Recommended Free Tools
Technology layoffs and unfilled specialist roles are not mutually exclusive. In a March 26, 2024 interview with Computerworld, Scott Rogers, then Udemy’s senior vice president of Instructor & Content Strategy, argued that employers were becoming more selective about current, demonstrable skills as generative AI changed job responsibilities and shortened the useful life of some technical skills. That interview describes 2023 and earlier conditions—not the labor market in August 2026—but its practical lesson remains relevant: learning a fashionable tool is not the same as becoming employable.
What explains layoffs alongside “skills shortages”?
Rogers’ explanation was a mismatch between the people available and the capabilities employers needed. A company might cut staff in one product group while recruiting data, cloud, cybersecurity or AI specialists elsewhere. A laid-off worker’s previous title may not map cleanly to a new job description, even when both roles are called “software” or “IT.”
Several forces can widen that gap:
- Requirements change faster than many workers can retrain.
- Companies slow general hiring while selectively recruiting for immediately productive specialists.
- Longer interview processes, location or work-authorization limits, compensation differences and automated résumé screening can keep qualified people out of consideration.
- An open vacancy is not evidence that every displaced worker is qualified, geographically suited or competitive for it.
The interview presents rapidly changing skills as one explanation, not proof that skills depreciation caused most layoffs. Udemy’s platform data also cannot establish the total economy-wide demand for any skill.
What Udemy’s GenAI numbers actually measure
Rogers told Computerworld that the first ChatGPT course appeared on Udemy 11 days after ChatGPT launched. During 2023, he said, more than 3.2 million learners enrolled across roughly 1,700 generative-AI courses and consumed more than 79 million minutes of GenAI instruction. ChatGPT, Midjourney and prompt engineering were among the popular subjects. Udemy’s 2024 trends reporting identified ChatGPT as its most-consumed global skill.
#1 Best Overall
Those are useful signals of interest among Udemy users, not employment statistics. “Learners,” “enrollments,” minutes watched and course popularity are different measures; none proves completion, competence, job placement or employer-validated proficiency. The related Udemy report drew on nearly 15,000 business customers and compared core business-learning activity from July 1, 2022, to July 1, 2023, according to the company’s announcement (methodology and report announcement).
Udemy’s later reports show continued growth in AI learning, but they should not be retroactively attributed to the 2024 interview. Its 2026 report, based on activity from July 1, 2024, through June 30, 2025, reported 11 million GenAI enrollments to date (Udemy’s 2026 report). Udemy also reported 859% year-over-year growth in GenAI-for-productivity learning and a 3,949% increase in LangChain consumption in its 2025 report (2025 report). These remain company-reported learning-activity measures.
GenAI is moving from specialist teams into ordinary jobs
Rogers identified data scientists, machine-learning engineers, AI researchers and product managers as early areas of investment. He also described a broader shift toward using GenAI in software engineering, data science, sales, marketing, finance and human resources—for example, financial modeling and analysis or drafting job descriptions.
Udemy launched its GenAI Skills Pack on March 13, 2024, after assessing more than 1,700 GenAI courses and organizing material into technical and business role pathways (Skills Pack announcement). The underlying idea is more durable than any particular product: employers increasingly want people who can apply AI inside a function, not merely explain what a language model is.
Is prompt engineering enough?
No. Prompting can be useful, but it is one component of a broader capability. A credible AI-ready profile combines:
- AI literacy: knowing model limitations, uncertainty and appropriate use.
- Tool fluency: using relevant workplace systems, APIs or copilots.
- Data competence: preparing, interpreting and protecting data.
- Domain knowledge: understanding the business or technical problem.
- Evaluation: checking accuracy, bias, security, privacy and reliability.
- Workflow design: deciding what to automate and where human review is required.
- Communication and judgment: explaining decisions and handling ambiguity.
- Proof: showing a working result rather than listing a course.
A portfolio project might document an AI-assisted reporting workflow, its test cases, error rates, privacy controls and human sign-off. That is stronger evidence than a prompt-engineering badge with no applied output.
Do practical skills matter more than a computer-science degree?
Rogers said companies were placing more emphasis on practical skills, experience and certifications, with greater flexibility around formal degree requirements. He also said Udemy had seen 10 million IT-certification enrollments during the prior 12 months. Those statements describe Udemy’s observations, not a universal hiring rule.
Degrees still matter differently by employer, occupation, seniority, geography and regulation. A certification is a signal, not a guarantee of competence, and a course completion certificate is generally weaker than a recognized professional certification, relevant work history or a substantial project. “Skills-based hiring” only improves access when employers genuinely assess skills instead of replacing degree filters with opaque credential filters.
A job-first learning plan for someone who has been laid off
- Choose two or three target titles. Do not begin with a course catalog.
- Collect 20 to 30 current postings in the geography where you can work.
- Group recurring requirements into core technical skills, AI-related tools, cloud or platforms, security and compliance, and communication or leadership.
- Find the smallest gap blocking candidacy. A narrow, repeated requirement is usually more valuable than a broad survey course.
- Build one work-like artifact. Examples include a GitHub application, cloud deployment, data-analysis case study, security assessment or documented business-process redesign.
- Add a certification only when postings repeatedly request it or it materially improves screening odds. Budget for exam fees, cloud usage, lab subscriptions and renewals.
- Rewrite your résumé around outcomes. State the tool, scale, business result and your role rather than listing hours of study.
- Use referrals and direct evidence alongside coursework. A course alone does not restore employability or guarantee an interview.
How to judge an online course before spending money
| Check | Why it matters |
|---|---|
| Match to target postings | Prevents spending on skills employers in your market do not request. |
| Update date and tool versions | AI interfaces, models and APIs can become obsolete quickly. |
| Instructor background | Relevant professional experience is more informative than popularity alone. |
| Labs, projects and assessments | Practice creates evidence; watching videos does not prove mastery. |
| Deployment and troubleshooting | Real jobs require operating systems under imperfect conditions. |
| Total cost | Include exams, cloud resources, lab access and possible renewals. |
| Credential recognition | Confirm whether the employer or professional body actually values it. |
Udemy’s pricing page distinguishes a Personal Plan from individual course purchases and says its plans draw from a catalog of more than 250,000 courses and reference 80 million learners. Those are changeable, company-reported figures; the page did not expose one universal price for every country, account, device or promotion. Individual buyers should check the live page, course update date, reviews, exercises and instructor history.
Rank #4
What employers should measure instead of hours watched
Organizations need a skills inventory tied to actual roles, then role-based pathways with hands-on practice, assessments and internal mobility. Training teams should evaluate assessment performance, work adoption, quality, productivity, risk reduction and movement into new responsibilities—not merely enrollments or completion percentages.
Udemy says its business offering includes more than 30,000 curated courses, 16 languages and more than 17,000 enterprise customers, with a Business Pro add-on offering assessments, labs, workspaces and learning paths (Udemy Business). These are vendor-reported claims. Buyers should also examine content-refresh cadence, privacy and security terms, learning-management integration, accessibility, support and whether the vendor can prove outcomes for their own workforce.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What job-creation forecasts can—and cannot—tell workers
Rogers cited World Economic Forum estimates that AI could displace 85 million jobs while creating 97 million new jobs. He also cited estimates of $2.6 trillion to $4.4 trillion in annual economic gains from GenAI and potential time savings of 60% to 70% on mundane tasks. These are forecasts quoted in the 2024 interview, not observed employment outcomes.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
New jobs may require different skills, appear on a different timetable or arise in different places from displaced work. Productivity gains can benefit employers without immediately increasing headcount. Roles may be redesigned rather than eliminated, and an aggregate net-job estimate cannot guarantee an individual transition or comparable pay.
How much confidence should you place in Udemy’s claims?
Udemy can show what people and business customers consume on its platform. It cannot, by itself, show the total labor-market demand for a skill, that a course caused a hire, that a credential improved performance or that GenAI will create more jobs than it displaces. Treat its figures as a window into learning activity, then verify career decisions against current job postings, employer requirements and demonstrable work.
For individuals, a targeted course or subscription can be a flexible, relatively low-cost way to close a defined gap; it is a poor substitute for live instruction, accredited education, mentorship or guaranteed placement. Employers comparing providers can also consider Coursera, LinkedIn Learning, Pluralsight and edX, matching the choice to credential needs, technical labs, integrations, assessments and measurable outcomes rather than catalog size alone.
The practical takeaway
The strongest lesson from Rogers’ interview is not “take a prompt-engineering course.” It is to combine AI fluency with domain expertise, data and security judgment, communication and evidence of results. That combination helps explain why layoffs and vacancies can coexist—and gives a job seeker a more reliable way to decide what to learn next.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
