You've probably noticed how loud and contradictory the AI-and-jobs conversation has gotten lately, doom headlines one week, dismissive shrugs the next. Google just ran the biggest real-world test yet of what AI is actually doing. Its own ATLAS report analyzed 14.65 million actual Gemini conversations across 800 occupations in 150 countries, and the answer is more interesting than the headlines suggest.

AI touches nearly 70% of jobs in some way. It's currently doing about a fifth of the tasks inside any given one. Today we break down what that gap actually means for how your job changes over the next year, plus fresh government data on where hiring is genuinely cooling versus where it's about to open up, a rare AI-adjacent layoff that isn't blamed on AI, and a copy-paste prompt for building your next salary counter-offer.

Let's get into it.

Job openings fell in the sectors you'd least expect. The government's June JOLTS report, out yesterday, showed openings dropping to 7.36 million, missing forecasts, with the steepest declines in healthcare (down 147,000) and leisure and hospitality (down 86,000), two sectors that have been hiring machines for years. Openings in transportation, warehousing, and federal government actually rose (BLS).

Zillow cut over 500 jobs and didn't blame AI. It's the company's biggest round of cuts this year, roughly 7% of headcount, and CEO Jeremy Wacksman pinned it on "disciplined cost structure" instead of the usual algorithm-did-it framing everyone's grown numb to (GeekWire). Refreshing, in a grim way.

Two AI infrastructure startups just raised big and are hiring on the record. Enterprise AI-agent company HappyRobot closed a $150 million round at a $1.2 billion valuation and says it's grown 5x since its last raise less than a year ago. Backend infrastructure startup Convex raised $57 million and explicitly opened hiring across product, engineering, and go-to-market the same day (BusinessWire, Convex).

The EU just started fining companies over AI hiring tools, and it reaches US employers too. As of August 2, any AI system used to screen, monitor, or fire employees is enforceable under the EU AI Act, fines up to €15 million or 3% of global revenue, and it applies to any company screening EU-based candidates, not just European ones (EchelonGraph).

ADP's July jobs report drops today, and the runway into it looks shaky. Hiring has slowed for five straight weeks according to ADP's weekly Pulse data, down to roughly 15,000 jobs a week from over 40,000 in early May (ADP). We'll have the actual number by the time you read this; watch for whether it breaks the streak.

Funding

Contracts

Apple just secretly added Starlink satellite support to iPhones through iOS 18.3.

One of the biggest potential winners? Mode Mobile.

Mode’s EarnPhone already reaches 490M+ users that have earned over $1B, and that’s before global satellite coverage. With SpaceX eliminating "dead zones," Mode's earning technology can now reach billions more in unbanked and rural populations worldwide.

Their global expansion is perfectly timed, and investors like you still have a chance to invest in their pre-IPO offering at $0.52/share.

With their recent 32,481% revenue growth and newly reserved Nasdaq ticker, Mode is one step closer to a potential IPO.

Tap into a $1T opportunity — invest now at just $0.52/share and get up to 20% bonus.

^{Please read the offering circular and related risks at invest.modemobile.com. This is a paid advertisement for Mode Mobile’s Regulation A+ Offering.}

^{Mode Mobile recently received their ticker reservation with Nasdaq ($MODE), indicating an intent to IPO in the next 24 months. An intent to IPO is no guarantee that an actual IPO will occur.}

^{The Deloitte rankings are based on submitted applications and public company database research, with winners selected based on their fiscal-year revenue growth percentage over a three-year period.}

For two years, the AI-and-jobs debate has run on vibes, layoff headlines here, shaky ROI numbers there, with nobody actually measuring what's happening on the ground. Google's own ATLAS report is the first study built from actual usage instead of speculation: 14.65 million anonymized Gemini conversations, mapped against 800 occupations in 150 countries.

The headline number is a gap, not a verdict. AI shows up in some form in roughly 68-70% of occupations. But inside any given job, it's currently handling about 21% of the tasks. Most people using AI at work aren't outsourcing their job to it, they're outsourcing a fifth of it, the parts that were already the most repetitive or research-heavy.

The surprising part is who's using it. Adoption among blue-collar trades, electricians and mechanics among them, came in higher than researchers expected. That tracks with a separate, less-noticed trend: the AI data center boom's real bottleneck right now isn't chips, it's electricians, and Google and Meta have started funding apprenticeship programs to close the gap (Fortune). The technology is showing up in more places than the "AI is a white-collar story" framing usually admits.

This reframes what "AI-proofing" your job actually looks like. It's not a binary question of whether your title survives. It's a task-by-task audit: which parts of your week are repetitive enough that AI already does them reasonably well, and which parts depend on judgment, relationships, or context AI doesn't have. The jobs that shrink fastest are the ones built almost entirely out of the first category.

Why it matters: The next 12 months won't look like whole departments vanishing overnight, they'll look like your job quietly losing its most repetitive 20-30% while the rest gets more valuable, not less. This week, write down the five tasks that ate the most of your calendar last month. For each one, ask honestly whether it's the kind of task Google's data says AI already handles. If two or three are, that's not a threat, it's time back, use it to build the judgment-heavy work that's actually growing in value.

Indeed Hiring Lab found that occupations most exposed to AI, the ones that saw the steepest posting declines from 2022 through early 2026, are now recovering faster than the broader labor market. US software development postings are up roughly 15% since early 2025 while overall postings fell 7% over the same stretch (Indeed Hiring Lab).

The catch is who's getting hired back. Seventy-one percent of that increase came from senior-level postings, and 37% explicitly mention AI skills in the title itself. The door reopened, but mostly for people experienced enough to direct AI rather than compete with it.

That pattern lines up with what's happening in professional services more broadly: firms are getting comfortable cutting mid-level, process-heavy roles while bidding aggressively for people who can supervise the work AI can't fully own yet.

Why it matters: If you're weighing whether to leave a corporate role for something more AI-native, the data says the market rewards demonstrated seniority over raw AI enthusiasm. Before you make a move, make sure your resume shows you've directed or reviewed AI output, not just used it, that distinction is what's separating the postings that are actually growing from the ones that keep shrinking.

Land your next role faster with Tailr, our Claude-powered job search and networking tool: Start searching

Build a negotiation counter-script instead of asking AI to "help me negotiate." Vague prompts get vague advice. This one forces the model to ground a counter-offer in real leverage instead of generic confidence-building. Paste this into ChatGPT or Claude, free tier is fine:

"I'm evaluating a job offer and want your help building a negotiation counter-script. Here's my situation: [paste offer details, base salary, bonus target, equity/RSU value and vesting, title, location, and your current comp]. The role is [job title] at [company/industry], based in [city]. Do three things: (1) Estimate a realistic market range for this role/level/location using public comp benchmarks, and flag anywhere you're uncertain rather than inventing precise numbers. (2) Identify the 2-3 strongest, most defensible negotiating points based on my background: [paste 3-4 bullets on relevant experience, competing offers, or unique skills]. (3) Write a 120-150 word counter-offer email that names a specific number, cites one piece of leverage without overreaching, and proposes a friendly next step."

Scientists found an immune cell that kills by exploding itself — Researchers studying flatworms identified "ruptoblasts," a brand-new immune cell type that self-detonates to kill dozens of surrounding bacteria at once, the fastest form of cell death ever documented. (Stanford Report)

A 2,500-year-old statue survived because Romans buried it face-down as a paving stone — A remarkably preserved marble statue from ancient Sardis owes its condition to being accidentally protected under a Roman street for 2,000 years. (Daily Sabah)

Your judgment doesn't peak until your late 50s, even though processing speed peaks in your 20s — New research separates raw cognitive speed from wisdom and emotional regulation, and they peak decades apart. (Beautiful Minds)