Media focuses on sensational stories of 'token maxing,' but a more systemic threat to the AI boom is the vast majority of expenditure on advanced AI coding tools failing to translate into products that reach users, indicating a massive productivity and ROI gap.

AI coding assistants can make engineers so hyper-productive that they mistakenly believe they can handle the entire product lifecycle alone. This leads them to ignore critical inputs like design, customer briefs, and collaboration, resulting in technically functional but ultimately useless products.

Some engineering teams use AI in a way that produces a high volume of code riddled with mistakes. This forces them to rewrite large portions, sometimes without AI assistance, ultimately slowing them down. The issue is not the tool, but the lack of best practices for its application.

As part of its 'token minimizing' strategy, Meta is encouraging employees to use its in-house tools like MetaCode over more advanced external models. This creates an awkward trade-off: potentially reducing employee productivity to lower the company's massive AI operational expenditure bill.

AI coding tools dramatically accelerate development, but this speed amplifies technical debt creation exponentially. A small team can now generate a massive, fragile codebase with inconsistent patterns and sparse documentation, creating maintenance burdens previously seen only in large, legacy organizations.

When companies see high AI tool usage without a corresponding increase in shipped features, it may not be tech failure. It could be that engineers are successfully automating their existing tasks to maintain previous output levels, effectively gaming productivity metrics.

While AI coding assistants appear to boost output, they introduce a "rework tax." A Stanford study found AI-generated code leads to significant downstream refactoring. A team might ship 40% more code, but if half of that increase is just fixing last week's AI-generated "slop," the real productivity gain is much lower than headlines suggest.

After achieving broad adoption of agentic coding, the new challenge becomes managing the downsides. Increased code generation leads to lower quality, rushed reviews, and a knowledge gap as team members struggle to keep up with the rapidly changing codebase.

AI tools can generate vast amounts of verbose code on command, making metrics like 'lines of code' easily gameable and meaningless for measuring true engineering productivity. This practice introduces complexity and technical debt rather than indicating progress.