Developers at large technology firms have grown disillusioned with using artificial intelligence for programming and complain of skill loss, according to 404media.
Coders say AI-generated code often contains errors. Reviewing and fixing it can take longer than writing from scratch.
“We are being forced to use AI agents to make sweeping changes across the entire codebase. Assessing the quality and safety of that volume is simply impossible, especially given that hundreds of programmers are doing the same thing,” said a UX designer at a technology company.
The expert added that the team is accruing “a pile of technical debt” that will be impossible to untangle once models become prohibitively expensive.
Tech executives are eager to tout the share of code generated by AI:
- In April, Google said 75%;
- in 2025, Microsoft CEO Satya Nadella put it at 30%;
- Anthropic — 90%;
- Meta boss Mark Zuckerberg predicted that within 12–18 months neural networks will write most of the code that improves AI itself;
Meanwhile, industry-wide layoffs continue — companies cite automation and cost-cutting.
The boost never came
404media writes that “the huge productivity leap delivered by artificial intelligence” has not led to more or better product.
Developers deny that AI helps them at work — yet they are required to integrate it into their workflows.
“Using an LLM in one form or another is a mandatory requirement. Its use is part of performance evaluation criteria. We are literally being flooded with AI tools, and the answer to any problem is to ‘try artificial intelligence first’,” said a programmer at one of the FAANG companies.
Because performance reviews are tied to adopting the technology, most developers use AI “to tick a box”.
An engineer at a fintech company said using LLMs is not mandatory but encouraged: developers are given access to Cursor.
A programmer at a small web-design firm stressed that AI assistants in the IDE have not boosted productivity — the code contains errors and every line has to be rechecked.
“Another developer works with me on contract. He generates huge volumes of code, leaving me with more than 1,000 lines of pull requests to review, and that takes a tremendous amount of time. As a result, I feel more tired and burned out than at any time in my life,” the engineer said.
A coder in fintech added that AI can generate more code than the team can review or explain.
“As a result, you either throw it away or ship it, fearing there may be elements of very low quality,” he explained.
AI is useful, sometimes
Developers concede that AI does handle some tasks — for instance, quickly assembling prototypes and implementing solutions in unfamiliar domains.
One engineer said LLMs are handy with large volumes of information: they find where a request is handled on the server, summarise data from logs and help search documentation on code changes.
Skills are deteriorating
As AI embeds deeper into workflows, developers are losing skills honed over years. Researchers call this phenomenon “cognitive debt” or “cognitive atrophy”.
A programmer at a small web-design company said he once could not remember how to implement an API in Laravel — and it “scared me to death”.
“It is like when mobile phones arrived and we stopped memorising numbers. For me it has grown into outsourcing the thinking process. My critical thinking and ability to sit down and think through a problem or a project have deteriorated,” commented a software developer in finance.
AI is here to stay
Most engineers agree that large language models will remain and continue to play a role in programming. The question is how the industry will cope with management’s current obsession with the technology, particularly when training new generations of developers.
“We are hiring junior programmers who rely on AI to perform the simplest tasks. They do not have the knowledge or experience to recognise when the outputs of neural networks contain errors or are inefficient,” said the UX designer.
In August 2025, Coinbase fired programmers who refused to use artificial intelligence in their work.