Amazon forces its engineers to use AI and measures who uses it the most. The employee response: anti-AI memes
AI gurus keep selling us that this technology is going to change the world and? the AGI is about to fall. From the outside, everything is promises of automation and productivity, but from the inside the feeling is very different. Yesterday we learned that the engineers who are working on Google AI They don’t stop making fun of her on internal channels. Well, exactly the same thing is happening on Amazon. Sloppenheimer. It’s one of the memes that Amazon employees have shared in the Slack channel called #actual-aws-memes. In it they mock tools like Kiro (Amazon Web Services’ code platform), Claude Code, and the AI agent Meshclaw. They tell it in 404mediawhere they have spoken with several company employees who wanted to remain anonymous. These employees admit that anti-AI memes started circulating quite some time ago (in late 2024), which is when the company started putting more pressure on them to adopt AI tools. Of course, although mockery of AI is very common, they assure that there is a variety of opinions on the matter. Kiro AI. It is the tool that receives the most ridicule. There are several memes that suggest Kiro is pretty mediocre, such as the one that includes the text “Kiro: I confirm I have the full picture” over an image of an iceberg with much of it underwater. Discontent with this tool reached such a point that Amazon closed Kirorank, an internal leaderboard that measured and rewarded employees’ use of Kiro. Amazon’s official version is that employees had already integrated AI into their daily lives and ranking was no longer necessary, but what really happened is that engineers started cheating. They automated absurd and totally useless tasks for the sake of climbing the rankings, adding to the company’s AI bill. Amazon is not the only one. As we said at the beginning, this anti-AI rebellion based on memes is not exclusive to Amazon, exactly the same thing is happening at Google. While CEO Sundar Pichai boasts that 75% of Google’s code is written with AIemployees use an internal channel to make fun of that very thing. For example, while Google announced news in the Google I/O 2026a meme appeared on the channel saying that they were announcing “new ways to slop.” Memes can be voted on with a reaction system and this immediately added 100 thumbs up. They also make fun of the AI bros who do not stop evangelizing about the benefits of AI and point out the enormous work it is to review the code made by AI, often riddled with errors. What companies say. There is an evident disconnection between the official discourse and the internal feeling of the employees. Amazon has responded to the leaks by trying to downplay them. In an email to 404media, Amazon assures that the negative comments come from a few individuals and do not represent the majority. For its part, Google sent a statement to the same media in which they said they encouraged their employees to test and criticize their internal tools, even through memes. And a curious thing, Google sent two almost identical statements, the only difference is that in the first they mentioned that it was “essential that we keep humans in the process, including supervision.” In the second that phrase had disappeared. Image | Xataka with Magnific In Xataka | From “tokenmaxxing” we have moved on to “tokenwasting”: the level of waste in AI is reaching unprecedented levels
