A recent bug in AWS’s new AI-powered IDE, Kiro, has caused problems for developers trying to get their work done. The issue is linked to the way tasks are consuming more requests than they should, which quickly eats up usage limits and frustrates users. This all came to light just days after AWS changed how it charges for Kiro, making many developers upset.

What’s Causing the Issues with Kiro

Last week, AWS admitted there was a bug in Kiro’s system. The problem is that some tasks are counting as multiple requests instead of just one. This means developers are reaching their limits faster than expected. Adnan Ijaz, AWS’s director of product management for Agentic AI, said they are working to fix the bug and hope to resolve it within a few days.

This bug has made it difficult for developers to use Kiro smoothly. Many have reported on social media and forums that their usage limits are being exhausted way too quickly. Some users even say that they’re consuming large amounts of credits without actively engaging with the IDE. AWS has promised to reset limits for anyone affected by the bug while they work on a fix.

How the Pricing Changes Sparked Discontent

The trouble started right after AWS announced its new pricing structure for Kiro. When the IDE was first introduced in public preview, AWS limited its use to prevent overload. They had also scrapped earlier plans for simple tiers and moved to a more complex system based on vibe and spec requests. The free tier caps at 50 vibe and 0 spec requests, while paid plans range from $20 to $200, with different limits on vibe and spec requests.

Many developers felt the new pricing was confusing and unfair. Social media posts and forum comments showed a lot of dissatisfaction. Some users complained that they were hitting their limits fast, making Kiro less useful. Several on GitHub called out the bug, saying it was causing them to burn through credits even when they weren’t actively using the IDE.

This situation echoes a similar episode with Cursor, another AI IDE, which also faced criticism after its own pricing changes in June. Back then, users were confused and angry about unexpected charges. Cursor later clarified its plans and offered refunds for surprise costs, but some developers remain unsure about what to expect from such pricing models.

Why Kiro Is Still Seen as Valuable Despite the Issues

Despite the problems and frustrations, some analysts believe Kiro offers unique advantages. A key point is Kiro’s approach to development, which focuses on defining entire applications or tasks through detailed specifications. This method is different from more traditional chat-based AI coding tools.

According to Jason Andersen, a principal analyst at Moor Insights and Strategy, this spec-driven approach means each request can be more complex and resource-intensive. It’s like bundling many smaller tasks into one big request, which uses more GPU power and costs more. That’s partly why Kiro’s pricing is higher compared to rivals.

Andersen notes that AWS’s pricing model was designed to reflect this increased resource use. The goal was to create a tool that’s powerful and flexible, even if it costs more. AWS also hopes that educating developers about how to use Kiro efficiently will help them get the most value. Still, many more insights and benchmarks about what users can expect are needed to better understand the costs involved.

In the end, while Kiro’s recent issues have caused some setbacks, its innovative approach could give it an edge in the long run if AWS can iron out the bugs and clarify its pricing model. Developers will likely keep an eye on how AWS manages these challenges in the coming weeks.

Inspired by

• https://www.infoworld.com/article/4042912/aws-blames-bug-for-kiro-pricing-glitch-that-drained-developer-limits.html

Sources

• gmpg.org
• github.com
• github.com
• github.com
• cursor.com