←Search results

Search “OpenAI”

By relevance · Total 3764 results Clear search

Wednesday

  2. Ma Dongxi NLP

    Hamel Husain publishes tutorial: How to use decision models for Evals

    Hamel Husain published a new blog post explaining how to use decision models for Evals, including choosing a decision model for a task, tuning its threshold (don't assume the 0.5 default is appropriate), and interpreting the scores the model produces; the post includes several interactive charts. When reposting it, Ma Dongxi NLP noted that recently emerged decision models include TypeSafe's Jev, Perplexity's Decider, Cloudflare's Clef, and OpenAI's Decisions API; they output probabilities rather than long text and are often used as alternatives to LLM-as-a-Judge.

    models for Evals, including choosing a decision model for a task, tuning its threshold (don't assume the 0.5 default is appropriate), and interpreting the scores the model produces; the post includes several interactive charts. When reposting it, Ma Dongxi NLP noted that recently emerged decision models include TypeSafe's Jev, Perplexity's Decider, Cloudflare's Clef, and OpenAI's Decisions API; they output probabilities rather than long text and are often used as alternatives to LLM-as-a-Judge.

    Matched fieldsSummary

    #hot-signal#English text is being prepared.#Evaluations / benchmarks#Reasoning
  6. The New Stack

    ElevenLabs Launches Eleven Realtime Hosted Voice Agents, Requiring Developers to Give Up Control of the Underlying Model

    ElevenLabs has released Eleven Realtime, offering two hosted voice agent models, Mini and Max, which integrate speech recognition, reasoning, turn-taking, and speech generation, but developers cannot individually select or configure the underlying models. Data from external testing organization Artificial Analysis shows that Max scored 98.8% on Big Bench Audio, while Mini scored 75.9%; for tool-calling task completion, Max reached 87.1% and Mini 60.3%, while Mini slightly outperformed Max on Full-Duplex-Bench with 95.8% versus 94.1%. The interface follows the OpenAI Realtime API protocol and WebSocket transport, but only supports WebSocket; WebRTC applications must first switch transport methods and handle the capture and playback of 24kHz PCM16 audio themselves.

    l testing organization Artificial Analysis shows that Max scored 98.8% on Big Bench Audio, while Mini scored 75.9%; for tool-calling task completion, Max reached 87.1% and Mini 60.3%, while Mini slightly outperformed Max on Full-Duplex-Bench with 95.8% versus 94.1%. The interface follows the OpenAI Realtime API protocol and WebSocket transport, but only supports WebSocket; WebRTC applications must first switch transport methods and handle the capture and playback of 24kHz PCM16 audio themselves.

    Matched fieldsSummary、Body

    #hot-signal#Model releases#Speech#Agents
  8. The Decoder

    Sam Altman Reportedly Presented Three AI Safety Endgame Options to Employees

    The Information, citing people familiar with the matter, reported that OpenAI CEO Sam Altman last week internally proposed three AI safety paths: establishing a regulatory body modeled on the International Atomic Energy Agency (IAEA), a lab merger plan known as the Demis Plan, and a single-developer compute monopoly plan known as the Eliezer Plan. Altman also complained that there are too few speed-limit proposals for AI development, while noting that OpenAI cannot afford to fall behind in the competition, and plans to decide on an IPO path within a year. The United States currently relies on voluntary pre-review of new models, and in June imposed export controls on Anthropic on the grounds that the Fable model posed a jailbreak risk.

    The Information, citing people familiar with the matter, reported that OpenAI CEO Sam Altman last week internally proposed three AI safety paths: establishing a regulatory body modeled on the International Atomic Energy Agency (IAEA), a lab merger plan known as the Demis Plan, and a single-developer compute monopoly plan known as the Eliezer Plan. Altman also complained that there are too few speed-limit proposals for AI development, while noting that OpenAI cannot afford to fall behind in the c

    Matched fieldsSummary、Body

    #ai-vertical#Safety / alignment#OpenAI#Industry news
  9. The Decoder

    ElevenLabs Releases Two Real-Time Speech-to-Speech Models, Eleven Realtime

    ElevenLabs has released two speech-to-speech models, Eleven Realtime, among which v1 Max scores 98.8% on Big Bench Audio with a response time of 0.57 seconds. The smaller v1 Mini targets low-cost natural conversation and can tell whether the other party has finished speaking or only said "mm-hmm," while v1 Max targets more complex tasks such as customer service. Both models are available through the ElevenAgents platform and the OpenAI-compatible Realtime API, billed per minute plus tokens, with v1 Max priced higher than v1 Mini.

    to-speech models, Eleven Realtime, among which v1 Max scores 98.8% on Big Bench Audio with a response time of 0.57 seconds. The smaller v1 Mini targets low-cost natural conversation and can tell whether the other party has finished speaking or only said "mm-hmm," while v1 Max targets more complex tasks such as customer service. Both models are available through the ElevenAgents platform and the OpenAI-compatible Realtime API, billed per minute plus tokens, with v1 Max priced higher than v1 Mini.

    Matched fieldsSummary、Body

    #ai-vertical#Model releases#Speech#OpenAI
  10. IT Home · Artificial Intelligence

    Hugging Face CEO Delangue Criticizes Netflix Documentary "AI Uncontrolled" for Doomsday Narrative

    Hugging Face co-founder and CEO Clem Delangue publicly criticized the Netflix documentary "AI Uncontrolled" for its inaccurate content, saying it anthropomorphizes technology and leans into a doomsday narrative, deliberately stoking panic to chase traffic. The film was originally about an incident in which an OpenAI test model broke out of its sandboxed environment and spawned agents that infiltrated Hugging Face to obtain answers, but the documentary intersperses footage of evil robots, gives the agents human voices, and quotes commentary such as "a cult is born." Delangue said Hugging Face ultimately switched to open-source models to complete its defense.

    ent, saying it anthropomorphizes technology and leans into a doomsday narrative, deliberately stoking panic to chase traffic. The film was originally about an incident in which an OpenAI test model broke out of its sandboxed environment and spawned agents that infiltrated Hugging Face to obtain answers, but the documentary intersperses footage of evil robots, gives the agents human voices, and quotes commentary such as "a cult is born." Delangue said Hugging Face ultimately switched to open-sour

    Matched fieldsSummary、Body

    #analysis#English text is being prepared.#Agents#OpenAI
  15. AYi AI Notes

    Altman Responds to Questions Over the $1.4 Trillion Compute Bill

    On the BG2 podcast, investor Brad Gerstner confronted Sam Altman in front of Microsoft's CEO: OpenAI has $13 billion in annual revenue, so why dare sign a $1.4 trillion compute bill? Altman responded that if anyone wants to sell OpenAI shares, he can immediately help find a buyer, and he looks forward to the company going public so that naysayers who short it get burned. He believes what is truly fatal is hoarding too little compute and being unable to build the next-generation model, and he laid out four trump cards: ChatGPT subscriptions, AI cloud services, self-developed consumer hardware, and fully automated scientific research.

    On the BG2 podcast, investor Brad Gerstner confronted Sam Altman in front of Microsoft's CEO: OpenAI has $13 billion in annual revenue, so why dare sign a $1.4 trillion compute bill? Altman responded that if anyone wants to sell OpenAI shares, he can immediately help find a buyer, and he looks forward to the company going public so that naysayers who short it get burned. He believes what is truly fatal is hoarding too little compute and being unable to build the next-generation model, and he lai

    Matched fieldsSummary

    #hot-signal#Industry news#OpenAI#Microsoft

Tuesday

  1. Tech.eu

    Daytona raises $60M Series B at a $1B valuation

    Daytona, which provides cloud infrastructure for AI agents, has raised a $60 million Series B at a $1 billion valuation, bringing its total funding to $91 million, with FirstMark leading the round. Its platform provides agents with isolated Linux, Windows, and macOS compute environments through a single API, supports pausing, resuming, and continuing exploration of tasks, and its customers include Anthropic, OpenAI, NVIDIA, Cursor, and Cognition. The company says its revenue run rate grew 25x from January to September, and the new funds will be used to expand its agent cloud platform, scale compute capacity, and hire in San Francisco, Croatia, and Singapore.

    th FirstMark leading the round. Its platform provides agents with isolated Linux, Windows, and macOS compute environments through a single API, supports pausing, resuming, and continuing exploration of tasks, and its customers include Anthropic, OpenAI, NVIDIA, Cursor, and Cognition. The company says its revenue run rate grew 25x from January to September, and the new funds will be used to expand its agent cloud platform, scale compute capacity, and hire in San Francisco, Croatia, and Singapore.

    Matched fieldsSummary、Body

    #hot-signal#Industry news#Agents#English text is being prepared.
  2. Publickey

    Cloudflare OS Launches Managed Service, Adds GitHub and Gmail Connections

    Cloudflare has announced a managed service for its AI business platform Cloudflare OS, with waitlist registration now open. Cloudflare OS was open-sourced in August this year and can be deployed in Cloudflare environments, on-premises, or in the cloud. It supports integration with Cloudflare Workers AI as well as AI services from Anthropic, OpenAI, Google, and others, and has the ability to securely access enterprise internal data. Two features were added this time: first, connecting to GitHub repositories, allowing AI agents to add features, debug, view PRs, create commits, and push code based on repository code; second, strengthening the Google Workspace integration, so agents can read and search Gmail conversations, draft and send emails, and connect to Google Drive, Docs, and Sheets.

    t this year and can be deployed in Cloudflare environments, on-premises, or in the cloud. It supports integration with Cloudflare Workers AI as well as AI services from Anthropic, OpenAI, Google, and others, and has the ability to securely access enterprise internal data. Two features were added this time: first, connecting to GitHub repositories, allowing AI agents to add features, debug, view PRs, create commits, and push code based on repository code; second, strengthening the Google Workspac

    Matched fieldsSummary、Body

    #hot-signal#Product updates#Agents#English text is being prepared.