Forgot your password?
typodupeerror
Programming

Amazon Engineers Want Claude Code, but the Company Keeps Pushing Its Own Tool (businessinsider.com) 40

Amazon engineers have been pushing back against internal policies that steer them toward Kiro, the company's in-house AI coding assistant, and away from Anthropic's Claude Code for production work, according to a Business Insider report based on internal messages. About 1,500 employees endorsed the formal adoption of Claude Code in one internal forum thread, and some pointed out the awkwardness of being asked to sell the tool through AWS's Bedrock platform while not being permitted to use it themselves.

Kiro runs on Anthropic's Claude models but uses Amazon's own tooling, and the company says roughly 70% of its software engineers used it at least once in January. Amazon says there is no explicit ban on Claude Code but applies stricter requirements for production use.
This discussion has been archived. No new comments can be posted.

Amazon Engineers Want Claude Code, but the Company Keeps Pushing Its Own Tool

Comments Filter:
  • Probably get better results than when they just assign someone to you.

    • I swear there used to be these places, you go there and learn how to code, you pay a bunch of money and get a piece of paper, and in the end, a coder with that piece of paper would get paid more than someone without the piece of paper... can't remember what those places were called.

      • by unrtst ( 777550 )

        can't remember what those places were called.

        The Cisco Network Academy?

        Joking aside, getting a comp-sci degree isn't the same as proof you know a certain skill. It's a foot in the door, and it's evidence of a solid basis should they prove via the interview to have the necessary domain knowledge (so maybe more starting salary), but one should still have the opportunity to interview anyone that will be their direct report (just my 2 cents... I've been assigned employees before and have not had the best experience with it).

        • by Anonymous Coward
          Sure. But in theory anyone with a comp-sci degree, even from thirty years ago, will interview better than someone who breaks down and cries out "Where is my Claude? I can't program without my Claude!"
      • by jhoegl ( 638955 )
        Last I heard ITTech was shut down.
  • by Anonymous Coward on Thursday February 12, 2026 @01:22PM (#65985070)

    I'm a hardware guy with limited programming experience. I can read and understand C for the most part but when it comes to coding, I am not a neckbeard. Claude Code has enabled me to finish a veritable shit-ton of projects that I have designed and built hardware for but been unable to get the firmware/software side working. And, that's just at home.

    At work, I'm an engineering manager and I've been able streamline our design and coding processes immensely by training my engineers on using Claude code to do some heavy lifting for them. The result is that my green engineers fresh out of school are productive much faster, and more importantly, I can reduce org costs by having a lower headcount. By the end of this year I expect I'll be able to reduce headcount by 25% in the engineering department, mostly in the senior staff as they are a) resistant to adoption of AI and b) expensive as hell for people who don't want to grow in their roles and use new tools.

    AI is the way of the future, whether we like it or not. The best course is to adopt it and let it make you better at your job, as it has for me. Those who refuse are going to whine and bitch and moan themselves into irrelevance by the end of this year.

    • AI works well today. Is it perfect? No, but I haven't seen any 1.0 code that hasn't had bug fix releases when coded "the old way" without AI.

    • by Afell001 ( 961697 ) on Thursday February 12, 2026 @01:54PM (#65985156)
      As you journey down this road of agentic adoption of AI is the degree of infiltration into a company's core, and then the "lift" of proprietary code that runs everything the company does. I foresee one of the greatest thefts in history as the companies involved in programming and training these models, and then hosting the data centers these agents "call home" literally walk away with all the distinct intellectual property that companies have built over time. This will truly be the end of proprietary code altogether. How long until all these $$trillion companies are now competing directly with the smaller companies that took advantage of AI adoption and transferred all their distinctiveness to the AI agents they trained, for them to hand it over every time they "called home"? It will be much akin to the US companies that were forced to cooperate with a Chinese partner company to enter the Chinese market, only to discover that the Chinese partner has taken the entirety of the Chinese market with free access to the distinctiveness of the US company, and no recourse for the US company because of Chinese sovereign rights.
    • by Anonymous Coward
      Be very careful with AI generated C code for hardware interfaces. C is well known for code that compiles but does very bad things, usually unintentionally.
    • At work, I'm an engineering manager and I've been able streamline our design and coding processes immensely by training my engineers on using Claude code to do some heavy lifting for them... By the end of this year I expect I'll be able to reduce headcount by 25% in the engineering department, mostly in the senior staff ...

      And you too are "senior staff", right? How long will it be before AI replaces you?

      AI is the way of the future, whether we like it or not. The best course is to adopt it and let it make you better at your job, as it has for me.

      I'm going to ignore nuance for a moment and just mention this: dystopian trends grow stronger faster with the help of collaborators. Do you really want to be so actively supporting AI, given its likely negative social, economic, psychological, and environmental impacts?

      I get that your personal situation may give you little choice in embracing AI at work. But must you be a cheerleader for it?

      • by Hadlock ( 143607 )

        > And you too are "senior staff", right? How long will it be before AI replaces you?
         
        I suspect you still need a human in the loop for 30-60% of things. It's gonna be a few years before companies are ready/confident enough to staff down that low. It doesn't matter if engineers cheerleader AI, management is busy building their own tools already, the toothpaste is well out of the tube at this point.

        • Most of the work is decision making and analysis. Push a deploy. Did it hose anything? Do you need to fix something? Or legal wants some new thing in some form but it requires coordination between different API providers and have to be rolled out piece meal. There's a lot of nuance and many moving pieces where humans have to at least supervise.

    • So, to summarize: your "green engineers" are getting better guidance from Claude Code than from you, who doesn't know how to code? Seems reasonable.

    • My seniors are loving AI and we're building agents to help each other. Our cadence has increased but we're likely gonna keep head count as is because the projects we're working on are still long term logistical scaling problems that involve lots of A/B testing and slow methodical roll outs.

      As just one example I had to build a project for a newer linux distro but I was getting a lot of errors. Before ML help I'd be googling for the problem and eventually figure it out. Sometimes it would take 1-2 weeks or mo

      • As just one example I had to build a project for a newer linux distro but I was getting a lot of errors. Before ML help I'd be googling for the problem and eventually figure it out. Sometimes it would take 1-2 weeks or more even.

        That's not a very specific example.

  • Jeez - I had just assumed that Bezos has a limo and a driver.

    But I wouldn't put it past Jeffy to have a stable of dray humans. After all, feudal lords gotta feudalize...

  • Does "Claude" code in PHP [reddit.com]?

  • We have to use Copilot at my company "because we paid a bunch of money to license it."
    I put 2/12/26 in a document and the summary said it's about an event that happens in 1926.
  • by ZipNada ( 10152669 ) on Thursday February 12, 2026 @03:05PM (#65985352)

    "Kiro runs on Anthropic's Claude models but uses Amazon's own tooling"

    The models require a front end that feeds them user queries, gives r/w access to local files, etc. and most people will want it to work from within the IDE they normally use. For millions of developers that IDE is Visual Studio Code. There are several companies out there that have forked VS Code to make a custom version that integrates with multiple LLM models. Amazon Kiro is one of them, so for many people this will integrate right in with their work flow; https://kiro.dev/ [kiro.dev]

    You can choose from multiple Anthropic models with Kiro but not other brands such as Gemini, GPT, which means less flexibility for the developer. I will say that in my opinion the Claude models are the top performers.
    https://kiro.dev/docs/cli/chat... [kiro.dev]

    Here is someone's comparison of Claude Code and Windsurf; https://aiforcode.io/tools/cla... [aiforcode.io]

    So I'm not sure what the big fuss is about. Claude Code is "Terminal-native CLI via npm. Native VS Code and JetBrains IDE extensions." Possibly it gives you a larger context window than some other options? Other than that it appears to be much the same.

    • by CAIMLAS ( 41445 )

      The power of claude code isn't in the tool, as much as it is the ecosystem of plugins and the ability to customize it for your specific workflow, and it's AI-native.

      It isn't some sloppy bolt on that thinks people will be writing the actual code. (They won't.)

  • The inevitable big company foot dragging is going to produce grim results, not in decades or years, but in quarters to months. The difference between "ChatGPT is my new Google" and people who actually figure out what to do with AI isn't a few percentage points, I'm roughly 10x more productive with Claude Code, and that's after shedding the poor tooling choice of putting Antigravity in the middle. Little things make ENORMOUS differences in 2026.
    Eating one's own dogfood is the right choice, but it's gotta k
  • by Tony Isaac ( 1301187 ) on Thursday February 12, 2026 @07:56PM (#65985862) Homepage

    The more I hear about the stunts Amazon pulls, in the way they treat their employees, the more I'm determined never to work for them.

  • I worked at Amazon Games, back when we were told we had to dogfood the nascent Lumberyard engine instead of using established industry-standard tech. If you can name any of the games that Amazon released in those 5 years you can maybe guess how that went.

186,000 Miles per Second. It's not just a good idea. IT'S THE LAW.

Working...