Developer Daniel Gilbert published a candid piece on May 9, 2026 detailing his experience with task paralysis and how AI tools like Claude Code both helped him break through creative blocks and hooked him into spending sprees that felt dangerously close to addiction. The article, shared via Hacker News, struck a nerve with developers who recognize the paradox of using AI to solve problems it might be creating elsewhere.

What Task Paralysis Actually Feels Like

Gilbert distinguishes between Analysis Paralysis—where his brain runs in circles—and Task Paralysis, where 'my brain doesn't run at all.' For developers with similar symptoms (he notes undiagnosed ADHD traits and family history), the inability to start even when a plan exists can be career-damaging. Gilbert admits he changes roles every 2-3 years because the same job for three decades feels impossible, yet acknowledges this isn't sustainable. 'I can navigate myself around a lot of technical fields,' he writes, 'but I have no special knowledge.' The irony: he's got ideas and strategy, but execution becomes the wall.

AI as the Kickstart Button

Gilbert dropped nearly €100 on Claude tokens (Max plan) to build both an iOS app and a game. For him, AI solves the motivation gap that derails projects before they begin. 'I have the ideas,' he explains. 'But boy is coding exhausting.' He describes how fighting motivation for every user story was affecting his work—until AI handled the grind while he focused on direction. This isn't unique; developers across Hacker News threads describe similar patterns where LLMs handle boilerplate and implementation details, letting human attention focus on architecture decisions.

The Dopamine Trap Nobody Warns You About

Here's where Gilbert gets uncomfortably honest: the rapid cycle between 'I have an idea' and 'this is the result!' creates intense dopamine hits. Claude Code ships working code fast—faster than any solo developer or small team could manage alone. But token limits exist, and when you hit them? Credit card comes out. Gilbert admits to buying Pro, then adding €20 in API credits, then upgrading to Max for the month while using tricks like /modus opusplan to stretch usage. 'You set yourself in a position where you throw endless money at your source of dopamine,' he writes, 'like a junkie running to their dealer.'

The Artistic Hypocrisy He Acknowledges

Despite embracing AI for coding, Gilbert refuses to use it for artistic work. He buys art or makes it himself 'the old way' because 'the effect AI has on artists is just too destructive.' This tension—raging against AI displacement in one domain while actively using it in another—highlights the messy reality many developers live. We know the technology hurts creators, but when it solves our personal productivity problems? Ethics get blurry fast.

Key Takeaways

  • Task paralysis differs from analysis paralysis: brain won't start versus brain running circles
  • AI tools can legitimately help neurodivergent developers execute on ideas they couldn't previously complete
  • Fast iteration cycles create dependency risks that token limits and pricing models exploit
  • The developer community needs honest conversations about AI addiction, not just productivity wins

The Bottom Line

Gilbert's disclaimer—'No AI was used to create this article'—is the most important line in his piece. It shows self-awareness about what he's participating in while refusing to pretend the system isn't rigged to hook him. This is hacker culture at its most honest: we build tools, and sometimes those tools build dependency in us. The question isn't whether AI helps—it clearly does—but whether we're honest enough with ourselves to spot when 'helpful' becomes 'habit-forming.'