Can AI Replace The Need To Learn Coding? What The Evidence Says
AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: Can AI Replace The Need To Learn Coding? What The Evidence Says on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get hardware and tech essentials delivered free with Prime

  • Fast, free delivery on millions of items
  • Prime Video, Amazon Music and more included
  • Member-only deals all year
Start your free Prime trial Free trial for eligible customers · Cancel anytime
As an affiliate, we earn on qualifying purchases.

TL;DR

AI now authors much of the code merged at Anthropic, but that company-specific figure does not show that AI can replace programming knowledge. Research on speed, hiring and learning points to a changing role for coding skills, with key questions still unresolved.

AI writes more than 80% of the code merged at Anthropic, according to the company’s May 2026 figures, but available research does not show that developers can safely dispense with coding knowledge. Evidence from workplace data, controlled studies and developer surveys instead points to a shift in the work: people still need to specify requirements, review output, test and debug code, and judge design choices.

Anthropic says Claude authored more than 80% of code merged into its codebase in May 2026, compared with the low single digits before Claude Code launched in February 2025. The company also reported that a typical engineer merged eight times as much code per day in the second quarter of 2026 as in 2024. It cautioned that lines of code measure quantity rather than quality and that the eightfold figure is “almost certainly an overstatement” of productivity gains. Anthropic’s results describe one AI company, not the software industry as a whole.

Controlled evidence on whether AI makes developers faster is mixed. In a randomized METR trial, 16 experienced open-source developers completed 246 tasks in projects they already knew. Using AI tools from February to June 2025 increased completion time by 19% in that trial, even though participants expected a time reduction and later believed they had worked faster. METR said its later data suggest developers may have received more help from AI by early 2026, but described those data as only very weak evidence about the size of any increase.

Other indicators raise questions about who gets to enter the field and how people learn. Indeed data put entry-level roles at 4.5% of US software-development postings in the first quarter of 2026, while senior roles accounted for 69.3%. Stanford researchers found employment among 22- to 25-year-olds in AI-exposed jobs was 19% below where it would otherwise have been, according to their analysis. Neither figure establishes that AI alone caused the change. In an Anthropic learning experiment, mostly junior developers who used AI while learning a new library scored 50% on a follow-up quiz, compared with 67% for participants who coded by hand.

At a glance
reportWhen: Evidence reported through November 2026
The developmentA review of 2026 evidence finds that AI code generation is widespread in some workplaces, while its effects on productivity, entry-level jobs and learning remain mixed.
Crypto market snapshot
Fear & Greed Index
10/100 — Extreme Fear
Bitcoin BTC$63,654▼ 1.6%
Ethereum ETH$1,804▼ 0.3%
Tether USDT$0.9992▲ 0.0%
BNB BNB$589.14▼ 0.5%
USDC USDC$0.9996▲ 0.0%
XRP XRP$0.9812▼ 2.1%
Solana SOL$74.48▼ 0.4%
TRON TRX$0.3131▲ 1.5%
Live data · CoinGecko · alternative.me (24h change)

Skills Still Needed to Review AI Code

The findings matter to students, career switchers, developers and employers because code generation is not the same as software engineering. A tool may produce code quickly, but someone must determine whether it answers the right question, works with an existing system, handles errors and protects users. The evidence supplied here does not establish that every developer needs the same depth of coding skill, or that AI will have the same effect in every job.

For learners, the experiment’s quiz results raise a practical concern: using AI to complete an exercise may not build the understanding needed to find faults later. For employers, a thinner share of entry-level postings could make it harder for new workers to gain experience, although the available figures do not prove that AI caused the change. The reported increase in software-development postings by November 2026 also cautions against interpreting fewer junior listings as evidence that demand has disappeared.

The most defensible takeaway is that coding knowledge remains useful, while the work is changing. Reading unfamiliar code, testing it, debugging failures and judging security are likely to matter alongside prompting and directing AI systems. That is an interpretation of the evidence, not proof of a settled long-term job-market outcome.

Products Worth Considering

Logitech MX Mechanical Wireless Illuminated Keyboard, Tactile Quiet, Graphite

Logitech MX Mechanical Wireless Illuminated Keyboard, Tactile Quiet, Graphite

Apple MacBook Pro 16-inch (M1 Pro, 10‑core CPU, 16‑core GPU, 16GB RAM, 512GB SSD) - Silver

Apple MacBook Pro 16-inch (M1 Pro, 10‑core CPU, 16‑core GPU, 16GB RAM, 512GB SSD) – Silver

As an affiliate, we earn on qualifying purchases.

From Code Generation to Oversight

In March 2025, Anthropic chief executive Dario Amodei predicted that AI would write 90% of code within three to six months and essentially all of it within 12 months. The company’s later internal figure shows substantial adoption at Anthropic, though it does not verify that his broader forecast was met across the industry. In the same remarks, Amodei said programmers would still need to specify what was being built, how it fit with other code and whether its design was secure.

Other public figures are less direct or cover 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 numbers measure different companies and activities, so they should not be treated as a single industry-wide rate.

Productivity studies also test different kinds of work. A 2022 experiment with 95 professional developers found that participants using GitHub Copilot completed a small HTTP-server task 55.8% faster on average. METR’s later trial involved real tasks in large codebases familiar to experienced developers. The different settings help explain why results may diverge; neither experiment alone settles how AI affects everyday software work.

““More than 80% of the code we merge into Anthropic’s codebase was authored by Claude.””

— Anthropic

Limits in the Available Evidence

The evidence does not establish how much AI improves developer productivity across the industry. Anthropic’s figures come from a company that builds and uses its own coding tools, and its output measure counts code rather than assessing its quality. METR’s initial randomized trial involved a small group of experienced developers working in familiar projects; its later study produced results with confidence intervals that included no effect, and METR called the signal unreliable.

It is also unclear how much AI contributes to changes in hiring. The share of entry-level postings and the employment finding for younger workers describe trends, but do not isolate AI as their cause. The Anthropic learning experiment points to a possible cost of relying on AI during practice, but the supplied account does not provide enough detail to determine how broadly the result applies to other learners, tasks or tools. Longer-term effects on training, job numbers and required skills remain uncertain.

What Future Studies Must Measure

The next useful evidence will need to compare AI-assisted work with unaided work across a broader range of developers, tasks and workplaces, while measuring quality, security, maintenance and time—not just lines of code. Researchers will also need to track whether newer tools produce more reliable speed gains than those tested in early 2025 and whether workers’ expectations match measured outcomes.

For the labor market, continued tracking of entry-level and senior postings, hiring and employment can help show whether current patterns persist. For education, studies that test what learners remember and can debug after using AI would clarify when assistance supports learning and when it replaces practice. Until those results are available, the evidence supports neither the claim that coding is obsolete nor a guarantee that current developer roles will remain unchanged.

Key Questions

Does the evidence show that AI can replace learning to code?

No. It shows that AI can generate substantial amounts of code in some settings, while people still need to direct and evaluate the output. The evidence does not establish that coding knowledge is unnecessary.

Does AI make developers faster?

Results vary by study. METR’s early-2025 randomized trial found developers took 19% longer with AI on the tested tasks, while a 2022 Copilot experiment found participants finished a small coding task 55.8% faster. The tasks and study conditions differed.

Are entry-level software jobs disappearing because of AI?

The supplied figures show entry-level roles made up 4.5% of US software-development postings in the first quarter of 2026. They do not establish that AI caused the low share or prove that entry-level jobs are disappearing.

Can using AI make it harder to learn programming?

One Anthropic experiment found mostly junior developers who used AI while learning a new library scored 50% on a follow-up quiz, compared with 67% for those who coded by hand. That result suggests a possible learning trade-off, but its reach beyond that experiment is unclear.

Source: ThorstenMeyerAI.com

Nothing in this article is financial or investment advice. Cryptocurrency and precious-metal investments carry significant risk — do your own research and consider a licensed advisor.
You May Also Like

The Co-Founder’s Black Hole — A Structural Read on Jack Clark’s Automated AI R&D Essay

Jack Clark predicts over 60% chance of fully automated AI research by 2028, raising concerns about institutional readiness and future risks.

The SSD Squeeze: Why Storage Joined the Party

Enterprise and consumer SSDs face record price hikes due to supply shortages driven by AI demand and wafer competition, reshaping the storage market.

Unexpected Decline: Pending Home Sales Fall for First Time in Five Months!

Notable shifts in pending home sales signal a changing market landscape; what could this mean for your next real estate decision?

Upgrade Your AI Storage With These Portable SSDs In 2026

Explore the best portable SSDs for AI storage in 2026, featuring Samsung T9, SanDisk Extreme PRO, and more. Find out which suits your needs.