Everyone's Using AI Now.
So Stop Competing on the Résumé.
Job hunting & résumé-building in the age of AI. The robots read first, everyone has the same tools, and "polish" is now free. Bring your drink — we're going to figure out what actually still works.
I'm Jorrel — and I've read a lot of résumés.
Founder and operator — I build companies for a living (currently Diem360 and Whale Labs; before that I co-founded a YC-backed startup, AnythingLLM). So I've sat on the hiring side of the table and I use AI every single day. This talk is both.
The important disclaimer: I'm not a recruiter or a career coach. I'm just someone who reviews applications, builds the AI tools doing the screening, and has made most of these mistakes personally — so you can get them at a discount.
Consider this "what the person on the other side of the inbox actually sees," not gospel.
Welcome back to Sip & Learn.
A casual series where we learn one genuinely useful thing over drinks. So far: money, then networking. Tonight: getting hired when the robots read first.
Cadence: Every Thursday in July · 6–8 PM · Bonsai Coffee Bar
- Interrupt with questions anytime
- Screenshot anything (especially the cheat sheet)
- No judgment — we've all mass-applied at 1am
[Drop any event-specific announcements, shoutouts, or sponsor thanks here.]
The Arms Race.
First, the uncomfortable part: the thing everyone's doing to "win" is exactly why nobody's winning.
The job market got an AI dogpile.
Applying got frictionless — one click, GPT writes it, an auto-applier fires it off. So volume exploded on both sides. This is the water we're all swimming in now:
Sources at the end. The point: you're not applying into a room. You're applying into a firehose.
How an ATS actually works.
Almost every company runs your application through an Applicant Tracking System (Lever, Greenhouse, Workday, Ashby…). It's less "AI judge," more filing cabinet with a search bar:
Apply
Your résumé drops into a database next to hundreds of others.
Parse
It's chopped into fields — titles, dates, skills. Tables & graphics garble this.
Filter
Knockout questions, plus recruiters searching by keyword, title, location.
Surface
A filtered, ranked shortlist is what a human actually opens.
"An AI robot read my résumé and rejected me."
Most résumés are never surfaced — filtered out or never searched. It's invisibility, not a verdict. Being parseable + keyword-matched is how you show up at all.
Insider note: I worked at Lever — the "black hole" is usually a search you didn't match, not a bot's judgment.
Which of these have you actually done?
Tap all that apply. No judgment — okay, a little.
tap a few and watch the room light up
AI Job-Hunt — Hall of Fame
Every one of these feels like an edge. Every one of these is now what everybody else is also doing.
It's all the same bug.
When everyone has the same tool, the output stops being a signal. A polished résumé used to say "this person is sharp." Now it says "this person has the internet." AI didn't make you stand out — it dissolved you into the flood. You optimized the one thing that no longer differentiates.
Good news: if polish is free for everyone, the edge just moved somewhere AI can't mass-produce. That's the rest of this talk.
Application roulette.
Spray-and-pray vs. a warm intro. Drag how many roles you'll go after, then flip the switch:
(cold ≈ 2–3% land)
equal this whole pile
Referred candidates interview at ~40–60% and get hired ~4× as often. 100 cold shots ≈ a handful of real conversations.
What Actually Moves the Needle.
Enough doom. Four moves that AI can't commoditize — because they're about proof, not polish.
Beat the robot — honestly.
Up to 75% of résumés are cut before a human ever sees them. The ATS isn't judging your genius — it's matching words. So give it the real ones:
- Mirror the actual job post. If it says "SQL," write "SQL," not "relational query languages." Use their nouns.
- Keep formatting boring. One column, real text, standard headers. No tables, text-boxes, or graphics the parser chokes on.
- Match, don't stuff. Keyword salad gets flagged now — detection of AI/gamed résumés jumped from 53% to 76% in two years.
This gets you past the filter. It does not get you hired. It's the turnstile, not the finish line.
Show the receipts.
Anyone can generate "results-driven self-starter." Nobody can generate your actual work. Proof is the one thing that doesn't come out of a prompt box.
AI can mass-produce
- "Passionate about driving impact"
- "Led cross-functional initiatives"
- "Strong communication skills"
Only you have
- A link to the thing you shipped
- "Cut load time 4.2s → 0.9s"
- A 90-sec Loom explaining your call
Portfolio, repo, deck, writeup, before/after number. One real artifact beats a page of adjectives.
Be referable.
The résumé gets you past the robot. A human gets you past the résumé. This is why last week's "lazy networking" wasn't a side quest — it's the whole cheat code.
Referrals are 30–50% of all hires and get hired ~4× as often. One warm intro is worth more than a hundred perfect PDFs.
Use AI as leverage, not a costume.
- "Write my whole résumé"
- Generate a cover letter you don't read
- Invent metrics you can't defend
- Auto-apply to everything
- Tailor real bullets to one real post
- Research the company & interviewer
- Run mock interviews out loud
- Rewrite your wins to be clearer
Rule: AI should help you say true things better — never help you say impressive things that fall apart in the interview.
Chase the real villains.
- Generic everything — same PDF to 200 roles
- Zero proof of work anywhere online
- No warm intro, ever (applying cold only)
- Metrics you invented and can't explain
- Using AI to draft and tailor (everyone does)
- Employment gaps or a non-linear path
- No fancy degree / no name-brand company
- Applying to fewer roles, but better
Sending 300 identical AI résumés while never asking for one intro is the most expensive kind of "hustle."
Do one thing this week.
Not 200 applications you'll regret by Friday. One move that compounds:
PICK ONE REAL ROLE →
Find one human who touches it.
Send one warm message. Add one piece of proof.
One tailored application with a warm intro beats a hundred cold sprays. Every time.
Depth > volume. The whole market is optimizing for volume — that's exactly why depth is open.
The whole talk on one card.
Mirror the job post
Real keywords, boring format. 75% get cut before a human. This is the turnstile.
Proof > adjectives
A link, a repo, a number, a Loom. AI can't mass-produce your actual work.
One warm intro > 100 PDFs
Referred = 40–60% interview rate, ~4× hires. The human gets you past the résumé.
Sharpen, don't fabricate
Tailor, research, rehearse. Say true things better — never fake things smoother.
Mindset: when polish is free, the edge is proof and people. Aim there.
Intermission. ↻
Refill your drink, stretch, heckle a stranger. Back in ~10 for the toolkit — the exact prompts that make AI actual leverage.
Taking August off. New dates drop on Instagram first — follow so you catch the next one.
@sawtelleneighborhoodcollective
The Toolkit.
The good half of "everyone has AI": used right, it's a personal recruiter, coach, and research team. Here's how to actually drive it.
The prompt cookbook.
Four prompts that do the un-fun 80%. Pick one, copy it, paste your real stuff in the brackets.
Always paste your real experience in. The AI's job is to sharpen the truth, not invent a nicer one.
Turn a claim into a receipt.
Same job, same person — one version is AI wallpaper, one is undeniable. Tap through:
The pattern: what you did → how you measured it → the number. If there's no number, add scope (how many, how often, how big).
Don't sound like a robot wrote it.
Recruiters can smell raw AI now — 76% say they can spot it. The tells that get you insta-skipped:
- "Delve," "tapestry," "leverage synergies," "in today's fast-paced world"
- Every bullet the same length and rhythm
- Big adjectives, zero specifics
- A cover letter that never names a real person or project
- Read it out loud — cut anything you'd never say
- Add one weirdly specific detail only you'd know
- Name the actual company, product, and problem
- Keep your own verbs; let AI trim, not replace
The goal isn't to hide that you used AI. It's to still sound like a person underneath it.
On the menu next.
One useful thing, every Thursday. Vote for what's next:
The AI interview, live
Practicing with a bot, take-home tests, and the new "talk to our AI screener" round.
LinkedIn without the cringe
A profile that gets you found — and posting that isn't insufferable.
Negotiating the offer
The 20-minute conversation worth the most money per minute of your life.
Personal brand & proof-of-work
Building the portfolio that does the applying for you.
Vote for the next topic — shout it out or DM @sawtelleneighborhoodcollective.
What role are you hunting?
Say it out loud. Let's find the warm intro in this room, live — someone here is one hop from where you're trying to go. Then grab another drink.
Sip & Learn · Job Hunting in the Age of AI
Come to the next one.
Follow along, send me the role you're chasing, or just say hi.
Thanks for sipping & learning. Go get one warm intro.
Where the numbers came from.
- 11,000 applications/min & +45% YoY, "applicant tsunami" — eWeek / LinkedIn data (2025)
- ~45% of applicants use AI (Canva survey); 77% of hiring pros see AI applications — Willo, Jan 2026
- 82% of AI-using firms screen résumés with it; ~71% allow some AI-only rejection — ResumeBuilder, Oct 2024
- AI-résumé detection 53% → 76% (H1 2024 → H1 2026); trust gap 70% vs 8% — Greenhouse, Nov 2025
- Up to 75% of résumés filtered before a human; referrals = 30–50% of hires; 2–3% cold vs 40–60% referred interview rates — NBER / BLS JOLTS via refer.me
Figures are directional, drawn from 2024–2026 surveys — the trend, not gospel. Full links in the notes accompanying this deck.