Kimi K3 rocks the AI industry as Moonshot AI undercuts closed-source American competitors on price
Jev's output (and optionally input) is in a predefined data format that's essentially plain JSON. Unlike interacting with LLMs, there are no extraneous words, long-winded thinking, or necessity to ask the model for brevity. Likewise, there's no need for global contextual prompts or saved memories, often necessary to try and coax LLMs to behave as if they were minimally deterministic. Beyond providing the state and questions, operations like input parsing, date handling, or database reading remain in your own code and are of no concern to Jev.
At first sight, this might look like a more straightforward interface to an LLM, but it's fundamentally different. Jev does not need or even want the entire context leading up to the question — providing extraneous information actually lowers the accuracy, and the context window is capped at a meager 64,000 tokens. Since the bot always produces a confidence percentage, it doesn't hallucinate in the familiar chatbot sense of creating statements and data out of nowhere.
The developer, rather than Jev, is responsible for making decisions based on the answers' confidence factors. Jev can still misclassify information, fall victim to adversarial attacks, or answer literal wording rather than meaning.
It's expected that the most common use case for Jev (and presumably future System One models) will be to wrap logic workflows around posted questions, as always having a confidence factor available makes it easy to integrate it into the decision steps of something like "if we're fairly certain the user requested a refund, and they prefer store credit, and we can see that they buy more PC gear around September, also offer them a 20% deal on an RTX 5090." The documentation has other usage examples like intent routing or citation checking .
Jev's strength isn't in acting like an agent or reasoning through a broad problem — TypeSafe makes it clear that open-ended tasks are better suited to an LLM, perhaps even integrated into code that also involves Jev. An example would be a monitoring system where Jev can use its provided info to assess if there's a serious system issue, and if so, bring in an LLM to examine logs, look for a cause, and produce a report. Likewise, Jev isn't trained on customer data and doesn't make inferences from anywhere other than the provided state (and its own training).
I'm by no means an AI engineer, but my developer layman's opinion is that if Jev works reasonably as promised, it might fix one of the major roadblocks to deeply integrating AI in software: dealing with chatbot LLMs, which, for practical purposes, are annoyingly amorphous blobs that may or may not behave and produce the desired output — never mind correct output — and slowly and expensively at that. Having a simple assessment/response/confidence interface that one can easily and cleanly integrate into code without requiring specific training or carefully crafted textual incantations feels far more natural and easy to use.
Bruno Ferreira Social Links Navigation Contributor Bruno Ferreira is a contributing writer for Tom's Hardware. He has decades of experience with PC hardware and assorted sundries, alongside a career as a developer. He's obsessed with detail and has a tendency to ramble on the topics he loves. When not doing that, he's usually playing games, or at live music shows and festivals.
Key considerations
- Investor positioning can change fast
- Volatility remains possible near catalysts
- Macro rates and liquidity can dominate flows
Reference reading
- https://www.tomshardware.com/tech-industry/artificial-intelligence/SPONSORED_LINK_URL
- https://www.tomshardware.com/tech-industry/artificial-intelligence/typesafe-ais-jev-offers-an-alternative-to-llms-that-claims-to-be-193x-faster-and-445x-cheaper-system-one-type-model-is-bespoke-for-probabilistic-decision-making#main
- https://www.tomshardware.com/my-account
- Apple Mac mini (Late 2026) Review: Strong performance, but losing its grip on value
- Intel suspends bug bounty program that paid up to $100,000 per flaw — new Intigriti disclosure program offers no rewards
- The PC gaming ray tracing obsession began with the first RTX 20 graphics cards released on this day in 2018 — the GeForce RTX 2080 and 2080 Ti led the charge, b
- Microsoft patents system to freeze games and inject ads during downtimes — watching commercials earns ad-free playtime credits
- Techie makes his own motherboard to refurbish 15-year-old HP ProLiant N40L tower server — Raspberry Pi CM5 adds NVMe and USB 3.0
Informational only. No financial advice. Do your own research.