![\"Banner](\"https://www.artificialintelligence-news.com/wp-content/uploads/2025/08/ai-expo-banner-2025.png\")
TypeSafe, founded by a ChatGPT co-inventor, has left stealth and is launching its Jev model to automate programmatic decisions with parallel sampling architecture. Software systems requiring automated deterministic logic can now bypass conversational language models following TypeSafe AI’s release of Jev, a specialised System One Model designed to execute structured probabilistic decisions directly inside production […]
\nThe post ChatGPT pioneer launches Jev model for programmatic logic appeared first on AI News.
", "content_html": "TypeSafe, founded by a ChatGPT co-inventor, has left stealth and is launching its Jev model to automate programmatic decisions with parallel sampling architecture.
\n\n\n\nSoftware systems requiring automated deterministic logic can now bypass conversational language models following TypeSafe AI’s release of Jev, a specialised System One Model designed to execute structured probabilistic decisions directly inside production codebases.
\n\n\n\nDeveloped during two years in stealth by OpenAI veteran and ChatGPT co-inventor Diogo Almeida, the model abandons text and string generation entirely.
\n\n\n\nInstead of autoregressively generating sequential text tokens, the system takes an unstructured state as an input and outputs type-safe structured values in a single parallel query. The architecture targets deterministic code integration, preventing syntactic type failures and output hallucinations by design.
\n\n\n\nDiogo Almeida, Founder at TypeSafe, said: “Think of Jev as a frontier-intelligence function call: unstructured state in, typed probabilistic decisions out.”
\n\n\n\nEngineers built the platform around an alternative training methodology termed Reinforcement Learning for Calibrated Decisions (RLCD).
\n\n\n\nConventional frontier models train via Reinforcement Learning with Human Feedback (RLHF) or Reinforcement Learning with Verifiable Rewards (RLVR) to satisfy human conversational preferences or isolated programmatic checks. TypeSafe trains Jev to produce calibrated probabilities on execution logic, ensuring returned confidence scores correlate directly with output accuracy.
\n\n\n\nDeparting from autoregressive token generation, Jev relies on a hardware-aware parallel sampler that evaluates and delivers all structured values simultaneously.
\n\n\n\nThe structural constraint restricts outputs to predefined schemas, eliminating the parsing pipelines and guardrail wrappers required when deploying traditional large language models across business automation flows. The internal sampler handles high-cardinality selections up to 255 discrete options, using a two-stage process that scores options independently prior to selection.
\n\n\n\nInternal evaluations conducted from the team’s West Coast servers record end-to-end response latencies between 70–500 milliseconds, contrasted against the 3–329 seconds observed across conversational frontier deployments.
\n\n\n\nPublished workflow evaluations run across fixed compute graphs (measuring multi-step decision branching against consensus baselines from GPT-6 Astra and Fable 5.1) recorded execution speeds up to 193.6 times faster.
\n\n\n\nInput processing costs are priced at $0.042 per million tokens, down from standard conversational rates spanning $0.20 to $10 per million tokens. Because parallel sampling generates structured states without autoregressive token generation passes, TypeSafe provides output tokens without metering charges.
\n\n\n\nProduction demonstrations show Jev resolving branching rules across high-speed game states and web traversal trees. In dynamic stress testing, engineers operated a real-time reactive bot across Doom game state structures at 10 queries per second, generating operating costs of approximately $7 per hour.
\n\n\n\nA secondary Wikiracing test evaluated link selection across dense encyclopaedia directories, finishing traversals in fewer steps than external non-reasoning models by evaluating options without generating hallucinated dead ends.
\n\n\n\nField testing confirms primary target applications across real-time feature extraction, petabyte-scale data workflows, output verification layers, and automated branching logic where hand-written rules prove brittle.
\n\n\n\nTypeSafe opened early developer access today and started onboarding engineering teams from its deployment waitlist.
\n\n\n\nSee also: Microsoft AI opens review on Humanist AI Code of Conduct
\n\n\n\nWant to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information.
\n\n\n\nAI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.
\nThe post ChatGPT pioneer launches Jev model for programmatic logic appeared first on AI News.
", "content_text": "TypeSafe\n, founded by a ChatGPT co-inventor, has left stealth and is launching its Jev model to automate programmatic decisions with parallel sampling architecture.\nSoftware systems requiring automated deterministic logic can now bypass conversational language models following TypeSafe AI’s release of Jev, a specialised System One Model designed to execute structured probabilistic decisions directly inside production codebases.\nDeveloped during two years in stealth by OpenAI veteran and ChatGPT co-inventor Diogo Almeida, the model abandons text and string generation entirely.\nInstead of autoregressively generating sequential text tokens, the system takes an unstructured state as an input and outputs type-safe structured values in a single parallel query. The architecture targets deterministic code integration, preventing syntactic type failures and output hallucinations by design.\nDiogo Almeida, Founder at TypeSafe, said: “Think of Jev as a frontier-intelligence function call: unstructured state in, typed probabilistic decisions out.”\nJev model architecture: How hardware-aware parallel sampling works\nEngineers built the platform around an alternative training methodology termed Reinforcement Learning for Calibrated Decisions (RLCD).\nConventional frontier models train via Reinforcement Learning with Human Feedback (RLHF) or Reinforcement Learning with Verifiable Rewards (RLVR) to satisfy human conversational preferences or isolated programmatic checks. TypeSafe trains Jev to produce calibrated probabilities on execution logic, ensuring returned confidence scores correlate directly with output accuracy.\nDeparting from autoregressive token generation, Jev relies on a hardware-aware parallel sampler that evaluates and delivers all structured values simultaneously.\nThe structural constraint restricts outputs to predefined schemas, eliminating the parsing pipelines and guardrail wrappers required when deploying traditional large language models across business automation flows. The internal sampler handles high-cardinality selections up to 255 discrete options, using a two-stage process that scores options independently prior to selection.\nInternal evaluations conducted from the team’s West Coast servers record end-to-end response latencies between 70–500 milliseconds, contrasted against the 3–329 seconds observed across conversational frontier deployments.\nPublished workflow evaluations run across fixed compute graphs (measuring multi-step decision branching against consensus baselines from GPT-6 Astra and Fable 5.1) recorded execution speeds up to 193.6 times faster.\nInput processing costs are priced at $0.042 per million tokens, down from standard conversational rates spanning $0.20 to $10 per million tokens. Because parallel sampling generates structured states without autoregressive token generation passes, TypeSafe provides output tokens without metering charges.\nFrom Doom bots to petabyte data: Real-world Jev deployment testing\nProduction demonstrations show Jev resolving branching rules across high-speed game states and web traversal trees. In dynamic stress testing, engineers operated a real-time reactive bot across Doom game state structures at 10 queries per second, generating operating costs of approximately $7 per hour.\nA secondary Wikiracing test evaluated link selection across dense encyclopaedia directories, finishing traversals in fewer steps than external non-reasoning models by evaluating options without generating hallucinated dead ends.\nField testing confirms primary target applications across real-time feature extraction, petabyte-scale data workflows, output verification layers, and automated branching logic where hand-written rules prove brittle.\nTypeSafe opened early developer access today and started onboarding engineering teams from its deployment waitlist.\nSee also:\nMicrosoft AI opens review on Humanist AI Code of Conduct\nWant to learn more about AI and big data from industry leaders?\nCheck out\nAI & Big Data Expo\ntaking place in Amsterdam, California, and London. The comprehensive event is part of\nTechEx\nand is co-located with other leading technology events including the\nCyber Security & Cloud Expo\n. Click\nhere\nfor more information.\nAI News is powered by\nTechForge Media\n. Explore other upcoming enterprise technology events and webinars\nhere\n.\nThe post\nChatGPT pioneer launches Jev model for programmatic logic\nappeared first on\nAI News\n.", "image": { "url": "https://virtues.guru/tech-assets/80_VirtualBox_Windows_Server_2008_SP2_11_02_2025_12_39_40_crop.png", "file": "80_VirtualBox_Windows_Server_2008_SP2_11_02_2025_12_39_40_crop.png", "title": "File:VirtualBox Windows Server 2008 SP2 11 02 2025 12 39 40 crop.png" }, "language": "en", "translations": { "pt": { "title": "Pioneiro do ChatGPT lança modelo Jev para lógica programática", "content_text": "TypeSafe, fundada por um dos inventores do ChatGPT, lançou seu modelo Jev para automatizar decisões programáticas com uma arquitetura de amostragem paralela. O Jev contorna os modelos de linguagem conversacionais e executa decisões probabilísticas estruturadas diretamente dentro dos repositórios de código de produção. O modelo é até 193,6 vezes mais rápido do que os modelos de linguagem grandes tradicionais e tem custos de processamento de entrada de US$ 0,042 por milhão de tokens.", "summary_en": "TypeSafe, founded by a ChatGPT co-inventor, has launched its Jev model to automate programmatic decisions with parallel sampling architecture. Jev bypasses conversational language models and executes structured probabilistic decisions directly inside production codebases. The model is up to 193.6 times faster than traditional large language models and has input processing costs of $0.042 per million tokens.", "translated_at": "2026-09-17T06:21:54.077494+00:00", "model": "translategemma:4b", "summary_model": "llama3.1:8b", "source": "json_summary" } } }