Should New Developers Still Learn To Code?
AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: Should New Developers Still Learn To Code? on ThorstenMeyerAI.com

Buying for a business?Offer from Amazon

Get business pricing on office and shipping supplies

  • Business-only prices and quantity discounts
  • Tax-exempt purchasing
  • Multiple users, one account, clear invoices
As an affiliate, we earn on qualifying purchases.

TL;DR

AI is writing a large share of code at some companies, but research does not show that programming knowledge is no longer useful. Entry-level hiring is tight, and studies suggest developers still need to specify, review, test and debug AI-generated work.

A new report published by ThorstenMeyerAI.com on November 18, 2026, finds that new developers still have reason to learn programming, even as AI writes much of the code at some companies. The report points to mixed evidence on productivity, fewer entry-level postings and research suggesting that using AI to learn can leave gaps in understanding.

Anthropic says that in May 2026, Claude authored more than 80% of the code merged into the company’s codebase, compared with a share in the low single digits before its Claude Code tool launched in February 2025. The company says engineers direct and review the generated work. It also reports that the typical engineer merged eight times as much code per day in the second quarter of 2026 as in 2024, but cautions that counting lines measures quantity, not quality, and likely overstates the productivity gain.

Research on whether AI makes developers faster is less settled. In a 2025 randomized trial, the nonprofit METR studied 16 experienced open-source developers completing 246 tasks in projects they knew well. Developers took 19% longer with AI, although they expected the tools to save time and later believed they had worked faster. METR says newer results may suggest greater speed-ups by early 2026, but calls that evidence very weak; its newer estimates include no effect within their confidence intervals.

The labor-market signals are also mixed. Indeed data cited in the report show that 4.5% of US software-development postings were entry-level in the first quarter of 2026, while 69.3% were senior. Stanford researchers found employment among 22- to 25-year-olds in AI-exposed jobs was 19% below where it otherwise would have been. At the same time, US software-development postings on Indeed were up about 23% year over year in November 2026, after reaching a low in May 2025.

At a glance
reportWhen: Report published November 18, 2026; fig…
The developmentA November 18, 2026 report weighs AI coding adoption against mixed productivity research and a shrinking share of entry-level software job postings.

Skills Beyond AI-Generated Code

For people entering software work, the evidence points to a change in what programming skill is for, not proof that it is no longer needed. If an AI tool produces code, someone still has to define the task, judge whether the result fits the existing system, find errors and decide whether the design is secure. Those responsibilities require enough technical understanding to spot plausible but incorrect output.

The report also cites an Anthropic learning experiment in which mostly junior developers used AI while learning a new library. On a follow-up quiz, they scored 50%, compared with 67% for those who coded by hand, with the largest gap in debugging. The result does not settle how every learner should use AI, but it is a warning that faster completion may come at the cost of practice and retention.

For employers, the distinction matters when setting entry-level expectations and training plans. If junior staff get fewer chances to write and repair code themselves, they may have fewer opportunities to build the judgment needed to review AI output. The cited data show pressure on entry-level hiring, but do not establish that AI alone caused it.

KitchenAid Artisan 5-Quart Tilt-Head Stand Mixer, Dried Rose (KSM150PS)

KitchenAid Artisan 5-Quart Tilt-Head Stand Mixer, Dried Rose (KSM150PS)

De'Longhi Magnifica Evo Fully Automatic Espresso Machine with Milk Frother, Silver

De'Longhi Magnifica Evo Fully Automatic Espresso Machine with Milk Frother, Silver

Ninja Foodi Smart 6-in-1 Indoor Grill and Air Fryer, 4 qt, FG551

Ninja Foodi Smart 6-in-1 Indoor Grill and Air Fryer, 4 qt, FG551

As an affiliate, we earn on qualifying purchases.

From Code Share to Job Skills

In March 2025, Anthropic chief executive Dario Amodei told an audience at the Council on Foreign Relations that AI could write 90% of code within three to six months and essentially all of it within a year. Anthropic’s later internal figures show substantial AI authorship at its own company, though the timing and extent of that adoption do not establish what is happening across the whole industry.

Other figures cited in the report are less extensive or come from different periods. Google chief executive Sundar Pichai said in October 2024 that AI generated more than a quarter of new code at Google, which engineers then reviewed and accepted. METR cited an estimate in February 2026 that Claude Code authored about 4% of GitHub commits. These figures measure different things at different organizations, so they should not be treated as directly comparable.

Productivity findings also depend on the work being tested. A 2022 experiment with 95 professional developers found that those using GitHub Copilot completed a small JavaScript HTTP-server task 55.8% faster. METR’s later trial involved real tasks in large codebases familiar to experienced developers. The contrast helps explain why results from a short, self-contained coding exercise may not predict performance on complex maintenance work.

“The programmer still needs to specify what is being built, how it fits with other code and whether this is a secure design or an insecure design.”

— Anthropic chief executive Dario Amodei, speaking at the Council on Foreign Relations in March 2025

What the Evidence Cannot Establish

The available figures do not show whether AI is the main reason for the decline in entry-level job postings or the employment pattern reported for younger workers. They also do not establish how much of Anthropic’s reported adoption or output gain will carry over to employers with different tools, workflows and codebases.

Productivity remains difficult to pin down. METR’s 2025 trial measured slower completion on its selected tasks, while the organization’s newer results are too uncertain to give a firm estimate of current speed gains. The report also draws on experiments with different tasks, participants and tools, so their results do not produce a single answer for every developer or workplace.

The Anthropic learning experiment offers a limited finding, not a general verdict on AI-assisted education. The source material does not establish whether different teaching methods, longer practice or other uses of AI would reduce the quiz-score gap. It also does not identify which programming languages, courses or career paths will face the most change.

Signals to Watch in Hiring

The next useful evidence will be sustained data on job postings, employment and hiring by experience level, alongside studies that test current tools on a wider range of real development work. Comparing those measures over time could help distinguish a temporary hiring slowdown from a longer-lasting change in the entry-level route into software careers.

For learners and employers, the report’s evidence supports keeping programming practice while adding work with AI tools. Learners can test generated code, trace bugs and explain design choices rather than accepting an output solely because it runs. Employers can track whether AI changes task completion and code quality in their own teams, while preserving opportunities for junior staff to build and demonstrate those skills. The report does not establish a single curriculum or hiring formula.

Key Questions

Should beginners still learn to code?

The report’s conclusion is yes, but with attention to reading, testing, debugging and reviewing code, as well as specifying what a program should do. AI adoption does not remove the need to understand and evaluate software.

Does AI make software developers faster?

The evidence cited is mixed. A METR trial in 2025 found experienced developers took 19% longer on selected tasks with AI, while a 2022 Copilot experiment found faster completion on a small coding task. METR says its newer evidence is too weak to quantify current speed gains reliably.

Are entry-level software jobs disappearing?

The report cites Indeed data showing that entry-level roles made up 4.5% of US software-development postings in the first quarter of 2026. That is evidence of a narrow entry-level share, but it does not prove AI caused the pattern or that such jobs are disappearing.

Can AI help people learn programming?

It can be used during learning, but the cited Anthropic experiment found that mostly junior developers who used AI to learn a new library scored 50% on a follow-up quiz, versus 67% for those who coded by hand. The study does not establish how all forms of AI-supported learning compare.

Source: ThorstenMeyerAI.com

This content is for general information only and is not financial, tax or legal advice. Consult a qualified professional for decisions about your money.
You May Also Like

4K Gaming Monitors Sound Great—Until Refresh Rate Enters the Chat

Narrowing down the perfect 4K gaming monitor requires understanding why refresh rate matters and how it impacts your gameplay experience.

Build, Rent, Or Quantize: Cutting Your Memory Bill Without Cutting Capability

A new framework shows how AI users can reduce memory costs by building, renting, or quantizing models — with quantization offering the most leverage.

Technology Operations Signal Monitor: PeerTube Is A Free, Decentralized And Federated Video Platform

PeerTube is confirmed as a free, decentralized, and federated video platform, signaling a shift in online video hosting and distribution.

2026-08-31 – Petra Tschudin: Current Challenges Of Monetary Policy

SNB’s Petra Tschudin highlights key issues facing monetary policy today amid rising coverage interest and ongoing economic uncertainties.