Myths in IT: 30 beliefs about getting in, money, careers and learning worth fact-checking

IT myths outlive framework versions. "You can't get in without a degree", "too late after 35", "everyone earns five grand", "AI will replace everyone": somebody heard it in a chat once, retold it, and off it went. Some of it is pure nonsense, some is half true, and some was true in 2019 and went stale long ago.

This hub has 30 myths in four groups. For each one: what is actually true, what the data says, and what to do. Where a myth is partly true we say so, because "all of it is false" is a myth too. Where we could not find a reliable source we mark it "unverified" instead of inventing a number.

For CVs read the developer CV guide, for ATS the ATS hub, for cover letters the cover letter hub, for communities the IT communities hub, and for freelance platforms the freelance hub. Here we only point to them.

About the numbers. The Ukrainian market is mostly measured by DOU (twice a year, winter and summer), the global one by the Stack Overflow Developer Survey, GitHub Octoverse, Indeed Hiring Lab and academic papers. This is a snapshot at November 2026. Salaries are monthly, in USD, net, and medians unless stated otherwise. A median means "half earn less, half earn more", not "this is what awaits you".

Quick table

Group 1. Getting into IT

1. "You can't get into IT without a computer science degree"

What is true. A degree helps at the start, especially at big companies abroad, where it is sometimes a formal visa requirement. But for most employers in Ukraine and many remote ones, what you can show decides.

Evidence. Stack Overflow Developer Survey 2025: a large share of professional developers hold a bachelor's degree (the figure comes from a podcast summary and we did not match it on the survey site (unverified)), and school is just one source of learning next to documentation (68% of respondents), online resources (59%) and Stack Overflow itself (51%). DOU counts specializations, not diplomas, so we found no honest Ukrainian number for "how many have no relevant degree".

Takeaway. Do not hide behind a degree and do not be ashamed of lacking one. Replace it with proof: projects, commits, reviews, a test task you passed. CV structure without a degree is in the CV guide and the career switcher template.

2. "You must be great at math"

What is true. Most web, mobile and backend work needs school math plus logic. Real math matters in ML, graphics, games, cryptography and signal processing.

Evidence. We found no public survey of "how much math on which roles" (unverified). Rule of thumb: front-end, QA and backend CRUD postings almost never ask for more than basic algebra, while Data Science and ML postings almost always do.

Takeaway. Pick a direction with that in mind instead of ruling out IT. Algorithmic thinking (loops, conditions, data structures) is trained by practice, not by a calculus textbook. The free ShortLesson lessons let you try logic on short tasks at no risk.

3. "Too late after 30, 35, 40"

What is true. It is never too late, but ageism exists and in some companies it is real. The older the newcomer, the more it matters to show concrete value from previous experience (domain, communication, management).

Evidence. DOU, summer 2026 (14,631 respondents): the median age of a Ukrainian IT specialist is 31. So 30 plus is the norm, not the exception. We found no age breakdown of "how many enter after 35" (unverified). An academic paper on discrimination in developer careers exists (arXiv 2510.22457); we did not go through it in detail.

Takeaway. Sell the combination: "an accountant who learned to code to automate things" is worth more than "one more junior". Aim at niches where your previous domain already counts.

4. "A 3-month course and you're a developer"

What is true. In three months you can learn to write working code. Becoming a hired developer in that time rarely happens, because employers need proof of skill, not a completion certificate.

Evidence. DOU, March to April 2026: the median job search in Ukraine takes 11 weeks, and that is across the whole market, not for beginners. For beginners competition grew: about 47 applications per entry-level vacancy in September 2026 versus 26 in January (DOU, secondary source, via dev.ua and kurs.com.ua).

Takeaway. Plan for 9 to 18 months from zero to a first job (a rule of thumb, not a statistic) and budget for that period. A course is a start, not a finish.

5. "Bootcamps guarantee a job"

What is true. Nobody can guarantee a job in IT. Some bootcamps do provide structure and community, but figures like "93% placement" are often marketing.

Evidence. Course Report 2025, as relayed, reports a high share of graduates employed full time, but the data is self-reported (unverified). CIRR-audited figures in secondary sources are lower than the advertised ones (unverified). A Yahoo investigation found a school advertising 74% with a real rate around 30%. We found no honest year-over-year trend.

Takeaway. Ask for three things: third-party audit, the definition of "employed" (full time in field or anywhere), and numbers for the latest cohorts. If the school will not give them, that is your answer. For honest reviews, see the communities hub.

6. "IT is only for introverts"

What is true. We found no reliable share of introverts among developers (unverified). What is visible is the work itself: development is teamwork with code review, standups, and explaining things to business and clients.

Evidence. DOU and Stack Overflow describe roles, not personalities. A reading of job postings: "communication" and "English" sit next to technical requirements almost always.

Takeaway. IT suits introverts who do not avoid communication entirely: train written communication (Slack, PR descriptions, docs), which is the main channel in remote teams anyway. Extroverts have plenty of room too, since talking to people gets more valuable with seniority.

7. "Women don't belong in IT"

What is true. The oldest myth on the list and the easiest to check.

Evidence. DOU, summer 2026: women are 31% of Ukrainian IT specialists, up from 29% a year earlier and 23% in 2021. In the global Stack Overflow survey the share of women among respondents is much lower (per older SO data, not checked for 2025 (unverified)), with differences in sample and country. The share of women among technical specialists in Ukraine is growing more slowly than overall (up 1 percentage point in a year per DOU).

Takeaway. Look for supportive communities (see the women and inclusive communities section of the global communities hub). Interview practice helps rehearse answers to questions women are asked more often, such as salary expectations.

Group 2. Market and money

8. "Everyone in IT earns $5000"

What is true. Not everyone earns that, and certainly not at the start.

Evidence. DOU, winter 2026 (data as of December 2025, 4,522 questionnaires, net): developer median $3,450, junior $900, middle $2,500, senior $4,500. About 9% of Ukrainian IT people earn over $6,000 (8.9% in December 2025, and the same in 2023). In the summer 2026 survey the developer median is about $3,500 (secondary source, via dev.ua).

Takeaway. Use the median of your level, not screenshots from Telegram. Count the path: junior $900 to middle $2,500 is a 2.8x jump over a few years, and that is real career arithmetic. See the Telegram hub for channels with salary cuts.

9. "The junior market is dead"

What is true. Not dead, but much harder than in 2021, and the numbers show it. The honest answer: dead for those who do not stand out, alive for those who are prepared and search well.

Evidence. DOU (September 2026, secondary source): 47 applications per entry-level vacancy versus 26 in January, and the number of such vacancies fell by about 31% in nine months. Front-end had about 275 applications per vacancy in September (secondary source). US: Stanford, "Canaries in the Coal Mine" (Brynjolfsson, Chandar, Chen): employment of young software developers was nearly 20% below its late 2022 peak as of July 2025; the August 2026 update sees no economy-wide mass displacement, but the drop for young workers in AI-exposed jobs remains. Indeed Hiring Lab: US software development postings are still about 27.5% below February 2020 (June 2026), and 71% of the net increase from May 2025 to May 2026 came from senior roles.

Takeaway. Do not compete as "one more front-end CV". Look at less crowded directions (QA, DevOps, data, hardware; DOU showed hardware at about 3.5 applications per vacancy), and let the first job be adjacent (support, QA, technical analysis). A practical plan is in the CV guide and the communities hub: referrals beat cold applications.

10. "AI will replace all developers"

What is true. It has not replaced everyone so far, but the entry level feels pressure and more tasks get automated. A myth in its absolute form, a risk in a moderate one.

Evidence. Stack Overflow 2025: trust in AI answer accuracy fell to 29% (from 40%), and 75.3% of respondents do not trust AI answers. METR (2025, randomized study of 16 experienced open-source developers, 246 tasks): with AI they were 19% slower while believing they were 20% faster (early-2025 tools, newer models may differ). Indeed: despite AI, US developer postings have grown since 2025 (+14.6% since late February 2025, secondary source).

Takeaway. Learn to work with AI as a tool and to check it, rather than fear or ignore it. Value shifts toward understanding the problem, architecture, review and communication. See myth 24 and 25.

11. "Remote is everywhere"

What is true. Remote really dominates Ukrainian IT, but the trend is not one-way: the office share ticked up.

Evidence. DOU, summer 2026: 66% work fully remote, while the share "only or mostly office" rose from 11% to 13%. Fully remote is more common abroad than in Ukraine. For juniors there are fewer remote openings because companies want mentoring close by (rule of thumb from observation (unverified)).

Takeaway. Do not dismiss hybrid in your city: for a first job it often teaches more. For remote, look for companies with internships and strong onboarding.

12. "Freelance is easier money"

What is true. Freelance is a different set of difficulties, not an easier one: finding clients yourself, setting prices, handling taxes, living without a steady flow.

Evidence. DOU on the $6K earners: freelance, startup, product and outstaffing companies raise the odds of such pay (secondary source), but that is a sample of those who made it, not a guarantee. Rates, fees and platform rules are covered separately in the freelance hub, including red flags.

Takeaway. A first freelance gig usually comes after 1 to 2 years of experience, not instead of it. Projects with no mentor and no client base are slow and nerve-wracking at the start.

13. "Salary grows automatically every year"

What is true. It does not. There are years without raises, and middles and seniors in the Ukrainian market have gone down too.

Evidence. DOU: the developer median in June 2026 is about $3,500, level with June 2024 (secondary source). Winter 2026: senior down $200 in half a year, junior down $30. The share earning over $6,000 did not change in two years (about 9%).

Takeaway. Ask for raises with arguments (results, DOU market medians for your role and level), do not wait. Once a year compare your rate with the median.

Group 3. Work and career

14. "Senior means N years of experience"

What is true. Years are needed, but level is set by scope of responsibility and independence, not the calendar. Two people with 5 years can be at different levels.

Evidence. DOU: after 10 years in IT the odds of earning over $6,000 rise markedly (secondary source), so years matter, but not linearly. There is no industry standard of "what a senior is"; titles differ between companies (we found no single standard (unverified)).

Takeaway. Keep a list of your own "senior" achievements: designed a system, mentored, owned a release, ran incident reviews. That goes into your CV by the formula in the guide.

15. "You must switch jobs every year"

What is true. Switching often raises pay, but not always, and it should not be done blindly.

Evidence. DOU, summer 2026: among middle-level developers, recent joiners often earn more than those who stayed. Among juniors the median is lower (they usually get a review after the first year), and among senior developers there is no difference. 37% of IT people in Ukraine left a company within a year (45% of those were dismissed, 46% left on their own).

Takeaway. Stay the first year: you learn and get a review. Switch when you have outgrown the role, not by the calendar. Keep your CV and letters ready in advance.

16. "Managers stopped coding, so they are worse"

What is true. "Worse" skips the question "at what". A manager solves different problems and code is not the main skill. But a technical manager who does not understand the context really does slow the team.

Evidence. Engineering managers in Ukraine earn up to $8,000 a month per DOU (secondary source, scroll.media), so the market values the role. We found no surveys on "manager quality versus coding level" (unverified).

Takeaway. Before choosing a track (IC or management), try running code reviews, delegating some code and working with people. If you do not want to go back to code afterwards, that is the answer.

17. "Only FAANG matters"

What is true. It is a nice name on a CV, but the path in IT is not limited to it. Most developers in the world and in Ukraine work elsewhere.

Evidence. DOU: several company types (startup, product, outstaffing) raise the odds of high pay, not only a big brand (secondary source). We found no DOU share of people who work at FAANG (unverified).

Takeaway. Choose by growth, mentors, stack and rate, not logo. One strong product on your CV is worth more than a year in a faceless big company.

18. "Certificates get you hired"

What is true. Employers rarely count a certificate without practice. Exceptions: cloud certifications (AWS, Azure, GCP), security, and outsourcing partnerships where certified people give the company a status.

Evidence. We found no independent data on the effect of certificates on hiring (unverified). Rule of thumb: in postings a certificate almost always sits in "nice to have", not "required".

Takeaway. The certificates worth getting are those that cover a requirement of a specific vacancy. The rest is time better spent on a project.

19. "75% of CVs are rejected by robots before a human sees them"

What is true. A popular number with no primary source. An ATS is a candidate database, not a judge. Auto-rejection exists, usually through knockout questions (work authorization, country), not for a "bad CV".

Evidence. Greenhouse support pages (vendor): auto-reject is configured by the employer, and they do not turn on AI scoring of candidates. We found no primary source for the 75% figure. Full breakdown in the ATS hub.

Takeaway. Write the CV for a human, keep it parser-friendly (plain structure, no tables in the middle), and answer knockout questions honestly.

20. "You must know everything before applying"

What is true. A job description is the employer's wish list, not a hard minimum. But do not ignore the "required" lines either.

Evidence. We found no public study of "what share of requirements those who got hired met". The often-cited claim that women apply only at a full match while men apply at a partial one has a murky origin (unverified).

Takeaway. If you cover roughly two thirds of the key requirements and can show how you will learn the rest, apply. In the letter and at the interview say plainly what you do not know and how you will close the gap fast. Templates in the cover letter hub.

21. "English isn't needed in Ukraine"

What is true. For some local employers basic English is enough, but the ceiling without it is lower and market access narrower.

Evidence. DOU: English level noticeably affects median pay. How much depends on level and role: for junior QA there is no significant difference, for managers each next level adds about $650 to $750 (secondary source), and for developers estimates differ, with no single number (unverified). DOU also names English among the factors that raise the odds of earning over $6,000.

Takeaway. B2 for reading and writing at a minimum, and speaking practice through mock interviews. For the interviews themselves you can rehearse phrasing in ShortLesson interview practice.

Group 4. Learning

22. "Learn the hottest framework first"

What is true. Popular means more openings but also more competition. Fundamentals (a language, HTTP, databases, git, debugging) transfer, frameworks do not.

Evidence. GitHub Octoverse 2025: TypeScript became the number 1 language by contributor count, up 66% in a year (GitHub partly ties this to AI tooling). Nearly 80% of new repositories use only six languages. DOU: front-end sits at about 275 applications per vacancy (secondary source), the most popular and the most crowded.

Takeaway. First one language and the fundamentals, then a popular stack aimed at your goal. Look at postings in your city and country, not at rankings.

23. "Tutorials are enough"

What is true. A tutorial gives an illusion of understanding: you reproduce someone else's solution instead of solving your own problem.

Evidence. Stack Overflow 2025: the top learning sources are documentation (68%) and online resources (59%), not step-by-step videos. We found no direct "tutorials versus projects" study (unverified).

Takeaway. The 1 to 1 rule: an hour of tutorial, an hour of your own code without peeking. Build something that does not match the lesson. ShortLesson lessons are short and hands-on rather than hour-long lectures.

24. "Copying from AI is cheating"

What is true. Using AI is not cheating in itself, most people do it. The problem starts when you do not understand the code you hand in.

Evidence. GitHub Octoverse 2025: 80% of new developers on GitHub use Copilot in their first week. Stack Overflow 2025: 44% of developers learn with AI (37% last year), and at the same time 61.3% want to fully understand their code.

Takeaway. Use AI as a tutor: ask it to explain, not only to write. Rule: if you cannot explain a line out loud, do not submit it. It shows quickly at an interview or in a test task.

25. "AI makes learning useless"

What is true. The opposite: to check what AI wrote you need to know more than to write it slowly yourself.

Evidence. METR: experienced developers with AI were 19% slower, because they spent time reviewing and fixing (early-2025 tools). SO 2025: only 29% trust AI accuracy.

Takeaway. Learn the fundamentals just as seriously and use AI to speed up. For tools see the series on AI assistants.

26. "LeetCode is all that matters"

What is true. For big product companies abroad, algorithm rounds really are common. For most Ukrainian outsourcing and product companies the interview is mixed: theory, a practical task, a conversation about experience.

Evidence. We found no public statistic of "what share of interviews include algorithm problems" (unverified). Rule of thumb: front-end and backend postings in Ukraine more often mention a test task and a technical conversation.

Takeaway. Pick your target company and prepare for its format. If the target is a top product company, you need LeetCode (30 to 60 minutes a day). Otherwise basic data structures and practice explaining a solution out loud are enough. Train in ShortLesson interview practice.

27. "All theory first, practice later"

What is true. That is the road to burnout. Theory without a problem does not stick, and practice without theory soon hits a ceiling.

Evidence. No direct data in our sources (unverified); the advice rests on documentation and online resources leading among learning methods (SO 2025), that is, learning by solving tasks.

Takeaway. Alternate: a small task, then the theory needed to solve it, then another task.

28. "Pet projects without users are worthless"

What is true. Projects with users are worth more, but a project without them still shows you can finish something.

Evidence. We found no surveys on the effect of pet projects on hiring (unverified). Octoverse shows new developers are very active on GitHub: "a new developer every second" (GitHub's estimate).

Takeaway. One well-finished project beats ten half-written ones: README, deployment, tests, a note on decisions. Show it on your CV by the formula in the guide.

29. "You can't find a job without a network"

What is true. You can, but slower. Referrals and personal messages are often faster than cold applications.

Evidence. DOU: median job search 11 weeks across the market, and the new data is titled "more searching, less optimism". We did not extract a DOU figure for the share of jobs found through referrals (unverified).

Takeaway. Start with Ukrainian and international communities, answer questions, share what you learned. Pair a cold application with a message to someone at the company.

30. "A rejection means you're a bad developer"

What is true. A rejection most often means the match of requirements and experience was partial, or a stronger candidate appeared, or the company changed plans.

Evidence. DOU: 17% of IT people lost a job last year, 37% left a company within a year, and 45% of those were dismissed. So layoffs hit strong people too.

Takeaway. Keep a table of applications and reasons (where you can learn them) and change one variable at a time: CV, letter, preparation. A review of trackers is in the tools series.

How ShortLesson helps

Most myths survive because they are hard to check. So we build short free lessons where a few minutes of practice replace an hour of theory, interview practice where you can rehearse answers out loud before the real thing, and career hubs with verified sources, where we say plainly what we do not know.

Sources

All data as of 2026-11-13. Anything not marked "secondary source" or "unverified" we read at the primary source or the publisher's page.

• DOU, salary survey, winter 2026 (via dev.ua): medians $3,450, junior $900, middle $2,500, senior $4,500.
• DOU, Portrait of the IT specialist 2026: 31% women, median age 31, 66% fully remote, 17% laid off.
• DOU, More searching, less optimism (job market 2026, part 1): 11 weeks of search, 37% left a company.
• DOU, Who earns $6K: share above $6,000 (about 9%), role of English and company type; figures via coverage, full text not read.
• DOU, Do IT people really take lower pay when changing jobs: pay of recent joiners by level.
• DOU, 47 applications per vacancy and Forbes Ukraine: junior competition (secondary source, report itself not opened).
• Stack Overflow, Developer Survey 2025 and press release: trust in AI, learning sources. Education and bootcamp figures come from a podcast (unverified).
• Stack Overflow, Domain expertise still wanted: 64% use AI to learn (February 2026, a separate survey, do not mix with 2025).
• GitHub, Octoverse 2025: TypeScript number 1, 80% of new developers use Copilot.
• Indeed Hiring Lab, The labor market is tilting toward seniority: US developer postings versus pre-pandemic, senior share of growth.
• Brynjolfsson, Chandar, Chen, Canaries in the Coal Mine?, Stanford Digital Economy Lab: young developers and AI exposure; August 2026 update.
• METR, Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity: 19% slower, 16 developers.
• Greenhouse support pages on auto-reject (see the ATS hub).
• Course Report and CIRR on bootcamps: through secondary coverage, self-reported data (unverified). Yahoo, biggest coding bootcamps may exaggerate.
• arXiv 2510.22457, A Multifaceted View on Discrimination in Software Development Careers: mentioned, not reviewed in depth.

No reliable source found (marked in the text): effect of certificates on hiring; share of algorithm rounds in interviews; share of introverts; "former-developer managers are worse"; FAANG share; bootcamp placement trends by year; age structure of entry into IT; Stack Overflow 2025 education from the primary source.