PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchiTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
AI coding tools have not been shown to eliminate entry-level software engineering jobs across the labor market. They can raise measured developer output, but hiring studies point in different directions: one U.S. vacancy study finds junior roles becoming relatively less common and more demanding, while a study of companies adopting GitHub Copilot finds an association with more software-engineer hiring, driven by entry-level hires. These studies measure different outcomes, so neither productivity gains nor one hiring estimate settles what will happen to junior jobs overall.
The clearest practical implication is a changing job profile: coding fundamentals still matter, but junior engineers may also need to show they can understand a problem, assess AI-generated code, communicate decisions, and take responsibility for the result. How much that shift affects hiring depends on the employer and the work it needs done.
What the evidence says about jobs—and what it does not
There are two separate questions: can AI help a developer produce more, and does a company then hire fewer or more developers? The first is about productivity; the second is about labor demand. A faster way to complete some tasks could mean fewer labor hours for a fixed amount of work. It could also let a team take on more work, create new tasks, or change how it divides work among employees. The studies available here measure pieces of that chain, not a single final effect for every employer.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →The evidence therefore does not support either blanket claim that AI has wiped out junior software jobs or the opposite claim that AI guarantees more entry-level hiring. Findings differ in population, time period, geography, and method. A controlled experiment where developers use an assistant is not interchangeable with a count of job ads or a study of hiring at firms that adopt a tool.
#1 Best Overall
What productivity studies tell us
More completed tasks in three field experiments
A June 2025 Microsoft Research paper pooled three randomized field experiments at Microsoft, Accenture, and an anonymous Fortune 100 company. Across 4,867 software developers, access to generative AI was associated with a 26.08% increase in completed tasks (standard error 10.3%). The authors report higher adoption rates and greater measured productivity gains among less-experienced developers.
This is evidence about output in the experiments, not a 26.08% increase in hiring, pay, or job availability. It also does not mean every developer or task receives the same benefit. The measure is completed tasks in those study settings; it does not, by itself, answer whether the resulting code was better, whether companies expanded output, or how many employees they needed afterward.
More code output in an Ant Group experiment
The Bank for International Settlements’ 4 September 2024 summary of an Ant Group field experiment examined CodeFuse, introduced in September 2023. It reports a 55% increase in lines of code. About one third of that increase was directly attributable to generated code; the rest was likely due to efficiency improvements elsewhere in the workflow. Statistically significant productivity gains were primarily among junior staff.
Lines of code are a measure of output volume, not quality, employment, or wages. More code is not necessarily more useful code: it may require review, testing, maintenance, or revision. The experiment is informative about how a coding assistant can change measured production, but it does not establish that employers will need 55% fewer developers or that junior hiring will rise by a corresponding amount. See the BIS paper summary for the experiment and its interpretation.
Rank #2
What hiring and vacancy studies find
Junior vacancies became relatively less common in U.S. job ads
A June 2026 IZA Discussion Paper by Samuel Westby, Alicia Sasser Modestino, and Peiran Cheng uses near-universe U.S. Lightcast online-vacancy data and event-study and difference-in-differences estimates around ChatGPT’s public release. The authors report a 14–15% relative decline in junior software developer vacancies versus senior vacancies. This is a change in the junior-to-senior comparison, not a finding that all junior vacancies fell by 14–15%.
The same paper finds experience requirements rose mainly within the same job titles. The remaining junior vacancy descriptions shifted toward problem solving, communication, and attention to detail rather than AI-specific skills. The authors summarize that result this way: “Remaining junior vacancies shifted toward problem solving, communication, and attention to detail, not AI-specific skills.” The paper, Generative AI and the Redefinition of Entry-Level Software Work, describes job ads and their requirements; it does not count every hire or establish that AI alone caused the observed changes.
Copilot adoption was associated with more software-engineer hiring in one firm study
A paper first published by Wiley on 22 April 2026 finds that firms adopting GitHub Copilot had an around 3–5% higher monthly probability of software-engineer hiring. The reported association was driven by entry-level hires. New hires had around 5% more non-programming skills, with no decrease in coding skills.
This is an association between firm adoption and hiring outcomes, not proof that adopting the tool caused the increase. The authors use proprietary LinkedIn and GitHub data and have industry affiliations, considerations readers should keep in mind when interpreting the result. It does not contradict the IZA vacancy finding directly: one study compares hiring probability at adopting firms, while the other examines the relative composition and stated requirements of U.S. online vacancies. Details are in the Wiley paper.
Broader early-career and coding-intensive employment measures
A U.S. Census Bureau Center for Economic Studies working paper published in April 2026 estimates a 12% decline in regression-adjusted employment among workers aged 22–24 in the most AI-exposed quintile of industry-state cells over the ten quarters after ChatGPT’s introduction. This is not a software-engineer-only estimate. The paper says hiring rates largely recovered by early 2025 on a smaller employment base, and notes possible earlier trend shifts around COVID. It should not be read as a direct causal estimate of AI’s effect on entry-level software engineering. The paper is You’re (not) Hired: Artificial Intelligence and Early Career Hiring in the Quarterly Workforce Indicators.
A 2026 paper from the Board of Governors of the Federal Reserve System estimates about 3% lower employment growth after ChatGPT’s introduction for an O*NET/CPS-defined group of coding-intensive workers, after controlling for industry-level shocks. That group is not limited to entry-level software engineers. The authors caution that they cannot control for task reallocation and other relevant factors, so this estimate does not isolate AI as the sole cause. Read AI and Coder Employment: Compiling the Evidence for the group definition and qualifications.
How to compare the findings
The figures below answer different questions. Comparing them is useful for understanding the range of possible effects, but they should not be averaged or treated as competing estimates of one universal number.
Free tools Windows power users keep installed
One-click scans. No signup required.
| Study and setting | Outcome and reported finding | What the finding can—and cannot—show |
|---|---|---|
| Microsoft Research, 2025; randomized field experiments at three companies; 4,867 developers | Completed tasks: 26.08% increase (standard error 10.3%). | Short-run task output in the experiment settings; not a hiring or labor-demand estimate. Less-experienced developers had higher adoption and larger productivity gains. |
| BIS summary of an Ant Group CodeFuse field experiment; tool introduced September 2023 | Lines of code: 55% increase; roughly one third directly attributable to generated code. | Measured output volume, not code quality, employment, or wages; statistically significant gains were primarily among junior staff. |
| IZA, 2026; near-universe U.S. online vacancy data; event-study and difference-in-differences estimates | Junior vacancies declined 14–15% relative to senior vacancies; experience requirements rose mainly within titles. | Relative vacancy composition and stated requirements, not an absolute count of all junior jobs or proof that AI alone caused the shift. |
| Census Bureau Center for Economic Studies, 2026; workers aged 22–24 across industry-state cells | Regression-adjusted employment declined 12% in the most AI-exposed quintile over ten quarters after ChatGPT’s introduction. | A broader early-career employment measure, not specific to software engineers; possible earlier trend shifts around COVID complicate causal interpretation. |
| Wiley, 2026; firms’ GitHub Copilot adoption and linked hiring data | Adoption associated with an around 3–5% higher monthly probability of software-engineer hiring, driven by entry-level hires. | An association using proprietary LinkedIn and GitHub data, not a causal guarantee for adopting firms or the whole market. |
| Federal Reserve, 2026; O*NET/CPS-defined coding-intensive workers | About 3% lower employment growth after ChatGPT’s introduction, controlling for industry-level shocks. | A broader occupation measure, not entry-level engineers alone; task reallocation and other factors remain relevant. |
| Chen et al., arXiv/HICSS, 2024/2025; survey of 32 industry professionals | Most participants said their organizations had not changed candidate evaluation for use or knowledge of code-generation tools; views on allowing tools in interviews were mixed. | Exploratory small-sample qualitative evidence, not a representative poll of employers. |
For example, the Microsoft and BIS results concern output among developers with access to particular assistants. The IZA study looks at the share and requirements of vacancies. The Copilot study links firm adoption with subsequent hiring outcomes. The Census and Federal Reserve papers examine broader employment groups, and the survey asks a small group of professionals about evaluation practices. Different measures, samples, and designs can produce different results without one automatically invalidating another.
What may change in a junior engineer’s work
The evidence points to a plausible shift in how some teams divide work, not a universal job description. If an assistant can draft code or accelerate routine tasks, a junior engineer may spend less time producing an initial implementation and more time understanding the requested change, checking the output, finding edge cases, integrating it with existing code, and explaining trade-offs. That is an interpretation of the productivity and skills findings, not a claim that every employer has already redesigned junior roles this way.
Code still matters. The Copilot study reports no decrease in coding skills among new hires, while the IZA vacancy analysis identifies problem solving, communication, and attention to detail—not AI-specific skills—as the changing emphasis in remaining junior ads. This combination suggests that tool familiarity is most useful when it supports sound engineering judgment rather than standing in for it.
- Problem solving: Clarify what a feature or fix needs to accomplish before asking for or writing code.
- Code review: Check whether generated code fits the project’s conventions, handles edge cases, and does what the task requires.
- Communication: Explain an implementation, raise uncertainty early, and make review feedback actionable.
- Attention to detail: Verify behavior with tests and inspect how a change interacts with existing code.
- Fundamentals: Keep developing programming and debugging skills so that an assistant’s plausible-looking answer can be assessed rather than accepted blindly.
These are sensible capabilities to build, not a checklist that guarantees employment. The studies do not test whether a particular portfolio, course, interview tactic, or amount of AI-tool use improves an individual candidate’s odds.
Recommended Free Tools
What the hiring evidence does not establish
None of these findings predicts a single net labor-market outcome. Productivity improvements can reduce the labor needed for a fixed workload, while lower production costs or faster delivery can make additional work viable. Teams may also reorganize tasks or require different skills. The cited studies do not establish which of these responses will dominate across software employers over the long run.
Best Value
Attribution is another limit. The IZA analysis examines changes around ChatGPT’s release, the Census paper discusses possible earlier shifts around COVID, and the Federal Reserve authors say relevant factors such as task reallocation cannot all be controlled. Hiring can also change with the broader business cycle and other workplace shifts. A change observed after the arrival of generative AI is not automatically a change caused only by AI.
Finally, the available evidence spans early tool rollouts and data through 2025 and 2026. Results depend on the assistant, employer, task, and period studied. They are informative about the changes seen in those settings, but they cannot settle how all companies will hire as tools and workflows evolve.
What candidates and employers can take from this
For candidates, treat AI fluency as a practical tool skill alongside—not instead of—programming fundamentals. Be prepared to explain how you reached an answer, how you tested it, and what you changed after review. That approach demonstrates judgment whether or not a team permits an assistant during an interview; the small Chen et al. survey found mixed views about tool use in interviews and does not represent all employers.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Scan for outdated or missing drivers - takes under a minute3Repair Windows errors before they cause bigger problemsFor employers, increased output per developer does not mechanically determine headcount. A hiring decision also depends on the volume and type of work, review and maintenance needs, and the capability mix a team wants. The evidence makes room for both fewer junior vacancies relative to senior roles in some settings and more entry-level hiring at firms adopting a coding assistant. It does not establish a universal outcome or a career guarantee.
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.
