Video Summary

I'm done coding with AI

Brett Codes

Main takeaways
01

After 18 months of heavy AI use, Brett decided to stop using AI for coding and everyday life to protect his craft and wellbeing.

02

AI coding tools sometimes fabricate facts and produce bloated, hard-to-maintain code that reduces developer ownership.

03

Workplace mandates and peer pressure pushed adoption, but speed gains came with trade-offs: apathy, worse software, and loss of learning.

04

Stepping away restored enjoyment of programming and a desire to write intentional, human-centered code.

Key moments
Questions answered

What prompted Brett to stop using AI for coding?

After a year of earnest AI use he experienced apathy, depression, and declining code quality; combined with incidents of AI fabricating facts and troubling real-world advice, he chose to stop.

Which concrete problems with AI tools does he highlight?

He cites hallucinations (made-up dependency versions), bloated and hard-to-maintain code from agentic generation, loss of intimate knowledge of projects, and degraded user experience.

How did workplace dynamics influence his decision?

Management and colleagues evangelized AI and mandated adoption, creating pressure to use tools despite his reservations about their accuracy and long-term effects.

What happened when he stopped using AI?

He regained clarity, started writing code by hand for enjoyment, and reconnected with the craft and purpose of software development.

Is he worried about consequences at work?

Yes—he acknowledges risking job loss or career changes but is willing to accept those outcomes to preserve his principles and wellbeing.

Brett's Experience with AI Coding Tools 00:01

"Over the last 18 months, my career, my craft, my hobby has been drastically changed by the introduction of AI coding tools."

  • Brett, a software developer with over 20 years of experience, shares that his relationship with coding has significantly transformed due to AI tools such as Claude Code.

  • After a year of actively using these AI technologies for coding, he decided to stop using them completely, finding them to have a negative impact on his feelings and the quality of code he produced.

  • He expresses a desire to discuss his mixed experiences with AI, hoping it will resonate with other programmers who may not fully recognize or consider the adverse effects of these tools.

Initial Encounters with AI Tools 02:04

"I first became aware of using AI... GitHub at some point launched Copilot in the Visual Studio Code editor."

  • Brett recounts his first encounter with AI tools when GitHub launched Copilot, which provides auto-completion and code suggestions while coding.

  • He finds such features distracting and claims they interrupt his thought process, ultimately deciding not to use them.

  • A second significant interaction involved a colleague who relied on an AI-powered chatbot for fixing coding issues, leading to confusion as the AI fabricated a non-existent software version.

Shift in Attitude Due to Peer Influence 04:17

"A person was hired as the product lead who was a true believer in AI."

  • Brett comments on the influence of a new product lead at his job who passionately advocated for AI tools.

  • Despite his skepticism, he observed peers using AI successfully and began to feel pressure to give these tools a chance as management pushed for widespread adoption, citing a prediction that AI would completely transform the workplace within six months.

Attempting to Embrace AI Tools 06:38

"I decided, 'Okay, I will try using these in earnest.'"

  • In response to management's mandate, Brett made a conscious choice to experiment with AI coding tools like the Zed editor after a few negative experiences.

  • He was initially impressed by Zed's capabilities and its integration of AI functionalities, enjoying the convenience of quickly generating small projects and software components.

  • As he utilized the AI more frequently, Brett began to feel that it increased his productivity, leading him to overlook the potential downsides.

Realizations and Consequences of AI Interactions 09:16

"I stopped using that... and I immediately felt like I was back, grounded in reality."

  • A pivotal moment occurred when Brett consulted an AI chatbot regarding health symptoms, which prompted him to visit the emergency department based on its suggestion.

  • After feeling embarrassed about relying on AI for medical advice and reflecting on the nature of his interactions, he concluded that the conversational style of AI communications could lead to misunderstanding and misconstruction of reality.

  • This experience, along with hearing about others' concerning interactions with AI, prompted him to discontinue using AI tools and regain a clearer perspective on reality.

Concerns About AI in Coding 10:54

"Using AI for coding has led me to feel existential dread, questioning my purpose after years of enjoying the craft."

  • The speaker expresses a growing discomfort with relying on AI for coding, feeling that it undermines their passion and purpose in software development.

  • Their increasing use of AI has led to doubts about their identity as a programmer, particularly after devoting two decades to honing their skills.

  • There is concern that AI tools, while improving efficiency, create a disconnect from the code itself, diminishing the emotional and intellectual satisfaction derived from writing code.

Impact of AI on Software Development 11:32

"These tools keep getting better, but handing off implementation to AI results in significant downsides for developers."

  • The improving capabilities of AI models and coding tools offer attractive features that many companies invest in, as they streamline development processes.

  • However, the speaker notes a troubling trend: as developers delegate more tasks to these tools, they lose touch with the quality and intricacies of the code being produced.

  • The reliance on AI leads to feelings of apathy towards one’s work, resulting in a disinterest in the software being developed compared to projects that originated from their own coding efforts.

The Harsh Realities of AI Dependence 16:43

"The trade-off for speed and quantity over quality and care is a mistake that we will see repercussions from."

  • The speaker warns that the quality of software suffers when developers lean heavily on AI code generation, leading to bloated features which can make software harder to use.

  • Increased code volume from AI agents makes it challenging for developers to maintain a nuanced understanding of their projects, contributing to a decline in job satisfaction and personal growth.

  • There is a negative impact not only on the software but also on user experience and business operations, as teams may struggle with maintaining complicated systems resulting from AI-generated code.

Personal Reflection and Resistance 19:16

"I decided I can't do it anymore. I started writing code by hand again for fun, which reminded me why I love coding."

  • After reflecting on the taxing implications of AI on both their mental well-being and professional identity, the speaker opts to cease reliance on AI for coding.

  • This decision leads to a renewed enjoyment in programming, allowing them to reconnect with the intrinsic satisfaction of solving problems and creating software manually.

  • The speaker illustrates how stepping back from AI has cultivated a sense of purpose and joy in their work, which had been lost during the past year and a half of heavy AI dependency.

Enjoying the Craft of Programming 21:08

"Enjoy the craft, be close to the code, be close to the software, make things by hand for humans."

  • The speaker emphasizes the importance of enjoying the actual process of coding and creating software, rather than solely focusing on optimizing for automated systems.

  • There is a clear call to embrace the craft of programming, suggesting that a more human-centered approach to coding is crucial for the fulfillment of both the creator and the end user.

Standing by Personal Beliefs 21:20

"Whatever comes of this decision, like if I get fired, if I can't find a job, if my career as I know it is over, so be it."

  • The speaker expresses a strong commitment to their beliefs regarding coding practices, even if it means facing significant career risks.

  • This viewpoint indicates a willingness to prioritize personal integrity and passion over professional security, suggesting that one must often stand firm in their values amidst industry changes.

Exploring Alternatives in the Field 21:51

"I hope that there's a way to exist in this field without that."

  • The speaker remains hopeful about finding a way to continue working in programming without relying on AI technologies.

  • This optimism reflects a desire to maintain a personal connection to coding while navigating an evolving tech landscape, hinting at the potential for alternative methods and practices within the programming community.