{ "sym": "Sk", "updated": "2026-08-06", "verdict": "A public portfolio with receipts still wins \u2014 shipped product, live URL, paying customers, a repo history with your fingerprints on it \u2014 because it is the only route employers already know how to read and it costs nothing but the work. But 2026 broke its monopoly: 67% of HR leaders say AI-generated applications are slowing their hiring and 65% say AI-polished resumes have made skills verification harder (Robert Half, n=2,000+, released Mar 10, 2026), and Gartner expects one in four candidate profiles worldwide to be fake by 2028. So the portfolio now needs a second, adversarial layer. A live human technical assessment (Karat-class interview engineering) is the highest-verifiability proof anyone can buy, and it's free to you because the employer pays. Vendor AI certifications \u2014 AWS AI Practitioner $100, Google Generative AI Leader $99, NVIDIA NCA-GENL $125, Claude Certified from ~$99 \u2014 are the cheapest credible floor when you need a recognizable token fast, with the caveat that Microsoft just retired AI-102 and AI-900 outright on Jun 30, 2026. LinkedIn Verified Skills (launched Jan 28, 2026 with Descript, Lovable, Replit and Relay.app) is the interesting new shape: verification derived from actual tool usage rather than an exam \u2014 free, unfakeable in principle, and completely unproven in practice. Credly and Open Badges 3.0 are the rails underneath all of it, not a signal on their own. On the thesis \u2014 that verified human skill appreciates as AI-generated claims inflate \u2014 the evidence is real but asymmetric: the inflation is thoroughly documented, the premium is not.", "top5": [ { "rank": 1, "name": "Public portfolio + receipts", "maker": "you (GitHub, a live product, named customers)", "url": "https://github.com", "docs": "https://github.blog/news-insights/octoverse/octoverse-a-new-developer-joins-github-every-second-as-ai-leads-typescript-to-1/", "pricing": "Free \u00b7 GitHub Free gives unlimited public repos \u00b7 a domain ~$12/yr \u00b7 the real cost is shipping something", "best_for": "Anyone who has actually built and shipped \u2014 the highest-recognition, lowest-cost proof, and the only one that also functions as marketing.", "why": "It is the route every hiring manager already reads without a decoder ring, and the substrate is enormous: 180M+ developers on GitHub, 36M of them joining in the 12 months to Oct 2025, across 395M public repositories (+19% YoY). Crucially it carries information no exam does \u2014 commit cadence, how you handle a bug report, whether the thing has users. LinkedIn's own framing of its 2026 verification push was that employers \"want to know what you can actually do,\" which is a portfolio argument even when LinkedIn is selling something else.", "watch": "This is the entry AI degraded most. A convincing repo, README, landing page and demo video are now a weekend of agent time, and 65% of hiring managers say AI-enhanced applications have made skills verification harder (Robert Half, Mar 2026). Depth and provenance are the defense \u2014 long commit history, real users, public writing, and the ability to be interrogated live about a decision you made 14 months ago. A portfolio you cannot defend in conversation is now a liability, not an asset.", "evidence": [ { "stat": "180M+ developers on GitHub; ~36M joined in 12 months; 395M public repos, +19% YoY (Octoverse, published Oct 28, 2025, updated Feb 28, 2026)", "src": "https://github.blog/news-insights/octoverse/octoverse-a-new-developer-joins-github-every-second-as-ai-leads-typescript-to-1/" }, { "stat": "65% of hiring managers say AI-enhanced resumes complicate skills verification; 42% now spend more time reviewing applications (Robert Half, n=2,000+ US hiring managers, fielded Nov 2025, released Mar 10, 2026)", "src": "https://press.roberthalf.com/2026-03-10-Robert-Half-survey-67-of-HR-leaders-report-AI-generated-applications-are-slowing-hiring" }, { "stat": "Gartner: 1 in 4 candidate profiles worldwide could be fake by 2028; 6% of 3,000 surveyed candidates already admit to interview impersonation (Jul 31, 2025)", "src": "https://www.hrdive.com/news/fake-job-candidates-ai/757126/" } ], "tile_note": "receipts beat r\u00e9sum\u00e9s" }, { "rank": 2, "name": "Live human technical assessment", "maker": "Karat (+ Byteboard) \u00b7 CodeSignal \u00b7 interviewing.io", "url": "https://karat.com", "docs": "https://karat.com/detect-ai-use-technical-interviews/", "pricing": "Free to you \u2014 the employer pays. Employer side: CodeSignal Hire from $79/mo annual (60 credits) \u00b7 $479/mo Grow (420 credits) \u00b7 $20/credit overage; Karat is enterprise-quoted. interviewing.io is free for candidates; Karat's Brilliant Black Minds gives free mock interviews.", "best_for": "Converting a claim into an observation \u2014 a trained interviewer watching you reason in real time, which is currently the only proof AI cannot sit in for.", "why": "It is the highest-verifiability route that exists, and 2026 made it structurally more valuable rather than less: 41% of enterprises surveyed say they have already hired and onboarded a fraudulent candidate (GetReal, n=668 security and risk leaders, Sep 2025), and 38% of employers are responding by adding interview rounds (Robert Half, Mar 2026). Karat has run 600,000+ technical interviews and serves 7 of the 25 largest public companies; its own guidance is blunt that live human interviews remain the strongest safeguard, because the tells are behavioral \u2014 instant optimized solutions with no iteration, explanations that don't match the code.", "watch": "You cannot buy it for yourself \u2014 it is employer-procured, so as a founder you're on the wrong side of the transaction unless you use the free channels. It's also the most expensive route in wall-clock time and the most stressful, and the anti-cheat arms race is live: an entire cottage industry of overlay tools now markets itself specifically against Karat, CodeSignal and Triplebyte assessments. Results are rarely portable across employers.", "evidence": [ { "stat": "600,000+ technical interviews conducted; clients include 7 of the 25 largest public companies (karat.com, Aug 2026)", "src": "https://karat.com/" }, { "stat": "41% of enterprises report having hired and onboarded a fraudulent candidate; 88% encounter deepfake/impersonation attacks at least occasionally (GetReal Security, n=668, Sep 2025)", "src": "https://www.getrealsecurity.com/resources/new-getreal-security-research-41-of-enterprises-surveyed-report-having-hired-and-onboarded-fraudulent-candidates" }, { "stat": "62% of candidates prefer in-person interviews; only 26% trust AI to evaluate them fairly (Gartner, n=3,000 candidates, Jul 31, 2025)", "src": "https://www.gartner.com/en/newsroom/press-releases/2025-07-31-gartner-survey-shows-just-26-percent-of-job-applicants-trust-ai-will-fairly-evaluate-them" } ], "tile_note": "a human watches you work" }, { "rank": 3, "name": "Vendor AI certifications", "maker": "AWS \u00b7 Google Cloud \u00b7 Microsoft \u00b7 NVIDIA \u00b7 Anthropic \u00b7 Databricks", "url": "https://aws.amazon.com/certification/certified-ai-practitioner/", "docs": "https://cloud.google.com/learn/certification/generative-ai-leader", "pricing": "Per exam: AWS AI Practitioner (AIF-C01) $100 \u00b7 AWS ML Engineer Associate (MLA-C01) $150 \u00b7 Google Generative AI Leader $99 \u00b7 Google Professional ML Engineer $200 \u00b7 NVIDIA NCA-GENL $125 \u00b7 Claude Certified Associate ~$99, Developer/Architect ~$125, Architect Professional ~$175 (Anthropic prices are aggregator-reported \u2014 see notes). Prep courses are free on Anthropic Academy, Google Cloud Skills Boost and Microsoft Learn.", "best_for": "Buying a recognizable, proctored, third-party-verified token in a weekend for around $100 \u2014 especially when a procurement form, partner tier or enterprise buyer needs a box ticked.", "why": "This is the cheapest credible floor in the category and the only one with genuine identity verification baked in: all of these are Pearson VUE proctored, ID-checked, and badge-issued through Credly. The AI-native end has real traction \u2014 Anthropic reports 36,000+ certified consultants across 1,300+ organizations as of Jul 2026 and 400,000+ people through Partner Academy training, and it added three new role-based certifications on Jul 23, 2026. Cost-to-signal is unmatched: $99\u2013$200 and 90\u2013120 minutes.", "watch": "Shallow and churn-prone. Google's Generative AI Leader explicitly targets people \"with or without hands-on technical experience\" \u2014 it certifies vocabulary, not capability. Microsoft retired both AI-102 and AI-900 on Jun 30, 2026, replacing them with AI-103 and AI-901, so a 2025 Azure AI badge is now a dead credential. Validity is short: NVIDIA's is 2 years, AWS and Google 3. And Anthropic's proctored exams are gated behind a Claude Partner Network company email \u2014 independent founders and consultants are locked out regardless of skill.", "evidence": [ { "stat": "AWS Certified AI Practitioner: $100, 65 questions, 90 minutes, valid 3 years, Pearson VUE proctored (aws.amazon.com, Aug 2026)", "src": "https://aws.amazon.com/certification/certified-ai-practitioner/" }, { "stat": "36,000+ consultants Claude-certified across 1,300+ organizations; 400,000+ trained via Anthropic Partner Academy; three new certifications announced Jul 23, 2026", "src": "https://claude.com/blog/four-role-based-claude-certifications" }, { "stat": "Microsoft retired AI-102 (Azure AI Engineer) and AI-900 (Azure AI Fundamentals) on Jun 30, 2026; successors AI-103 and AI-901 entered beta Apr 21, 2026", "src": "https://learn.microsoft.com/en-us/credentials/certifications/azure-ai-engineer/" } ], "tile_note": "cheap, proctored, recognized" }, { "rank": 4, "name": "LinkedIn Verified Skills", "maker": "LinkedIn (Microsoft), with tool partners", "url": "https://www.linkedin.com", "docs": "https://techcrunch.com/2026/01/28/linkedin-will-let-you-show-off-your-vibe-coding-chops-with-a-certificate/", "pricing": "Free \u2014 but you need a paid seat on a partner tool to generate usage (Replit, Lovable, Descript, Relay.app; GitHub, Zapier and Gamma announced as coming)", "best_for": "AI-tool operators who want proof attached to the profile employers actually look at, generated by work you were doing anyway.", "why": "Announced Jan 28, 2026, this is the first mainstream credential derived from telemetry rather than a test: partner tools assess \"real usage patterns, product outcomes, or demonstrated proficiency within the tool,\" and the credential updates automatically as you improve. It sits on the distribution nobody else has \u2014 the profile that recruiters open first \u2014 and it is structurally harder to fake than an exam, because faking it requires actually using the product. LinkedIn is pitching it against the degree directly: employers \"want to know what you can actually do.\"", "watch": "Almost entirely unproven. LinkedIn has published no earner counts, no employer-recognition data, and no detail on what qualifies as proficient \u2014 and it has form here: its previous multiple-choice Skill Assessments were quietly retired. The signal is also vendor-shaped, not skill-shaped (\"proficient in Lovable\" is a tool fact, not an engineering fact), and the partner list is a short list of AI startups whose own longevity is uncertain. Treat it as a free garnish on a portfolio, not a substitute for one.", "evidence": [ { "stat": "Launched Jan 28, 2026 with Descript, Lovable, Replit and Relay.app; GitHub, Zapier and Gamma announced as future partners", "src": "https://techcrunch.com/2026/01/28/linkedin-will-let-you-show-off-your-vibe-coding-chops-with-a-certificate/" }, { "stat": "Verification is derived from \"real usage patterns, product outcomes, or demonstrated proficiency within the tool, not assumptions or tests\"; credentials auto-update (Fortune, Jan 28, 2026)", "src": "https://www.fortune.com/2026/01/28/linkedin-ai-skills-verification-profile-skills-mismatch" }, { "stat": "AI-related keywords appeared in 4.2% of US job postings by end-2025 (Indeed Hiring Lab, cited in TechCrunch, Jan 2026)", "src": "https://techcrunch.com/2026/01/28/linkedin-will-let-you-show-off-your-vibe-coding-chops-with-a-certificate/" } ], "tile_note": "verified by usage" }, { "rank": 5, "name": "Portable credential rails", "maker": "Credly by Pearson \u00b7 1EdTech Open Badges 3.0 \u00b7 Accredible \u00b7 skill.re / Certificate.School", "url": "https://info.credly.com/", "docs": "https://www.1edtech.org/standards/open-badges", "pricing": "Free to the earner. Issuer side: Accredible from $45/mo (50 recipients, unlimited credentials each) \u00b7 Credly enterprise-quoted \u00b7 skill.re free to browse and join, 195 of its 255 courses end in a serial-numbered verifiable certificate", "best_for": "Making a credential you already earned machine-verifiable, portable and revocable \u2014 and, if you run a program, issuing credentials that survive contact with a skeptical employer.", "why": "This is the plumbing that turns a PDF into something checkable: Credly has issued 123M+ credentials from 3,700+ issuers across roughly 200,000 distinct credentials, and Open Badges 3.0 (released May 29, 2024) makes badges W3C Verifiable Credentials 2.0-compatible \u2014 cryptographically signed, wallet-portable, and carrying the assessment criteria and evidence inside the badge itself. Every vendor cert above already lands here, so you get the rails whether you chose them or not. The skill.re-class layer \u2014 cohort courses ending in a serialized, publicly verifiable certificate \u2014 is the cheap on-ramp for skills nobody sells a proctored exam for.", "watch": "The rails are not the signal. A verifiable badge proves the issuer issued it, not that the issuer's bar was high \u2014 and course-completion certificates from small platforms carry close to zero employer recognition. Open Badges 3.0 adoption is also slower than the spec suggests: 2.0 remains the most widely used version in 2026, so the wallet/decentralized-identity story is still mostly forward-looking. Credly publishes no public pricing.", "evidence": [ { "stat": "Credly by Pearson: 123M+ credentials managed, 3,700+ issuers, ~200,000 credentials in network, 650,000+ shared monthly (info.credly.com, Aug 2026) \u2014 up from 100M credentials and 48M earners at Jan 7, 2025", "src": "https://info.credly.com/" }, { "stat": "Open Badges 3.0 released May 29, 2024; badges are W3C Verifiable Credentials, carrying assessment criteria, evidence and skills-framework alignment", "src": "https://www.1edtech.org/1edtech-article/new-open-badges-30-standard-provides-enhanced-security-and-mobility/411060" }, { "stat": "Open Badges 2.0 \"remains the most widely used and accepted standard\" in 2026 \u2014 3.0 uptake still limited (VirtualBadge status review, 2026)", "src": "https://www.virtualbadge.io/blog-articles/open-badges-3-0-what-is-the-status-in-2026" } ], "tile_note": "the rails, not the signal" } ], "matrix": { "cols": [ "Cost to you", "Time to earn", "Employer recognition", "Verifiability", "Durability", "AI-fakeable?", "Who pays" ], "rows": [ [ "Public portfolio + receipts", "Free (~$12/yr domain)", "Months\u2013years", "Highest", "Medium \u2014 inspectable, not attested", "Compounds forever", "Surface yes, depth no", "You (in time)" ], [ "Live human interview", "Free to candidate", "1\u20134 hours", "High (hiring only)", "Highest", "Per-employer, rarely portable", "No", "Employer" ], [ "Vendor AI cert (AWS/Google/NVIDIA)", "$99\u2013$200/exam", "1\u20133 weekends", "Medium\u2013high", "High \u2014 proctored, ID-checked", "2\u20133 yrs, retirement risk", "No (proctored)", "You" ], [ "Claude Certified (Anthropic)", "~$99\u2013$175/exam", "1\u20132 weekends", "Medium, rising fast", "High \u2014 Pearson VUE + Credly", "Model-generation short", "No (proctored)", "You / your partner org" ], [ "LinkedIn Verified Skills", "Free (needs partner tool)", "Continuous", "Unproven", "Medium \u2014 vendor telemetry", "Auto-updating, platform risk", "Hard \u2014 requires real usage", "Tool vendor" ], [ "Async coding assessment", "Free to candidate", "60\u2013120 min", "Medium", "Medium \u2014 active arms race", "Single-process", "Yes, actively", "Employer" ], [ "Course certificate (skill.re-class)", "Free\u2013$49/mo", "Hours\u2013weeks", "Low", "Medium \u2014 serial + verify page", "Indefinite, low weight", "N/A (attendance, not skill)", "You" ], [ "Credly badge / Open Badges 3.0", "Free to earner", "Instant on issue", "Inherits the issuer's", "High \u2014 cryptographic", "Tracks the credential", "No (issuer-signed)", "Issuer" ] ] }, "rules": [ { "if": "You have shipped something with real users", "then": "Lead with the receipts and stop buying credentials. Nothing in this table outranks a live product with named customers \u2014 but make the provenance legible (commit history, public writing, changelog), because 65% of hiring managers now say AI-polished applications made verification harder." }, { "if": "You have the skill but no shipped artifact yet, and you need a signal this month", "then": "One proctored vendor exam \u2014 AWS AI Practitioner ($100) or Google Generative AI Leader ($99) \u2014 is the fastest $100 of credibility available. Budget one weekend, and check the retirement calendar first: Microsoft killed AI-102 and AI-900 outright on Jun 30, 2026." }, { "if": "You sell AI implementation services to enterprises", "then": "Claude Certified via the Anthropic Partner Network \u2014 36,000+ certified across 1,300+ orgs as of Jul 2026 makes it the emerging table stakes in that buying conversation. You will need a partner-org company email; a personal address won't get you into the exam." }, { "if": "You live in AI tools all day and want free, always-on proof", "then": "Turn on LinkedIn Verified Skills through Replit, Lovable, Descript or Relay.app. It costs nothing and rides the profile recruiters already open \u2014 but treat it as garnish, not as your primary evidence, until LinkedIn publishes recognition data." }, { "if": "You're the one issuing credentials \u2014 a course, a bootcamp, an internal academy", "then": "Issue on Open Badges 3.0 rails (Accredible from $45/mo, or Credly if you're already in the Pearson ecosystem) so the badge is cryptographically checkable. Assume most employers still read Open Badges 2.0 in 2026 and design for that." } ], "field": [ { "name": "GitHub", "maker": "Microsoft", "note": "The default portfolio substrate \u2014 180M+ developers, 395M public repos (Octoverse, Oct 2025); announced as a future LinkedIn Verified Skills partner", "url": "https://github.com", "oss": false, "entry": "free", "status": "active" }, { "name": "Karat", "maker": "Karat, Inc.", "note": "Interview-engineering network; 600,000+ technical interviews run, 7 of the 25 largest public companies as clients; added an AI Readiness assessment line in 2026", "url": "https://karat.com", "oss": false, "entry": "employer-quoted", "status": "active" }, { "name": "Byteboard", "maker": "Karat (ex-Google Area 120)", "note": "Project-based, real-world coding interview; acquired by Karat and now folded into its platform", "url": "https://www.byteboard.dev", "oss": false, "entry": "employer-quoted", "status": "acquired" }, { "name": "Triplebyte", "maker": "Karat", "note": "The original take-a-test-once, skip-to-onsite network; acquired by Karat in 2022 and largely absorbed \u2014 the brand now survives mainly as 'Karat NextGen' assessments referenced by third parties", "url": "https://triplebyte.com", "oss": false, "entry": "free to candidates", "status": "acquired" }, { "name": "CodeSignal", "maker": "CodeSignal", "note": "Assessments + interviews + a learning arm; Certified Assessments are IO-psychologist-validated. Hire from $79/mo annual; Learn free with Cosmo+ at $24.99/mo", "url": "https://codesignal.com", "oss": false, "entry": "$79/mo (employer) \u00b7 free (learner)", "status": "active" }, { "name": "HackerRank", "maker": "HackerRank", "note": "Largest developer-assessment install base; publishes the annual Developer Skills Report \u2014 the 2026 edition was not resolvable at its expected URL when checked", "url": "https://www.hackerrank.com", "oss": false, "entry": "employer-quoted", "status": "active" }, { "name": "TestGorilla", "maker": "TestGorilla", "note": "350+ role, cognitive and personality tests; free tier with 10 monthly credits, Core $142/mo \u2014 the accessible option for small teams verifying non-engineering skill", "url": "https://www.testgorilla.com", "oss": false, "entry": "free \u00b7 $142/mo", "status": "active" }, { "name": "Codility", "maker": "Codility", "note": "Long-standing async coding assessment with plagiarism/AI-similarity detection; enterprise-quoted", "url": "https://www.codility.com", "oss": false, "entry": "employer-quoted", "status": "active" }, { "name": "Vervoe", "maker": "Vervoe", "note": "AI-graded job simulations across non-technical roles \u2014 the work-sample approach outside engineering", "url": "https://vervoe.com", "oss": false, "entry": "employer-quoted", "status": "active" }, { "name": "interviewing.io", "maker": "interviewing.io", "note": "Anonymous mock interviews with FAANG engineers, free to candidates, with a fast-track to real technical screens \u2014 the one live-assessment route you can self-serve", "url": "https://interviewing.io", "oss": false, "entry": "free", "status": "active" }, { "name": "Credly by Pearson", "maker": "Pearson", "note": "The dominant badge registry \u2014 123M+ credentials, 3,700+ issuers; where AWS, Microsoft, NVIDIA and Anthropic badges land. No public pricing", "url": "https://info.credly.com", "oss": false, "entry": "enterprise-quoted", "status": "active" }, { "name": "Accredible", "maker": "Accredible", "note": "Credly's main challenger; priced per recipient not per credential \u2014 Launch $45/mo for 50 recipients with unlimited credentials each", "url": "https://www.accredible.com", "oss": false, "entry": "$45/mo", "status": "active" }, { "name": "Open Badges 3.0", "maker": "1EdTech", "note": "The standard, released May 29, 2024 \u2014 badges as W3C Verifiable Credentials 2.0, carrying criteria and evidence. Adoption real but slow; 2.0 still dominant in 2026", "url": "https://www.1edtech.org/standards/open-badges", "oss": true, "entry": "free (spec)", "status": "active" }, { "name": "W3C Verifiable Credentials 2.0", "maker": "W3C", "note": "The cryptographic substrate Open Badges 3.0 sits on; the reason a 2026 badge can be checked without phoning the issuer", "url": "https://www.w3.org/TR/vc-data-model-2.0/", "oss": true, "entry": "free (spec)", "status": "active" }, { "name": "BCdiploma", "maker": "BCdiploma", "note": "Blockchain-anchored credential issuer; earned 1EdTech Open Badges 3.0 certification \u2014 a rare confirmed 3.0 implementation", "url": "https://www.bcdiploma.com", "oss": false, "entry": "vendor-quoted", "status": "active" }, { "name": "Certificate.School", "maker": "Certificate.School", "note": "Issuance/verification layer behind skill.re \u2014 serial + QR per certificate, public no-login verify portal, REST API, revocation. Operator not disclosed on site", "url": "https://certificate.school", "oss": false, "entry": "free tier", "status": "active" }, { "name": "skill.re", "maker": "skill.re", "note": "The canon pick: cohort-based courses where 195 of 255 end in a serialized verifiable credential graded Aware\u2192Visionary. Free to browse and join; small, young, no published employer recognition", "url": "https://skill.re", "oss": false, "entry": "free to join", "status": "active" }, { "name": "LinkedIn Verified Skills", "maker": "LinkedIn (Microsoft)", "note": "Usage-derived AI-skill verification launched Jan 28, 2026 with Descript, Lovable, Replit, Relay.app; GitHub, Zapier, Gamma announced. No earner counts published", "url": "https://www.linkedin.com", "oss": false, "entry": "free", "status": "active" }, { "name": "LinkedIn Skill Assessments", "maker": "LinkedIn (Microsoft)", "note": "The old multiple-choice badge test \u2014 retired; LinkedIn's help centre now carries a 'no longer available' notice. The precedent worth remembering before over-investing in its replacement", "url": "https://www.linkedin.com/help/linkedin/answer/a1690529", "oss": false, "entry": "\u2014", "status": "dead" }, { "name": "AWS Certification", "maker": "AWS", "note": "AI Practitioner (AIF-C01) $100, 90 min, 3-yr validity; ML Engineer Associate (MLA-C01) $150, 130 min \u2014 passing MLA auto-recertifies AIF", "url": "https://aws.amazon.com/certification/", "oss": false, "entry": "$100\u2013$150/exam", "status": "active" }, { "name": "Google Cloud Certification", "maker": "Google Cloud", "note": "Generative AI Leader $99 (90 min, no prerequisites, explicitly non-technical); Professional ML Engineer $200 (2 hrs), recently rewritten from Vertex AI to Gemini Enterprise Agent Platform", "url": "https://cloud.google.com/learn/certification", "oss": false, "entry": "$99\u2013$200/exam", "status": "active" }, { "name": "Microsoft Credentials", "maker": "Microsoft", "note": "In active churn: AI-102 and AI-900 retired Jun 30, 2026; successors AI-103 (Azure AI App and Agent Developer) and AI-901 entered beta Apr 21, 2026, plus AB-620 AI Agent Builder Associate", "url": "https://learn.microsoft.com/en-us/credentials/", "oss": false, "entry": "~$99\u2013$165/exam", "status": "renamed" }, { "name": "NVIDIA Certification", "maker": "NVIDIA", "note": "NCA Generative AI LLMs (NCA-GENL) $125, 1 hour, online-proctored, 50\u201360 questions \u2014 but only a 2-year validity, the shortest of the majors", "url": "https://www.nvidia.com/en-us/learn/certification/", "oss": false, "entry": "$125/exam", "status": "active" }, { "name": "Claude Certifications", "maker": "Anthropic", "note": "Four role-based certs (Associate/Developer/Architect Foundations, Architect Professional), Pearson VUE proctored, Credly-badged; 36,000+ certified across 1,300+ orgs (Jul 2026). Partner Network company email required", "url": "https://claude.com/blog/four-role-based-claude-certifications", "oss": false, "entry": "~$99\u2013$175/exam", "status": "active" }, { "name": "Databricks Certification", "maker": "Databricks", "note": "Generative AI Engineer Associate plus a new Certified Context Engineering Associate \u2014 the first mainstream cert aimed at agent-context design. Exam pricing not published on the hub page", "url": "https://www.databricks.com/learn/certification", "oss": false, "entry": "vendor-quoted", "status": "active" }, { "name": "Google AI Essentials / Career Certificates", "maker": "Google (via Coursera)", "note": "Under 5 hours, $49/mo Coursera subscription after a 7-day trial; broad name recognition, shallow proof \u2014 attendance, not assessment", "url": "https://grow.google/certificates/ai-essentials/", "oss": false, "entry": "$49/mo", "status": "active" }, { "name": "Kaggle", "maker": "Google", "note": "Competition tiers (Novice\u2192Grandmaster) are the most adversarially-earned skill signal that exists in ML \u2014 public leaderboards, no self-report. Free; narrow to data/ML", "url": "https://www.kaggle.com", "oss": false, "entry": "free", "status": "active" }, { "name": "Polywork", "maker": "Polywork", "note": "Timeline-style 'what I actually did' profile \u2014 the right idea, but activity on the site is sparse and no 2026 product news surfaced", "url": "https://www.polywork.com", "oss": false, "entry": "free", "status": "fading" }, { "name": "GetReal Security", "maker": "GetReal Security", "note": "Deepfake-resistant identity verification for hiring \u2014 the adjacent layer that proves you are you before anyone assesses what you can do", "url": "https://www.getrealsecurity.com", "oss": false, "entry": "enterprise-quoted", "status": "active" } ], "signals": [ { "fact": "67% of HR leaders say AI-generated applications are slowing hiring, 20% by more than two weeks; 65% of hiring managers say AI-enhanced resumes complicate skills verification; 84% report heavier workloads (Robert Half, n=2,000+ US hiring managers, fielded Nov 2025, released Mar 10, 2026)", "src": "https://press.roberthalf.com/2026-03-10-Robert-Half-survey-67-of-HR-leaders-report-AI-generated-applications-are-slowing-hiring" }, { "fact": "Gartner predicts 1 in 4 candidate profiles worldwide will be fake by 2028; 6% of 3,000 surveyed candidates already admit to interview impersonation; only 26% of applicants trust AI to evaluate them fairly (Jul 31, 2025)", "src": "https://www.hrdive.com/news/fake-job-candidates-ai/757126/" }, { "fact": "41% of enterprises surveyed have already hired and onboarded a fraudulent candidate; 88% encounter deepfake or impersonation attacks at least occasionally, 45% frequently (GetReal Security, n=668 IT/security/risk leaders at 1,000+-employee firms, Sep 2025)", "src": "https://www.getrealsecurity.com/resources/new-getreal-security-research-41-of-enterprises-surveyed-report-having-hired-and-onboarded-fraudulent-candidates" }, { "fact": "Credential volume is compounding: Credly went from 100M credentials and 48M earners across 4,000+ issuers (Jan 7, 2025) to 123M+ credentials and ~200,000 distinct credentials from 3,700+ issuers (Aug 2026) \u2014 more credentials, slightly fewer issuers", "src": "https://www.prnewswire.com/news-releases/with-100-million-digital-credentials-issued-through-credly-pearson-fosters-a-future-proof-workforce-for-enterprises-in-the-ai-era-and-beyond-302343746.html" }, { "fact": "Skills for the most AI-exposed jobs are changing more than twice as fast as for the least exposed \u2014 a 75% wider gap than the prior year; 'professionalised' roles are growing twice as fast as 'democratised' ones with 42% higher wage growth since 2021 (PwC AI Jobs Barometer, Jun 15, 2026)", "src": "https://www.pwc.com/gx/en/issues/artificial-intelligence/ai-jobs-barometer.html" }, { "fact": "Counter-evidence: employers' announced shift to skills-based hiring produced almost no change in who actually gets hired \u2014 degree-requirement removals translated into a negligible share of real hires (Burning Glass Institute + Harvard Business School Project on Managing the Future of Work, Feb 2024)", "src": "https://www.burningglassinstitute.org/research/skills-based-hiring-2024" } ], "notes": "Ranking charter: verifiability against an adversary first, employer recognition second, cost last \u2014 which is why a free portfolio outranks every paid credential, and why a live human interview outranks a $100 exam. No affiliate consideration; picks are editorial.\n\nOn the thesis \u2014 'verified human skill appreciates as AI-generated claims inflate' \u2014 the evidence is genuinely asymmetric and we're saying so. FOR: the inflation is thoroughly documented and dated (Robert Half Mar 2026, Gartner Jul 2025, GetReal Sep 2025), employers are visibly spending more per candidate (42% more review time, 38% more interview rounds), and PwC's Jun 2026 barometer shows AI-exposed roles bifurcating toward professionalisation with materially higher wage growth. AGAINST: nobody has published a clean measurement of a premium attached to a *verified* credential as distinct from the underlying skill; the Burning Glass/HBS finding that skills-based-hiring announcements barely moved actual hiring is the strongest counter, and it predates the AI-slop wave; and Credly's own volume growth (100M\u2192123M credentials in ~19 months) means credential supply is inflating too, which cuts against scarcity. Net read: the *demand* for verification is rising fast and is well-evidenced; the *price* of a verified credential is not yet demonstrable. Anyone claiming a specific % salary premium for AI certifications in 2026 is quoting SEO content, not research \u2014 we found no primary source and cite none.\n\nConflicts and weak sources, stated plainly. (1) Anthropic exam prices ($99 Associate / $125 Developer / $125 Architect Foundations / $175 Architect Professional) come from a third-party guide, not Anthropic \u2014 Anthropic's own posts say only that 'getting started is free' with discounted first attempts for tiered partners. Treat the numbers as indicative. (2) Anthropic's certified-consultant count moved from '10,000+ since March' (partner-hub post) to '36,000+ across 1,300+ organizations' (Jul 2026 certifications post); we cite the later, larger, more specific figure. (3) Microsoft's AI-102/AI-900 retirement dates (Jun 30, 2026) and successor codes (AI-103, AI-901, AB-620) come from a specialist tracker; the retirement itself is confirmed first-party by the retired-credential banner on learn.microsoft.com, and Microsoft's own AI-103 certification page did not resolve when checked. (4) Karat's 600,000+ interview figure is self-reported on its own site; its guidance page separately says 'more than half a million', so the number is approximate. (5) Open Badges 3.0 adoption: 1EdTech's own materials are promotional and give no adoption count; the claim that 2.0 still dominates in 2026 is single-sourced to a vendor blog and should be re-checked. (6) Triplebyte's exact current operating state under Karat is unclear \u2014 its site still serves the old candidate flow, Karat's own site never mentions it, and third-party tooling refers to 'Triplebyte (Karat NextGen)'. We flag it 'acquired' rather than 'dead'.\n\nScope boundaries. This element covers how *your* skill gets proven. Buying assessment software to screen *other people* is a hiring-stack question \u2192 Hr. Communities where reputation accrues informally \u2192 Co. Advisors and mentors who vouch for you \u2192 Mn. The AI coding tools whose usage LinkedIn now verifies \u2192 Ca and Vb. Identity verification (are you a real human) is a prerequisite here but belongs to the trust stack, not this element; GetReal is listed only as the adjacent layer.\n\nOne structural note founders should read twice: three of the five routes here are employer-funded or employer-gated (live interviews, async assessments, and the partner-gated Anthropic exams). If you are a solo founder rather than a job seeker, your practically available set is portfolio + receipts, a ~$100 vendor exam, LinkedIn Verified Skills, and course credentials \u2014 which is exactly why the portfolio ranks first.", "sources": [ "https://press.roberthalf.com/2026-03-10-Robert-Half-survey-67-of-HR-leaders-report-AI-generated-applications-are-slowing-hiring", "https://www.hrdive.com/news/fake-job-candidates-ai/757126/", "https://www.gartner.com/en/newsroom/press-releases/2025-07-31-gartner-survey-shows-just-26-percent-of-job-applicants-trust-ai-will-fairly-evaluate-them", "https://www.getrealsecurity.com/resources/new-getreal-security-research-41-of-enterprises-surveyed-report-having-hired-and-onboarded-fraudulent-candidates", "https://github.blog/news-insights/octoverse/octoverse-a-new-developer-joins-github-every-second-as-ai-leads-typescript-to-1/", "https://karat.com/", "https://karat.com/detect-ai-use-technical-interviews/", "https://codesignal.com/pricing/", "https://www.testgorilla.com/pricing/", "https://aws.amazon.com/certification/certified-ai-practitioner/", "https://aws.amazon.com/certification/certified-machine-learning-engineer-associate/", "https://cloud.google.com/learn/certification/generative-ai-leader", "https://cloud.google.com/learn/certification/machine-learning-engineer", "https://learn.microsoft.com/en-us/credentials/certifications/azure-ai-engineer/", "https://vladtalkstech.com/microsoft-learning-and-credential-news/microsoft-certification-retirements-2026/", "https://www.nvidia.com/en-us/learn/certification/generative-ai-llm-associate/", "https://claude.com/blog/four-role-based-claude-certifications", "https://www.anthropic.com/news/services-track-partner-hub", "https://www.databricks.com/learn/certification", "https://techcrunch.com/2026/01/28/linkedin-will-let-you-show-off-your-vibe-coding-chops-with-a-certificate/", "https://www.fortune.com/2026/01/28/linkedin-ai-skills-verification-profile-skills-mismatch", "https://info.credly.com/", "https://www.prnewswire.com/news-releases/with-100-million-digital-credentials-issued-through-credly-pearson-fosters-a-future-proof-workforce-for-enterprises-in-the-ai-era-and-beyond-302343746.html", "https://www.accredible.com/pricing", "https://www.1edtech.org/standards/open-badges", "https://www.1edtech.org/1edtech-article/new-open-badges-30-standard-provides-enhanced-security-and-mobility/411060", "https://www.virtualbadge.io/blog-articles/open-badges-3-0-what-is-the-status-in-2026", "https://skill.re/", "https://certificate.school/", "https://www.byteboard.dev/", "https://www.pwc.com/gx/en/issues/artificial-intelligence/ai-jobs-barometer.html", "https://www.burningglassinstitute.org/research/skills-based-hiring-2024", "https://grow.google/certificates/ai-essentials/", "https://interviewing.io/" ], "element": { "number": 55, "name": "Verified Skill", "group": "Human Elements", "essential": false, "edition": "v2026.Q3", "revision": "r7", "license": "CC BY 4.0 \u2014 cite elems.ai", "url": "https://elems.ai/e/sk.html" } }