{
  "id": 11239523,
  "title": "Amazon releases its own Jev clone as decision models flood the web",
  "url": "https://urgent.news/2026/10/01/amazon-releases-its-own-jev-clone-as-decision-models-flood-the-web",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-01T16:49:22.000Z",
  "source": {
    "name": "TechCrunch",
    "slug": "techcrunch",
    "url": "https://techcrunch.com/2026/10/01/amazon-releases-its-own-jev-clone-as-decision-models-flood-the-web/"
  },
  "original_language": "en",
  "account": "Amazon has launched its own decision-making model, Strands Decider 2B, as a reaction to the growing demand for more specialized AI intelligence. This model was released alongside OpenAI's similar offering, targeting AI developers who prioritize decision-making processes over the capabilities of large language models. Strands Decider 2B is a cost-effective, high-speed tool designed to sort through predefined options and provide an assessment of the confidence in its chosen option. It is fully open-source, allowing for local deployment.\n\nThe model was conceived by Amazon's distinguished engineer Marc Brooker, who was inspired by TypeSafe’s Jev after experimenting with building his own version. This project gained significant traction, briefly topping the Jevbench ranking for its class. Amazon subsequently refined the model and incorporated it into their Strands Labs, which focuses on developing new tools and protocols for AI agents.\n\nBrooker's motivation for creating this model stemmed from discussions with AWS customers, who often required a decision-making tool in their workflows but did not need the capabilities of a fully-featured Large Language Model (LLM). Strands Decider 2B is constructed on the foundation of the Qen3.5-2B LLM, but instead of producing text, it generates calibrated choices. This approach enables customers to make decisions based on a structured workflow, with the added benefits of confidence scores, lower latency, and potentially lower costs.\n\nSimilar decision models have emerged in response to the TypeSafe Jev model, which is named after economist William Stanley Jevons due to its focus on the increasing demand for computer intelligence as costs decrease. The proliferation of such models highlights the high level of interest in this area, but it also raises questions about the practical value they provide. Brooker emphasizes the importance of striking a balance between optimizing decision-making speed and preserving the model's intelligence, particularly when it comes to understanding different languages and retaining a broad knowledge base.\n\nWhile Brooker does not anticipate the frontier AI labs dominating this niche due to smaller markets and associated costs, he does not envision TypeSafe as facing immediate competition from other players. CEO and founder Diogo Almeida agrees, stating that the current models more closely resemble innovative architectural implementations than the work of teams dedicated to creating highly intelligent systems.",
  "summary": "Amazon Web Services' Strand Labs has released the latest Jevalike decision model, Strands Decider 2B.",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}