburgh’s Katz Graduate School of Business.

In a team of 100, that’s 13 to 14 more people leaving than you’d normally expect, in a single year, purely because the company forced everyone back into the office.

And most of them are senior, highly skilled employees. These are the people with the most other options and the least patience for a rigid mandate.

Meanwhile, job vacancy duration increased by 23% and hire rates dropped by 17% at RTO-enforcing companies.

Employers are simultaneously making roles harder to fill and less attractive to the exact candidates who’d be best at filling them, while telling those same candidates to build a pitch about improving the team from the inside 🤨

You can’t reinvent a workflow for a team you were never let into in the first place.

It’s a Wrap

Every person on that YouTube episode was describing something true about their own hiring process, at their own company, under their own constraints.

In a way, the advice going around from founders and CEOs right now isn’t dishonest.

Yamini Rangan really is looking for explorers at HubSpot. Grant Lee really does value a generalist designer at Gamma. Aaron Levie really does want people who understand what an agent is doing underneath the prompt.

None of them were lying or talking about a job market that doesn’t exist.

What they were describing, though, is the view from inside a company that already decided to hire someone. Most job seekers right now are stuck trying to get a single human to open the application in the first place.

What none of them addressed is the layer standing between a job seeker and the room where any of that advice could matter.

Why didn’t they bring up automated filters, sheer application volume, keyword-matching systems, or Frankenstein postings written by someone who never read the advice their own CEO gave on a video?

You can be the exact explorer HubSpot says it wants and still never make it past a filter tuned for a completely different set of keywords.

So, if you feel frustrated by the hiring advice experts are giving online as though you’re somehow not doing any of it already, I get it.

I understand the feeling that all these experts are out of touch either with the real situation or their own internal hiring processes.

I just hope someone opens their eyes before they lose the chance at acquiring the talent they need.

Keep building. Keep growing. For yourself and your ambitions, not for anyone else.

I’ll see ya on the next one.

Bye 🍀