🔍 Read the full analysis: Why 'System One' AI Could Be A Major Step Forward In Practical Applications on ThorstenMeyerAI.com
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
TypeSafe has announced Jev, a new AI model within its System One framework, designed to produce structured decisions instead of text. This development aims to improve automation speed, accuracy, and cost-efficiency in enterprise applications, representing a significant shift from traditional large language models.
TypeSafe has introduced Jev, a new AI model that produces structured, typed decisions rather than text, marking a significant shift in enterprise AI applications. This development aims to improve decision automation, reduce costs, and increase reliability, making AI more practical for inside-business processes.
Jev is the first public release of TypeSafe’s System One models, designed for decision-making within software systems. Unlike traditional large language models (LLMs) that generate free-form text, Jev provides structured outputs such as categories, confidence scores, and probabilities, enabling software to act directly without parsing text.
The model is built around a new training approach called Reinforcement Learning for Calibrated Decisions (RLCD), which aims to address issues like overconfidence and mode dropping common in LLMs trained with Reinforcement Learning from Human Feedback (RLHF). Jev’s responses are designed to be schema-compliant, with a focus on choices, scores, and yes/no probabilities, making it behave more like a function call than a conversational agent.
TypeSafe claims that Jev can answer questions in 70 to 500 milliseconds at a cost of approximately $0.042 per million tokens, which it asserts is over 190 times faster and over 440 times cheaper than traditional LLM workflows. The company emphasizes that most internal decisions in software are System 1 judgments—quick, intuitive decisions—thus Jev is optimized for automating such judgments efficiently.
Jev vs. LLMs: who should make the call?
Jev, from TypeSafe AI, is a “System One” model. It doesn’t write text. It returns a typed decision with a confidence score that your software can act on directly.
Same support ticket, two kinds of answer
“This ticket appears most likely related to billing, although it could also concern account settings or a recent plan change. I would suggest reviewing the invoice history before…”
A person reads it, or code has to parse the prose.
team: "billing"Software reads it and acts. Nothing to parse.
How they differ
| LLM | Jev | |
|---|---|---|
| Output | Text written for people | A choice, a score or a yes/no probability |
| Speed | Seconds per call | 70–500 ms* |
| Price | Input and (pricier) output tokens | $0.042 per million input tokens, output free* |
| Knows when it’s unsure | Often sounds confident when wrong | Confidence score on every answer |
| Explains its answer | Yes | No, which matters for audits |
| Best at | Reasoning, writing, open questions | Routing, tagging, scoring, duplicate checks |
* Vendor-reported. TypeSafe also claims up to 194× faster and 445× cheaper on its own selected workflows.
Accuracy is something you build
Jev is far cheaper and faster, but not more accurate than frontier models. How you phrase the question matters a lot.
TypeSafe’s benchmark scores agreement with two frontier models rather than verified ground truth. The five-question result used weights fitted on 1,000 labelled examples.
The real idea: a confidence dial you control
“duplicate listing”, confidence 0.62
Raise the threshold for fewer mistakes and more manual review. Lower it for more automation and more risk.
Only use Jev when all four hold
Good fits
- Routing tens of thousands of support tickets a day
- Flagging duplicate listings in a product catalogue
- Replacing a keyword filter that mis-tags half its matches
Poor fits
- Drafting customer emails or release notes
- Reviewing a few high-stakes contracts a month
- Anything that needs a written explanation
Implications for Enterprise Automation
The introduction of Jev signifies a potential paradigm shift in enterprise AI, moving from text-based interactions to decision-oriented automation. Its speed and cost advantages could dramatically expand the scope of tasks that can be fully automated, reducing reliance on human oversight and decreasing operational costs. This approach also addresses reliability concerns associated with traditional LLMs, offering a more predictable and schema-compliant output that is easier to integrate into existing software systems. If widely adopted, Jev could reshape how companies implement AI in workflows, customer support, and decision-making processes, emphasizing structured, calibrated decisions over conversational AI.enterprise decision automation software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Evolution of Enterprise AI and Decision Automation
Over the past three years, the AI landscape has been dominated by large language models like GPT and Claude, which excel at generating human-like text but face limitations in reliability and cost for enterprise use. Many companies rely on these models for tasks like customer support, content creation, and decision support, often paying high costs for reasoning capabilities. However, issues such as overconfidence, hallucinations, and the need for human oversight have persisted. TypeSafe’s System One approach, exemplified by Jev, emerges from a recognition that most internal business decisions are quick judgments that do not require free-form text but benefit from structured, schema-based responses. This shift is partly driven by the realization that traditional RLHF training techniques may not be optimal for automation, prompting the development of alternative methods like RLCD. The launch of Jev with $40 million in funding, led by DCVC and developed by AI veteran Diogo Almeida—who previously co-invented RLHF—underscores industry interest in this new paradigm shift towards decision-focused AI systems.“Traditional large language models are not well-suited for automation because they produce unreliable, overconfident text. Jev’s structured decisions are a more practical approach for enterprise workflows.”
— Diogo Almeida, Co-inventor of RLHF
structured decision-making AI tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Limitations and Validation of Jev’s Performance
While Jev shows promising benchmarks, its accuracy is measured against frontier models like GPT-6 Astra and Claude Fable 5.1, not necessarily correctness. Independent testing indicates lower accuracy on complex tasks, such as a 2,000-email phishing test where Jev scored 62.6% versus 81.3% for Claude. Probabilities can also be overconfident or underconfident depending on question types. The company acknowledges that accuracy depends on task design, and results may vary based on workflow specifics. It remains unclear how Jev will perform across diverse real-world scenarios, especially outside controlled benchmarks.As an affiliate, we earn on qualifying purchases.
Next Steps for Adoption and Validation
TypeSafe plans to expand testing of Jev in real enterprise environments to validate its performance at scale. Industry observers will closely monitor how well Jev integrates into existing workflows, especially regarding accuracy and reliability. The company also intends to refine its training methods and benchmarks to better reflect real-world decision-making challenges. Broader adoption will depend on independent validation of accuracy, robustness, and the ability to handle complex, ambiguous cases. Future updates may include broader API integrations and additional decision types, further expanding Jev’s applicability in enterprise automation.
enterprise automation tools with confidence scoring
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
How does Jev differ from traditional large language models?
Jev produces structured, typed decisions with probabilities instead of free-form text, enabling direct automation without parsing or interpretation. It is designed for decision-making tasks rather than conversational AI.
What are the main advantages of Jev for enterprise use?
Jev offers faster response times (70-500 ms), lower costs (about $0.042 per million tokens), and more reliable, schema-compliant outputs, making it suitable for automating routine decisions within software systems.
What are the current limitations of Jev?
Its accuracy varies depending on task complexity and benchmark conditions. Independent tests show lower performance on some tasks, and probabilities can be overconfident or underconfident. Its performance in diverse real-world scenarios remains to be fully validated.
Will Jev replace traditional LLMs entirely?
Jev is designed for specific decision-making tasks within enterprise workflows. While it may reduce reliance on text-generation models for certain applications, it is unlikely to replace all traditional LLMs, especially those used for creative or conversational purposes.
Source: ThorstenMeyerAI.com
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
![Kimi-K3 Technical Report [Pdf]](https://goodsidekick.com/wp-content/uploads/2026/07/kimi-k3-technical-report-pdf-featured-560x300.jpg)
