AIToday

Software developer argues LLMs cannot replace programmers due to fundamental architectural limitations

Hacker NewsMay 12, 20261 min read

Summaries like this, in your inbox every morning.

Sign up free →

3 Key Points

  1. A software developer with decades of experience contends that LLMs, built on transformer architecture, are 'too stupid' and structurally incapable of performing serious programming tasks, citing repeated failures when working with unfamiliar concepts like SBP (a framework used only by their team at okTurtles).

  2. The developer observes that LLMs lack sufficient training data to understand new or niche concepts, and therefore cannot learn to work with technologies that were not widely represented in their training set—a limitation the developer believes will persist as new frameworks and patterns emerge.

  3. Despite this technical assessment, the developer notes that non-technical audiences remain enthusiastic about AI-generated code and pricing-focused marketing, unaware of the structural gaps that become apparent during real-world development work.

Discussion

No discussion yet for this article

Stay ahead with AI news

Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.

Get Started Free

Free · takes 30 seconds · unsubscribe anytime

1 minute a day. The AI essentials.

200+ sources · Email / LINE / Slack

Get it free →