AI and Jobs in Software

May 9, 2026, 16:50
USTechnology
A2

AI and Jobs in Software

Introduction

Many companies now use AI to write software. This makes work faster, but it changes jobs for people.

Main Body

Big companies like Alphabet and Anthropic use AI to write a lot of code. Some companies use AI for 90% of their work. Now, human workers do not write all the code. They check the AI's work instead. Because of AI, some people lose their jobs. In the US, many tech workers lost jobs in April. Company leaders want to spend money on AI instead of people. They think AI can do the work of middle-level engineers. Some people learn to code using AI without going to college. They can make helpful tools for medicine or jobs. But AI is not perfect. Humans must still watch the AI to make sure it is correct.

Conclusion

AI helps companies make software faster. However, it also means fewer jobs for software engineers.

Learning

The 'Instead Of' Pattern

In the text, we see a very useful phrase for A2 students: instead of.

We use this when we change one thing for another.

Examples from the text:

  • Spend money on AI →\rightarrow instead of →\rightarrow people.

How to use it in your life:

  • I drink tea →\rightarrow instead of →\rightarrow coffee.
  • I walk to work →\rightarrow instead of →\rightarrow taking the bus.

Simple Word Pairs

Notice how the author connects ideas using simple opposites:

FasterFewer
AI makes work faster.There are fewer jobs.

Tip: 'Fewer' is used for things we can count (like jobs, apples, or people).

Vocabulary Learning

software (n.)
Computer programs that help you do tasks
Example:The software helps me edit photos.
jobs (n.)
Positions where people work for pay
Example:There are many jobs in the city.
work (v.)
To do a job or task
Example:I work at the library.
faster (adj.)
Quicker than something else
Example:The new machine is faster than the old one.
human (adj.)
Relating to people, not machines
Example:Humans can solve problems that computers cannot.
learn (v.)
To gain new knowledge or skills
Example:I learn new skills every day.
college (n.)
An institution where people study after high school
Example:She goes to college to study biology.
B2

The Use of Agentic AI in Software Development and Its Effect on Jobs

Introduction

Many companies are now using agentic artificial intelligence to automate software engineering. This trend is leading to higher productivity, but it is also changing employment patterns within the technology sector.

Main Body

The use of agentic AI has become a key goal for company leaders. Data shows that integration varies across firms: Anthropic reports that AI generates 90% of its code, while Alphabet and DoorDash report 50% and 66% respectively. Other companies, such as Chime and Airbnb, have seen a significant increase in how quickly they release code. Consequently, the role of human engineers is changing; instead of writing all the code themselves, they are now focusing on managing and supervising autonomous AI tools. At the same time, there is a clear link between AI investment and job cuts. According to analysis from Challenger, Gray & Christmas, AI was the main reason for U.S. job losses for two months in a row. The tech sector is the most affected, with total layoffs increasing by 33% compared to last year. Some executives, such as those at Uber, have emphasized that they prefer investing in AI over hiring more people to increase output. Furthermore, Mark Zuckerberg suggested that by 2025, AI will be able to perform the tasks of mid-level engineers, which has created a feeling of instability among workers. Despite these challenges, AI tools are making it easier for people without traditional degrees to learn programming. Open-source platforms and AI assistants allow individuals to gain technical skills independently. For example, some self-taught developers are now contributing to medical algorithms or using AI for specialized tasks. However, these tools are not perfect. Some experimental techniques have been shown to reduce the logical reasoning of AI models, which proves that high-quality results still require professional human oversight.

Conclusion

In summary, the current situation is a mix of two trends: AI is making product development faster and more efficient, but it is also causing a decrease in traditional software engineering jobs.

Learning

⚡ The 'Logic Link': Moving from Simple to Complex Sentences

At the A2 level, you usually write short, separate sentences. To reach B2, you must connect your ideas to show how they relate. The provided text is a goldmine for Connectors of Result and Contrast.

🛠️ The Transition Tool: "Consequently"

In the text, we see: "...they release code. Consequently, the role of human engineers is changing."

Instead of using "so" (which is very basic), B2 speakers use Consequently or Therefore.

  • A2 Style: AI is fast, so people lose jobs.
  • B2 Style: AI is increasing productivity; consequently, traditional roles are evolving.

⚖️ Balancing Opposites: "Despite" vs "However"

Notice how the author switches from bad news (job losses) to good news (learning opportunities):

  1. Despite [Noun/Phrase]: Used to show a surprising contrast at the start of a sentence.

    • Example: "Despite these challenges, AI tools are making it easier to learn."
    • Pro Tip: Never put a full sentence (subject + verb) immediately after "Despite". Use a noun phrase.

  2. However: Used to pivot the conversation after a full stop.

    • Example: "...these tools are not perfect. However, some techniques..."

🚀 Quick Upgrade Guide

Try swapping your A2 words for these B2 alternatives found in the text:

A2 (Basic)B2 (Professional)Context from Text
A lot ofA significant increase"...a significant increase in how quickly..."
ChangeTransform/Evolve"...employment patterns... are changing"
ShowEmphasize/Prove"...have been shown to reduce..."

The B2 Mindset: Stop thinking in "dots" (sentence. sentence. sentence.) and start thinking in "bridges" (Sentence →\rightarrow Connector →\rightarrow Sentence).

Vocabulary Learning

agentic (adj.)
Having the power to act independently and make decisions
Example:The agentic AI can choose which code snippets to write on its own.
automation (n.)
The use of machines or software to perform tasks without human intervention
Example:Automation has reduced the time needed to test software.
productivity (n.)
The amount of work produced per unit of time
Example:Higher productivity means more features released each month.
employment (n.)
The state of having a job or being hired
Example:Employment rates in tech have fluctuated due to AI.
integration (n.)
The process of combining different systems or components into one
Example:Integration of AI tools into existing workflows can be challenging.
significant (adj.)
Notably large or important
Example:There was a significant increase in code releases after adopting AI.
supervising (v.)
Overseeing and directing the work of others
Example:Human engineers are now supervising AI-generated code.
autonomous (adj.)
Capable of operating independently without external control
Example:Autonomous AI tools can debug code without human input.
investment (n.)
The act of putting money or resources into something with expectation of return
Example:Companies are making large investments in AI research.
layoffs (n.)
The termination of employees from a company
Example:Layoffs increased by 33% in the tech sector last year.
executives (n.)
High-level managers or directors in a company
Example:Executives at Uber prefer AI over hiring more staff.
instability (n.)
Lack of stability or certainty
Example:Workers feel instability as AI replaces mid-level roles.
open-source (adj.)
Software whose source code is freely available for anyone to use or modify
Example:Open-source platforms allow developers to share AI models.
independently (adv.)
On one's own, without assistance from others
Example:Learners can acquire skills independently using online tutorials.
experimental (adj.)
Based on testing new ideas that have not been proven
Example:Experimental techniques can sometimes reduce AI’s logical reasoning.
logical reasoning (n.)
The ability to think in a clear, rational, and systematic way
Example:Logical reasoning is essential for debugging complex systems.
professional (adj.)
Relating to a paid occupation or skilled work
Example:Professional oversight ensures AI outputs meet quality standards.
oversight (n.)
Supervision or monitoring to ensure correctness
Example:Human oversight is needed to correct AI mistakes.
product development (n.)
The process of creating and improving products
Example:AI speeds up product development cycles.
efficient (adj.)
Achieving results with minimal waste of time or resources
Example:Efficient workflows reduce development time.
C2

The Integration of Agentic AI in Software Development and its Correlation with Workforce Restructuring

Introduction

Corporate entities are increasingly adopting agentic artificial intelligence to automate software engineering, leading to measurable gains in productivity and a simultaneous shift in employment trends within the technology sector.

Main Body

The adoption of agentic AI has transitioned from a specialized utility to a primary performance metric for executive leadership. Institutional data indicates a broad spectrum of integration: Anthropic reports that AI generates 90% of its code, while Alphabet and DoorDash report figures of 50% and approximately 66%, respectively. Other firms, such as Chime and Airbnb, have noted substantial increases in code shipment velocity, with Chime's AI-developed code rising from 29% to 84% over a four-month period. This systemic shift is characterized by a transition toward 'agentic workflows,' where human engineers pivot from primary authorship to the orchestration and supervision of autonomous digital task forces. Parallel to these productivity gains, a correlation has emerged between AI investment and workforce reduction. Analysis from Challenger, Gray & Christmas indicates that AI was the primary driver of U.S. job cuts for two consecutive months, accounting for 26% of layoffs in April. The technology sector remains the most affected, with total year-to-date cuts increasing by 33% compared to the previous year. While some executives, such as those at Uber, explicitly state a strategy of prioritizing AI investment over headcount growth to increase throughput, others, like Mark Zuckerberg, hypothesize that AI will effectively function as mid-level engineers by 2025. This institutional pivot has fostered a climate of professional instability, with a significant portion of the workforce anticipating role elimination due to automation. Despite these systemic pressures, the democratization of programming tools continues to facilitate non-traditional career trajectories. The emergence of open-source platforms and AI-assisted learning has enabled individuals to acquire advanced technical competencies independently of formal academic credentials. This is evidenced by the ability of self-taught developers to contribute to high-level research, such as FDA-cleared medical algorithms, or to utilize AI for niche applications like automated job procurement and vulnerability testing. However, the efficacy of these tools remains variable; experimental prompt engineering techniques, such as 'caveman speak,' have been found to degrade the logical reasoning capabilities of models like Claude, suggesting that high-fidelity output still requires sophisticated human oversight.

Conclusion

The current landscape is defined by a duality where AI-driven operational efficiency is accelerating product development while simultaneously precipitating a contraction in traditional software engineering employment.

Learning

The Architecture of Nominalization and 'Conceptual Density'

To move from B2 to C2, a student must stop describing actions and start describing phenomena. The provided text is a masterclass in nominalization—the process of turning verbs (actions) into nouns (concepts). This is the hallmark of high-level academic and corporate English, as it allows the writer to pack complex causal relationships into a single sentence without relying on repetitive conjunctions.

🔍 The 'Action' vs. 'Phenomenon' Pivot

Observe how the text avoids simple subject-verb-object structures in favor of dense noun phrases. Compare these two versions of the same idea:

  • B2 Level (Action-oriented): Companies are using AI to do their work, and because of this, they are cutting jobs.
  • C2 Level (Concept-oriented): *"A correlation has emerged between AI investment and workforce reduction."

In the C2 version, "investing in AI" becomes AI investment and "reducing the workforce" becomes workforce reduction. By transforming these actions into nouns, the writer can now treat them as variables in a mathematical or logical correlation.

🛠️ Dissecting the 'Systemic Shift' Syntax

Analyze this specific phrase:

*"...a transition toward 'agentic workflows,' where human engineers pivot from primary authorship to the orchestration and supervision of autonomous digital task forces."

The C2 Linguistic Engine here:

  1. The Pivot: Instead of saying "engineers stop writing and start managing," the author uses "pivot from [Noun A] to [Noun B]."
  2. Abstract Layering: "Primary authorship" and "orchestration and supervision" are not just words; they are conceptual categories.
  3. The Modifier Stack: "Autonomous digital task forces" uses three adjectives to modify one noun, creating a precise, technical image without needing a lengthy descriptive clause.

🎓 Application for Mastery

To synthesize this in your own writing, replace causal verbs (because, so, leads to) with relational nouns (correlation, catalyst, precipitating factor, duality).

  • Instead of: "AI makes software development faster, but it also makes engineers lose their jobs."
  • C2 Synthesis: "The acceleration of product development is simultaneously precipitating a contraction in employment."

Key C2 Vocabulary extracted for this phenomenon:

  • Precipitating (acting as a catalyst)
  • Contraction (the process of becoming smaller/reducing)
  • Democratization (making something accessible to all)
  • High-fidelity (exactness/precision of output)

Vocabulary Learning

agentic (adj.)
Capable of acting independently and making decisions.
Example:The new software system is agentic, allowing developers to set their own priorities.
orchestration (n.)
The arrangement and coordination of components to achieve a desired outcome.
Example:Effective orchestration of microservices is essential for scalable cloud applications.
supervision (n.)
The act of overseeing or directing the work of others.
Example:Human engineers now focus more on supervision rather than coding.
autonomous (adj.)
Operating independently without external control.
Example:The autonomous robot completed the task without human intervention.
systemic (adj.)
Relating to or affecting an entire system.
Example:The systemic shift in workflow design required a company-wide training program.
correlation (n.)
A mutual relationship or connection between two or more variables.
Example:There is a strong correlation between AI adoption and workforce reduction.
headcount (n.)
The total number of employees in an organization.
Example:The firm reduced its headcount by 15% to cut costs.
throughput (n.)
The amount of work completed within a given period.
Example:Throughput increased by 30% after implementing the new automation tool.
pivot (v.)
To shift focus or direction in strategy or operations.
Example:The company pivoted from product development to service delivery.
instability (n.)
The state of being unpredictable or lacking stability.
Example:The rapid changes in the market created a sense of instability among employees.
democratization (n.)
The process of making something accessible to all people.
Example:The democratization of programming tools has lowered the barrier to entry.
non-traditional (adj.)
Not following conventional or established patterns.
Example:Non-traditional career trajectories are now common in tech.
open-source (adj.)
Software whose source code is freely available to the public.
Example:Open-source platforms enable collaborative development worldwide.
advanced (adj.)
Highly developed or sophisticated.
Example:Advanced algorithms can detect subtle patterns in large datasets.
competencies (n.)
Skills or abilities required to perform a job effectively.
Example:The training program focused on building technical competencies.
FDA-cleared (adj.)
Approved by the Food and Drug Administration for medical use.
Example:The new diagnostic tool is FDA-cleared and ready for deployment.
procurement (n.)
The act of acquiring goods or services.
Example:Automated procurement systems streamline vendor selection.
vulnerability (n.)
A weakness that can be exploited to compromise security.
Example:Regular vulnerability testing protects the system from cyber attacks.
efficacy (n.)
The ability to produce a desired or intended result.
Example:The efficacy of the new algorithm was verified through rigorous testing.
variable (adj.)
Not constant; subject to change or variation.
Example:Performance metrics can be variable depending on workload.
experimental (adj.)
Based on or involving trial, testing, or research.
Example:Experimental prompt engineering techniques are still under investigation.
degrade (v.)
To lower in quality, value, or effectiveness.
Example:Poorly designed prompts can degrade the model's logical reasoning.
logical (adj.)
Based on clear reasoning and sound principles.
Example:Logical consistency is crucial for reliable software behavior.
high-fidelity (adj.)
Of very high quality or accuracy, especially in reproducing details.
Example:High-fidelity simulations provide realistic training scenarios.
sophisticated (adj.)
Highly complex or refined in design or execution.
Example:Sophisticated security protocols protect sensitive data.
duality (n.)
The state of having two distinct aspects or functions.
Example:The duality of AI as both tool and collaborator is evident.
precipitating (v.)
Causing an event or situation to happen suddenly.
Example:Rapid automation precipitating job cuts has sparked debate.
contraction (n.)
A reduction in size, number, or scope.
Example:The industry experienced a contraction in hiring last quarter.
operational (adj.)
Relating to the functioning or execution of a system.
Example:Operational efficiency can be measured by output per hour.
efficiency (n.)
The ability to achieve desired results with minimal waste or effort.
Example:Improving efficiency often involves streamlining processes.