Internal Dissent at Amazon over AI Tool Usage
Amazon’s initiative to prioritize its proprietary AI coding assistant, Kiro, over more efficient third-party tools, such as Anthropic’s Claude Code, has triggered a notable backlash within the organization.
Approximately 1,500 engineers have rallied behind Claude Code, expressing concerns regarding diminished productivity and questioning the compatibility of AWS Bedrock sales efforts while being curtailed from utilizing its own offerings.
The contention arises from internal directives implemented late last year, which encourage teams to employ Kiro for production coding while simultaneously imposing restrictions on competing platforms—including Claude Code, OpenAI’s Codex, and Cursor—unless formal approval is obtained.
Engineers Criticize Kiro’s Limitations
The frustration among the workforce is palpable. Internal discussions, which have been scrutinized by Business Insider, reflect sentiments that Claude Code far surpasses Kiro in performance; mandating the usage of a less effective tool is seen as a hindrance to expeditious development.
One engineer articulated, “A tool that can’t keep pace with rivals offers no real innovation. Without competitive strength, Kiro’s only survival mechanism becomes forced adoption rather than genuine value.”
This predicament is particularly acute for engineers engaged in promoting AWS Bedrock, Amazon’s platform that facilitates access to third-party AI offerings, including Claude Code.
Questions have arisen regarding the credibility of sales engineers who must advocate for a product that is not permitted for their own use. As one sales engineer pointedly queried, “Customers will ask why they should trust or use a tool that we did not approve for internal use.”
Amazon’s Response to Employee Concerns
In response to these allegations, Amazon asserts that approximately 70% of its software engineers engaged with Kiro at least once in January.
A spokesperson conveyed to Business Insider that the organization is “seeing incredible improvements in efficiency and delivery from Kiro” and noted that customer growth is “rapidly accelerating.”
While the company denies there is an outright prohibition on Claude Code, it has established “stricter requirements” for its use in production, accompanied by a structured process for soliciting exceptions.
An internal memorandum, first revealed by Reuters in November, formalized this policy. Authored by senior VPs Peter DeSantis and Dave Treadwell, it endorsed Kiro as Amazon’s “recommended AI-native development tool” and clarified that the company does not intend to support additional third-party AI development tools in the future.
The Paradox of Investment in Competitive Technology
This scenario is further complicated by Amazon’s substantial $8 billion investment in Anthropic, the entity behind Claude Code, which has ballooned in value to over $60 billion.
Anthropic also stands as a significant AWS customer, committed to leveraging Amazon’s cloud infrastructure and Trainium chips.
Reports show that some employees’ frustrations heightened after Claude Code seemingly passed security and legal assessments for production use—only for that language to be subsequently retracted.
As one employee succinctly stated, “The trend is clear. More and more people would like to start using Claude Code and be officially supported by Amazon.”
Source link: Timesofindia.indiatimes.com.
