Straight answers

Because most current layoffs are not distress — they're restructuring. Three forces stack: (1) correction of 2020–21 over-hiring, when big tech grew headcount 40–60% in two years; (2) investor pressure for 'efficiency' after 2022's rate hikes ended the free-money era — Wall Street now rewards layoffs with stock bumps; and (3) AI reallocation, where companies cut in one place to fund AI hiring and datacenter spending in another. A profitable company cutting 5% while posting record earnings is doing portfolio math, not survival math.

For visible, remote-friendly postings — yes, routinely. LinkedIn's own data and recruiter surveys through 2024–25 show popular postings drawing 300–1,000+ applications within days, partly because AI tools let candidates mass-apply (some send thousands of applications), which floods every opening and forces employers to lean on AI screening — an arms race that makes the front-door application channel close to useless. The practical takeaway isn't despair; it's that referrals, direct outreach, and being findable (portfolio, open source, content) now dramatically outperform applying cold.

In 2020–21, tech companies extrapolated pandemic demand (e-commerce, streaming, remote tools) into a permanent new normal and hired accordingly — Meta nearly doubled, Amazon added ~800k roles, Zoom, Peloton and Shopify scaled for a world that partially reverted. When demand normalized and interest rates rose in 2022, that excess headcount became the first thing cut. It matters now because it trained companies and investors that big layoffs are survivable and even rewarded — lowering the barrier to the AI-driven cuts that followed.

Across 2022–25, the most affected functions were recruiting/HR (hiring stopped, so the people who hire went first), customer support (first target of AI agents), manual QA, routine content production, data entry and back-office operations, middle management (the 'flattening'), and entry-level roles across the board. The most protected: security, AI/ML engineering, data engineering, senior engineers with system ownership, and anything combining technical skill with regulated-domain expertise.

Both, and the mix is shifting. In 2022–23, 'AI' was mostly narrative cover for over-hiring corrections. By 2024–25 the direct cases became undeniable: Duolingo's translators, Stack Overflow's traffic, Chegg's business model, Salesforce's 4,000 support roles, IBM's unfilled back-office positions. But the biggest effect is quieter than layoffs: jobs that don't get created. A team that would have hired 10 hires 6; the missing 4 never appear in any layoff tracker. That's why hiring-rate data matters more than layoff counts.

Think in tasks, not job titles. Work is durable when it involves: physical presence and dexterity; accountability someone must legally hold (medicine, engineering sign-off, audits); high-stakes judgment under ambiguity; deep relationships and trust; or frontier expertise where training data is thin. Work is exposed when it's routine, digital, well-specified and reviewable — regardless of how skilled it once was. A useful test: if you could delegate a task to a very fast, well-read intern with no accountability, AI will eventually do it.

Probably, but the timing is the problem. Historically, technology transitions destroy specific jobs quickly and create new categories slowly — the gap is typically 5–10 years, and the people displaced are rarely the people hired into the new roles. We already see the new titles emerging (AI operations, agent orchestration, evals, AI governance). The honest answer: the aggregate may balance out; your individual outcome depends on whether you move toward the new work or wait for the old work to come back. It won't.

Yes — but 'learning to code' is no longer the product; it's the entry fee. AI writes code, which means the value moved to what surrounds the code: knowing what to build, decomposing problems, reviewing and debugging AI output, system design, security, and shipping complete products. People who learn to code with AI (understanding what they're doing, at higher speed) are gaining an advantage. People who learn to code by memorizing syntax and doing what AI does — slower — are preparing for 2019.

As a knowledge base, yes; as a job guarantee, no — that changed. New-grad hiring has roughly halved from 2019, and degree-holders now compete with bootcamp grads, self-taught AI-native builders, and each other. The degree still helps for visa pathways, big-company filters and fundamentals (which matter more, not less, when AI writes the easy code). But treat it as one input: employers now consistently weight demonstrated shipping ability — real projects, real users, real contributions — above credentials.

Meaningfully longer than 2021's instant market. Surveys through 2024–25 put typical searches at 3–6 months, senior and niche roles often longer, with new grads facing the longest odds. Plan finances for 6 months, treat searching as a pipeline problem (many parallel threads, not sequential applications), and expect the successful channel to be a referral or direct relationship rather than a portal application. If you're employed and worried: start the search before you need it — employed candidates still convert much better.

In order: (1) Get the paperwork right — severance terms, equity deadlines (often 90 days to exercise options), health coverage, and don't sign anything on day one. (2) File for unemployment immediately; it exists for exactly this. (3) Take one week to process — panic-applying produces bad applications. (4) Line up 3–5 references and get LinkedIn recommendations while memories are fresh. (5) Then run a structured search: define 2–3 target role types, activate your network before applying cold, and put a daily structure on it. Treat the layoff as a market event, not a personal verdict — 2022–25 laid off half a million skilled people; hiring managers know this.

Sometimes, and it costs little to try professionally. More negotiable than people assume: the departure date (extending tenure for vesting cliffs), equity treatment (accelerated vesting or extended exercise windows), healthcare continuation, outplacement support, and the reference/rehire designation. Cash amount is the hardest to move in a mass layoff (companies fear precedent), but individual circumstances — long tenure, recent high ratings, protected-class concerns — create leverage. If the numbers are large, an hour with an employment lawyer is usually worth it.

Tier 1 (strongest demand signal): building with AI — LLM integration, RAG, agents, evals; not just using ChatGPT. Tier 2 (durable premiums): security, cloud/platform engineering, data engineering. Tier 3 (multipliers on everything else): domain depth in a regulated industry, communication and stakeholder skills, and a public track record. The pattern across all tiers: skills that either build the AI wave, secure it, feed it data, or supply the judgment it lacks.

Climb the ladder: (1) Assistant level — drafting, summarizing, explaining unfamiliar code or concepts. (2) Workflow level — standing prompts and templates for recurring tasks; AI in your IDE, inbox and docs. (3) Automation level — scripts and agents that complete whole tasks (report generation, data cleaning, first-pass code review) with you as editor. (4) Product level — shipping AI features for customers. Each level up is rarer and more valuable. Document what you automate — 'I built an AI workflow that saved my team X hours/week' is a resume line that works in every function.

The data says experience is currently protective, not disqualifying — layoffs hit junior and volume roles hardest, and mid-career professionals bring exactly what AI lacks: judgment, domain context, and knowing how organizations actually work. The genuine risk for experienced professionals is different: identity attachment to a workflow that's ending. The adaptation is usually smaller than feared — your expertise stays the product; AI changes the delivery mechanism. A 45-year-old accountant who runs AI-assisted close processes is more valuable than either the old version or a junior with prompts.

The layoff tracker is hand-curated from primary sources: company announcements, SEC filings and credible reporting (Reuters, Bloomberg, CNBC, the Financial Times and peers). Each event links its source, and counts reflect what was reported at announcement time. The AI-attribution flag is applied conservatively — only when the company itself or credible reporting tied the cut to AI. Aggregate context draws on public research from Challenger Gray, LinkedIn's Economic Graph, and layoffs.fyi. We track announced events; actual final numbers sometimes differ.

It's a transparent, weighted scoring model — not a black box and not a prophecy. It combines: your role's task-level AI exposure (based on published task-automation research), current hiring demand for the role, industry stability, experience level, the share of your work that's routine, your AI fluency, and protective skills like domain depth, leadership and security. Every weight is disclosed on the methodology section of the calculator page. Treat the score as a structured way to think about your position — the recommendations matter more than the number.

Yes. The tracker, calculator, skills guides, predictions and all articles are free. The project exists because navigating this market shift is hard enough without paywalls between people and the information they need. If you find it useful, the best support is sharing it with someone whose career it might help.