AI Was Going to Replace You. Then the Bill Arrived.

Sponsored by

Bad news is good business. Not everyone buys it.

Every morning, financial news follows the same script. Headlines panic, coverage catastrophises, and somewhere inside the noise is the story that actually matters — the one that tells you where the opportunity sits, not just where the fear is pointing.

Most sources have stopped looking. The alarm is easier to sell.

The Daily Upside was created by Wall Street insiders for readers who crave real insight over recycled anxiety. Five minutes of global business and finance, before the noise sets the agenda — just the facts, context, and analysis your decisions need.

Join 1M readers — including managing directors and principals at some of Wall Street’s largest institutions — who trust The Daily Upside to filter through the chaos.

The upsides are always there. We’ll find them before breakfast.

Start reading for free today

Welcome to today's SCALIS EarlyCareers newsletter! 🚀

If you've been doomscrolling layoff announcements and quietly wondering whether there's a seat left for you, read this one carefully.

The layoffs are real, and a lot of them got pitched to investors as "AI efficiency." That was the headline. Here's the part that didn't trend: the companies furthest out on the AI limb are now staring at the invoice and flinching.

This week Uber's president and COO admitted on a podcast that it's very hard to connect the company's exploding use of AI coding tools to actual features reaching customers. His words: "that link is not there yet." This came after reports that Uber tore through its entire 2026 AI coding-tools budget in four months, after literally running an internal leaderboard ranking teams by how much AI they used. Around the same time, Microsoft reportedly pulled most of its internal Claude Code licenses just months after pushing employees to adopt them, quietly moving engineers to a cheaper option. An Nvidia exec put the quiet part out loud: for his team, the cost of compute now runs well beyond the cost of the employees.

Now, read the nuance, because it matters. Nobody is abandoning AI. Uber is going all in on self-driving. Microsoft is keeping its multibillion-dollar AI partnership. The shift isn't "AI doesn't work." It's that AI just got recategorized from "magic replacement for payroll" to "variable expense that has to justify itself like everything else." And the moment a company has to justify the bill, the thing it rediscovers the value of is human judgment.

That's the market you're job hunting in right now. Here's how to stand in the right spot.

100+ Claude Code hacks to ship code 10X faster

Top engineers at Anthropic say AI now writes 100% of their code.

Are you using AI to write yours?

These 100+ Claude Code hacks show you exactly how. Sign up for The Code and get:

• 100+ Claude Code hacks — free

• The Code newsletter — learn the latest AI tools and skills to code faster in 5 mins a day

Claim your free playbook

There are two job markets right now. Stop applying to the wrong one.

The first market is contracting: rote, repeatable, fully-automatable work that AI genuinely absorbed. Pure data entry, first-draft copy churn, tier-one ticket triage. Applying there right now is rowing upstream against a current that isn't turning around.

The second market is quietly expanding and almost nobody is naming it: the "make the expensive AI actually pay off" roles. When a company burns its annual AI budget in four months and can't draw a line to value, somebody has to own that gap. Quality control on AI output, deciding which tasks are worth the tokens, catching the expensive mistakes before they ship.

Your move today: audit your last three applications. Were you positioning yourself as someone who does the task, or someone who owns the outcome and the tools that produce it? The first competes with a model. The second supervises one. Only one of those is getting more valuable this quarter.

Reposition from "I can do X" to "I make X worth what it costs"

A model can write the email, draft the report, summarize the call. What it can't do is be accountable when it's wrong, read the room, or decide which of three plausible answers won't blow up a customer relationship. That accountability is the product now.

So sell it. Rewrite the top of your resume around judgment, not tasks. Instead of "Managed content calendar," try "Owned content quality and brand voice across AI-assisted production, cutting review cycles in half." Same role. One version competes with ChatGPT on speed. The other gets hired to supervise it.

Go line by line and ask: would a competent model claim this bullet too? If yes, rewrite it to foreground the part only a human is on the hook for.

Read the job description like a recruiter, not a candidate

Postings are leaking the correction in real time. When you see "must be comfortable with ambiguity," "cross-functional judgment," "own quality of AI-assisted output," or a role that quietly reappeared a few months after a team got cut, that's a company that tried to automate something, watched the cost-to-value math fall apart, and is hiring the human back with a new title.

Build a watchlist. Pick five companies you'd want to work for and check their careers pages weekly for roles that fuse an old function with new oversight language. Those rehires are the least competitive openings on the board, because most applicants don't yet realize they exist.

What to actually say when they ask about AI in the interview

They will ask. The losing answers are "I'm scared of it" and "I automate everything." The first sounds obsolete. The second sounds like you'd cheerfully automate your own seat away, and after the year hiring managers just had, that lands badly.

Here's a framing that works. Adapt it to your role:

That answer tells a hiring manager three things at once: you're not threatened, you're not reckless with their budget, and you understand the exact thing their last AI rollout was missing.

Know your leverage, because right now you have more than you think

A company that just learned its AI bet was expensive and hard to justify is a company freshly, painfully aware of what good human judgment is worth. Most candidates walk into offer conversations like they're lucky to beat a robot. You're not. You're offering the thing they just discovered they overpaid to avoid.

At the offer stage, anchor on outcomes you've owned end to end. Ask what the team tried to automate and where it fell short, then position yourself as the answer to that specific gap. That's not desperation. That's reading the moment better than the person across the table.

One more thing: don't measure your own value the way they measured AI's

Here's the deeper lesson hiding in the Uber and Microsoft stories. These companies ran internal leaderboards rewarding employees for using the most AI, then acted shocked when the bill was huge and the value was murky. They optimized for volume instead of outcomes, and it cost them.

Do not make the same mistake with your job search. Sending 200 applications is a usage leaderboard. It feels productive and proves nothing. Ten precise, tailored applications to roles where you can clearly draw the line to value will out-convert a hundred sprays every time. Measure your search the way smart companies are now learning to measure AI: by output that matters, not activity that looks busy.

Babbel's Memorial Day Sale

Longer days + long weekend = more time to learn Spanish! Although you may hardly need it: with Babbel, all it takes is 10 minutes a day. Start learning now and save 65% on all languages in the app, forever.

Start Learning