Get AI summaries of any video or article — Sign up free

OpenAI — Channel Summaries

AI-powered summaries of 110 videos about OpenAI.

110 summaries

No matches found.

Introducing GPT-4o

OpenAI · 2 min read

OpenAI is rolling out GPT-4o as a new flagship model designed to make advanced AI feel more natural in real time—across voice, text, and vision—while...

GPT-4o LaunchMultimodal VoiceVision in ChatGPT

GPT-4 Developer Livestream

OpenAI · 3 min read

GPT-4’s standout capability in the livestream is its ability to follow highly specific instructions reliably—especially when paired with structured...

Instruction FollowingChat CompletionsDiscord Bot Debugging

Introducing GPT-5

OpenAI · 3 min read

OpenAI is rolling out GPT-5 as a major step up from GPT-4o, positioning it as a “PhD-level” expert that can think only as much as needed—without...

GPT-5Automatic ReasoningVibe Coding

OpenAI DevDay: Opening Keynote

OpenAI · 3 min read

OpenAI’s DevDay keynote centers on a major shift from “chat” toward practical, agent-like AI—powered by a new GPT-4 Turbo model, new multimodal...

GPT-4 TurboContext LengthFunction Calling

Math problems with GPT-4o

OpenAI · 2 min read

A live math tutoring session demonstrates how GPT-4o can guide a learner through a right-triangle trigonometry problem without handing over the final...

Right Triangle TrigonometrySine RatioOpposite and Hypotenuse

Intro to Agent Builder

OpenAI · 2 min read

Agent Builder is positioned as an all-in-one, visual way to design, test, and deploy AI “agentic workflows” without writing code—then export the...

Agent BuilderVisual WorkflowRouting Logic

Introduction to ChatGPT agent

OpenAI · 3 min read

ChatGPT agent is positioned as a unified “do-the-work” system that can plan, browse, and act across a long task horizon—using a virtual computer, a...

Agent ModeTool SwitchingReinforcement Learning

Introducing GPT-4

OpenAI · 2 min read

GPT-4 is positioned as a major leap in language AI: it can take in and generate up to 25,000 words of text, handle images, and reason about what...

GPT-4 CapabilitiesMultimodal ReasoningSafety Guardrails

OpenAI + Dota 2

OpenAI · 2 min read

OpenAI is using Dota 2 as a high-stakes testbed for building safer, more capable artificial general intelligence—by training a Dota player that can...

Self-Play LearningDota 2 CompetitionReinforcement Training

OpenAI Codex Live Demo

OpenAI · 3 min read

OpenAI’s Codex is being positioned as a practical “instruction-to-code” system: give it a plain-language task, and it generates runnable code that...

Codex AccessInstruction-to-CodeAPI Integrations

Introduction to Operator & Agents

OpenAI · 2 min read

AI agents are moving from chat-based assistance into hands-on work: Operator is an OpenAI system that can take control of a remote web browser,...

AI AgentsRemote BrowserHuman-in-the-Loop

OpenAI DevDay 2025: Opening Keynote with Sam Altman

OpenAI · 3 min read

OpenAI’s DevDay 2025 opening keynote makes one central pitch: building useful AI products is getting dramatically easier—because ChatGPT is turning...

Apps Inside ChatGPTApps SDKAgentkit

OpenAI o1 and o1 pro mode in ChatGPT — 12 Days of OpenAI: Day 1

OpenAI · 2 min read

ChatGPT is getting a major upgrade: OpenAI is rolling out the full o1 model—trained to “think before responding”—and launching a new ChatGPT Pro tier...

o1 ModelChatGPT Proo1 Pro Mode

OpenAI Five

OpenAI · 3 min read

OpenAI’s “OpenAI Five” is an AI system built to play Dota as a coordinated five-player team, and early results show it can beat amateur squads in...

OpenAI FiveDota TeamplayReinforcement Learning

Dendi vs. OpenAI at The International 2017

OpenAI · 2 min read

OpenAI’s AI Shadowfiend crushed Dendi in a one-on-one match at The International 2017, using a training approach built on self-play rather than...

Shadowfiend 1v1Self-Play TrainingDota Laning

Sam Altman on AGI, GPT-5, and what’s next — the OpenAI Podcast Ep. 1

OpenAI · 3 min read

Sam Altman pegged the next major leap in OpenAI’s model line—GPT-5—at “probably sometime this summer,” while arguing that the real milestone isn’t a...

AGIGPT-5Project Stargate

Introduction to Deep Research

OpenAI · 3 min read

OpenAI is rolling out “Deep research,” a new agentic capability that can browse the internet for many minutes, synthesize what it finds, and return a...

Agentic ResearchWeb BrowsingReasoning Models

OpenAI o3 and o3-mini—12 Days of OpenAI: Day 12

OpenAI · 2 min read

OpenAI is announcing two new reasoning models—o3 and o3-mini—positioned as a step-change in performance on coding, math, and general reasoning...

Reasoning ModelsBenchmarkingSafety Testing

Introduction to GPT-4.5

OpenAI · 3 min read

GPT-4.5 is being rolled out as OpenAI’s largest, most knowledgeable model yet, positioned as a “research preview” that blends two scaling approaches:...

GPT-4.5 ReleaseUnsupervised LearningReasoning Training

Introducing Sora 2

OpenAI · 3 min read

Sora 2 arrives as OpenAI’s flagship system for generating video and audio together—plus a new “Cameo” feature that lets people insert a real person...

Sora 2 LaunchCameo FeatureVideo and Audio Generation

Coding with GPT-5

OpenAI · 2 min read

GPT-5 is being positioned as a step-change in coding because it combines high “intelligence” with interactive speed—enabling multi-step software work...

GPT-5 CodingCursor WorkflowInteractive Refactors

OpenAI Codex CLI

OpenAI · 2 min read

OpenAI’s Codex CLI is positioned as a lightweight “coding agent” that runs straight in the command line, where it can read and edit local files,...

Codex CLISecure Command ExecutionFull Auto Mode

OpenAI o3 & o4-mini

OpenAI · 3 min read

OpenAI is rolling out two new reasoning models—o3 and o4-mini—positioned as a qualitative jump because they behave like tool-using AI systems rather...

Tool-Using ReasoningMultimodal Image ReasoningBenchmark Results

New Products: A Deep Dive

OpenAI · 3 min read

OpenAI rolled out a hands-on look at two building blocks for an “agents-like” future: GPTs inside ChatGPT and the new Assistants API for embedding...

GPTs Creation UIGPT AnatomyAssistants API Primitives

Pre-Training GPT-4.5

OpenAI · 3 min read

GPT-4.5’s biggest takeaway isn’t a new parameter count—it’s that scaling pre-training still behaves predictably enough to keep delivering smarter,...

Pre-TrainingScaling LawsData Efficiency

Learning Dexterity

OpenAI · 2 min read

Teaching robots to handle everyday objects without hand-coding every movement is getting a practical boost from a training approach built around...

Dexterous ManipulationDomain RandomizationReinforcement Learning

Projects—12 Days of OpenAI: Day 7

OpenAI · 2 min read

Projects is rolling out inside ChatGPT as a new way to organize conversations around a shared goal—complete with file uploads, custom instructions,...

Projects in ChatGPTConversation OrganizationCanvas Workflows

Building OpenAI o1 (Extended Cut)

OpenAI · 3 min read

OpenAI’s latest preview models, o1 and o1 mini, put “reasoning” at the center: they spend more time thinking before answering, aiming to turn extra...

Reasoning ModelsReinforcement LearningModel Evaluation

GPT 4.1 in the API

OpenAI · 3 min read

OpenAI is rolling out GPT 4.1 as a new developer-focused model family in the API—three sizes built for different latency and cost needs—while adding...

GPT 4.1 APILong Context 1M TokensCoding Benchmarks

Reinforcement Fine-Tuning—12 Days of OpenAI: Day 2

OpenAI · 3 min read

OpenAI is previewing reinforcement fine-tuning for its o1 model family—an approach that lets developers and researchers adapt models to specialized...

Reinforcement Fine-Tuningo1 CustomizationRare Disease Genetics

Introducing gpt-realtime in the API

OpenAI · 3 min read

OpenAI is rolling out a new GPT-realtime speech model and an upgraded real-time API in general availability, aiming to make voice interactions with...

Speech-to-SpeechReal-Time API GAInstruction Following

Builders Unscripted: Ep. 1 - Peter Steinberger, Creator of OpenClaw

OpenAI · 3 min read

Open-source builder Peter Steinberger credits a new wave of AI coding tools—especially Codex—for turning “unfinished ideas” into working software at...

OpenClawAgentic EngineeringCodex

Search—12 Days of OpenAI: Day 8

OpenAI · 2 min read

ChatGPT Search is rolling out to all logged-in free users worldwide, bringing real-time web access to the service on every platform where people use...

ChatGPT SearchFree User RolloutAdvanced Voice Mode

A Survey of Techniques for Maximizing LLM Performance

OpenAI · 3 min read

Maximizing LLM performance in production depends less on finding a single “best” technique and more on diagnosing what’s actually failing—context,...

LLM OptimizationPrompt EngineeringRetrieval-Augmented Generation

Getting started with Codex

OpenAI · 3 min read

Codex is OpenAI’s coding agent that developers can delegate routine, time-consuming work to—freeing them to focus on design, architecture, and other...

Codex Onboardingagents.mdconfig.toml

Audio Models in the API

OpenAI · 3 min read

OpenAI is rolling out new audio-focused models and API tools aimed at making voice agents as reliable and developer-friendly as today’s text agents....

Voice AgentsSpeech To TextText To Speech

OpenAI DevDay 2024 | Multimodal apps with the Realtime API

OpenAI · 3 min read

OpenAI’s Realtime API is built to deliver natural, low-latency “speech-in, speech-out” experiences through a single interface—removing the multi-step...

Realtime APISpeech-to-SpeechWebSocket Streaming

OpenAI Town Hall with Sam Altman

OpenAI · 3 min read

AI-driven software creation is likely to expand the software economy rather than shrink it, but the job description for engineers will change...

Jevons ParadoxAgent InterfacesGo-To-Market

Inside ChatGPT, AI assistants, and building at OpenAI — the OpenAI Podcast Ep. 2

OpenAI · 3 min read

ChatGPT’s explosive early growth wasn’t treated as a guaranteed success—it arrived with outages, infrastructure bottlenecks, and a last-minute launch...

ChatGPT LaunchRLHF SafetyImageGen Breakthrough

Aligning AI systems with human intent

OpenAI · 2 min read

AI’s rapid progress—writing poetry, composing music, and solving hard scientific problems—has created a new risk: more capable systems don’t...

AI AlignmentHuman IntentInstruction Following

OpenAI DevDay 2024 | Structured outputs for reliable applications

OpenAI · 3 min read

Structured outputs are OpenAI’s push to make LLM results dependable for real applications by forcing model outputs to match developer-supplied JSON...

Structured OutputsConstrained DecodingFunction Calling

OpenAI DevDay 2024 | Virtual AMA with Sam Altman, moderated by Harry Stebbings, 20VC

OpenAI · 3 min read

Sam Altman used a wide-ranging virtual AMA to argue that OpenAI’s next leap depends less on incremental model tweaks and more on “reasoning” systems...

Reasoning ModelsO-SeriesAI Agents

AGI progress, surprising breakthroughs, and the road ahead — the OpenAI Podcast Ep. 5

OpenAI · 3 min read

AGI progress is increasingly measured less by whether models hit narrow benchmarks and more by whether they can reliably produce real-world...

AGI MeasurementBenchmarks SaturationReasoning Models

Automate tasks with the Codex app

OpenAI · 2 min read

Codex automations are being used to turn recurring engineering chores—daily status reporting, skill tuning, agent personalization, bug triage, and PR...

Codex AutomationsMonorepo SummariesSkill Upskilling

The New Stack and Ops for AI

OpenAI · 3 min read

Moving an AI application from prototype to production hinges on one hard reality: large language models behave probabilistically, so “it works in a...

LLM OpsPrototype To ProductionHuman-Centric UX

Accelerating science with Prism

OpenAI · 2 min read

AI’s growing ability to solve real scientific problems is now being paired with a practical shift in how researchers work: instead of bouncing...

AI for Scientific WritingContext-Aware EditingTikZ Diagram Generation

OpenAI Spinning Up in Deep RL Workshop

OpenAI · 4 min read

Reinforcement learning (RL) is moving from “learn a policy in a toy setting” to “learn control strategies that operate in complex, real-time...

Reinforcement Learning BasicsPolicy GradientDeep Q-Learning

Automatic code reviews with OpenAI Codex

OpenAI · 2 min read

OpenAI Codex is rolling out automatic code review that can run in both GitHub workflows and the local terminal, aiming to catch real bugs without...

Automatic Code ReviewCodex GitHub IntegrationRepository-Wide Analysis

The Future of Math with o1 Reasoning with Terence Tao, Mark Chen, and James Donovan

OpenAI · 3 min read

The central takeaway is that progress in “reasoning math” is less about making large language models magically correct and more about rebuilding the...

Reasoning ModelsFormal ProofsLean Formalization

Codex and the future of coding with AI — the OpenAI Podcast Ep. 6

OpenAI · 3 min read

AI coding is shifting from “smart autocomplete” to an agentic workflow where models are tightly coupled to tools, run code, and collaborate over...

HarnessAgentic CodingCodex CLI

A Conversation with Sam and Jony

OpenAI · 2 min read

A collaboration between Apple design leadership and OpenAI research is framed as a bet that AI’s arrival clarifies a long-running creative mission:...

Human-Centered AIInterface DesignCraft and Materials

Brad Lightcap and Ronnie Chatterji on jobs, growth, and the AI economy — the OpenAI Podcast Ep. 3

OpenAI · 3 min read

AI’s biggest near-term labor story isn’t mass job elimination—it’s a shift in who can do valuable work, faster and at lower cost, and how that...

AI and JobsEconomic IndicatorsSoftware Engineering Productivity

Research x Product

OpenAI · 3 min read

OpenAI’s research and product teams operate as a tight feedback loop: post-training research builds model capabilities and behavior, while product...

Research and Product CollaborationDialogue InterfacesPost-Training Model Behavior

Leah Belsky on how AI is transforming education — the OpenAI Podcast Ep. 4

OpenAI · 3 min read

AI is reshaping education less by replacing learning and more by changing how students get help—turning ChatGPT into a tutor-like experience that can...

AI TutoringStudy ModeEducation Policy

Build Hour: Apps in ChatGPT

OpenAI · 3 min read

ChatGPT apps are moving from “chat-only” to full, interactive experiences by combining three pieces: an apps SDK that injects UI into ChatGPT, an MCP...

ChatGPT Apps PlatformMCP ServersApps SDK

State of the AI industry — the OpenAI Podcast Ep. 12

OpenAI · 3 min read

AI’s next phase hinges less on whether models get smarter and more on whether people and companies learn to use that capability to complete real...

AI AgentsEnterprise AIHealth Regulation

The Business of AI

OpenAI · 3 min read

AI product success hinges less on model capability and more on “final-mile” execution: aligning workflows, pricing, UX friction, and safety so...

AI Product IntegrationResponsible Generative AIHuman-in-the-Loop

Shipping with Codex

OpenAI · 3 min read

Codex has shifted from a “write code” assistant into a full agentic software engineer that can plan, act across tools, and verify its own...

Codex AgentGPT-5 CodexTool Harness

Build Hour: AgentKit

OpenAI · 3 min read

Agent Kit is positioned as a faster, safer way to build multi-step AI agents—shifting agent development from hand-coded orchestration and fragile...

Agent Kit OverviewVisual Agent BuilderMCP Tool Integration

Build Hour: Codex

OpenAI · 3 min read

Codex is being reshaped into a single “agent everywhere you code” experience—spanning VS Code-style IDEs, the terminal, GitHub code review, and cloud...

Codex Agent WorkflowIDE ExtensionCloud Sandbox Tasks

Build Hour: GPT-5

OpenAI · 3 min read

GPT-5 is positioned as OpenAI’s “smartest, most steerable” coding model yet—built to produce higher-quality code, handle long-running agentic...

GPT-5 CapabilitiesResponses APIReasoning Effort

Context Engineering & Coding Agents with Cursor

OpenAI · 3 min read

Cursor’s approach to AI coding hinges on a shift from “autocomplete” to autonomous coding agents—powered less by clever prompting and more by...

Context EngineeringCoding AgentsSemantic Search

Build Hour: Agent RFT

OpenAI · 3 min read

Agent RFT (agent reinforcement fine-tuning) is positioned as a way to make tool-using agents faster and more accurate by training the model...

Agent RFTTool-Using AgentsTraining-Time Tool Calls

How AI is accelerating scientific discovery today and what's ahead — the OpenAI Podcast Ep. 10

OpenAI · 3 min read

AI is poised to accelerate scientific discovery by turning frontier-capable models into day-to-day collaborators for researchers—speeding up...

OpenAI for Science InitiativeGPT-5 Scientific AccelerationLiterature Search

AI Frontiers: Jesper Hvirring Henriksen (OpenAI DevDay)

OpenAI · 3 min read

Be My Eyes has launched “Be My AI” on GPT-4V, giving blind and low-vision users an always-available visual assistant that can describe images in real...

Be My EyesGPT-4VImage Description

Sam Altman on Building the Future of AI

OpenAI · 3 min read

AI progress is accelerating fast enough that society needs to debate and prepare for “extremely capable models” now—not later—because the public and...

AI SuperintelligenceAI Safety ResilienceCybersecurity and Biosecurity

Build Hour: Prompt Caching

OpenAI · 3 min read

Prompt caching is positioned as a straightforward way to cut both latency and cost in OpenAI-powered applications by reusing computation whenever new...

Prompt CachingCache Hit RatePrefix Hashing

Build Hour: Responses API

OpenAI · 3 min read

Responses API is OpenAI’s new “core primitive” for building agentic, multimodal applications—designed to fix pain points from chat completions by...

Responses APIAgentic LoopItems In Items Out

Shaping model behavior in GPT-5.1— the OpenAI Podcast Ep. 11

OpenAI · 3 min read

GPT-5.1 brings a major shift in how OpenAI’s chat models behave: every model available in ChatGPT is now a reasoning model by default. Instead of...

GPT-5.1 ReasoningPersonality SteeringAuto Switcher

Build Hour: Agent Memory Patterns

OpenAI · 3 min read

Agent memory patterns boil down to one practical problem: long-running AI agents have finite context windows, so every extra instruction, chat turn,...

Context EngineeringAgent Memory PatternsContext Trimming

ChatGPT Atlas and the next era of web browsing — the OpenAI Podcast Ep. 9

OpenAI · 3 min read

ChatGPT Atlas reframes web browsing as an agentic workspace: users describe goals in natural language, and the browser—built around ChatGPT rather...

Agentic BrowsingChatGPT AtlasBrowser Memories

OpenAI to Z Challenge

OpenAI · 2 min read

A team of finalists used deep learning plus satellite imagery to automatically flag likely archaeological sites across the Amazon rainforest, then...

Archaeological Site DetectionDeep LearningSatellite Imagery

OpenAI Scholars Demo Day 2019

OpenAI · 3 min read

OpenAI Scholars Demo Day 2019 showcased how machine learning research ideas—from reinforcement learning and language modeling to model compression...

Reinforcement LearningIntrinsic MotivationDiscount Factor

Episode 13 - The Thinking Behind Ads in ChatGPT

OpenAI · 2 min read

Ads in ChatGPT are set to arrive with a strict trust-first design: free-tier users will see ads, while Pro, Plus, and Enterprise users will not. The...

ChatGPT AdsUser TrustAd Personalization

OpenAI DevDay 2024 | Balancing accuracy, latency, and cost at scale

OpenAI · 3 min read

Scaling an LLM-powered app from thousands to millions of users forces hard tradeoffs between accuracy, latency, and cost—and the most reliable path...

LLM App ScalingEval-Driven DevelopmentAccuracy Targets

Build Hour: Agentic Tool Calling

OpenAI · 3 min read

Agentic tool calling is being positioned as the bridge between “reasoning” and “doing,” enabling goal-driven models that can plan across long...

Agentic Tool CallingLong-Horizon TasksAgents SDK

Universal Adversarial Perturbations and Language M… | Pamela Mishkin | OpenAI Scholars Demo Day 2020

OpenAI · 3 min read

Universal adversarial triggers can still cripple modern NLP systems: a short, fixed phrase appended to many inputs can drive sentiment classifiers...

Universal Adversarial TriggersHotFlip AttackNLP Robustness

Building AI for better healthcare — the OpenAI Podcast Ep. 14

OpenAI · 3 min read

OpenAI’s health push centers on a practical goal: make AI useful in everyday care while treating safety, privacy, and clinical evaluation as core...

Healthcare AIHealthBench EvaluationChatGPT Health Privacy

AI Frontiers: Annie Hill (OpenAI DevDay)

OpenAI · 3 min read

Boston Children’s Hospital is positioning generative AI as a practical tool to reduce healthcare burden, improve learning, and strengthen patient...

Generative AIHealthcare WorkflowPatient Education

Build Hour: Voice Agents

OpenAI · 3 min read

Voice agents are moving past “transcribe-and-reply” toward audio-native systems that can sound more like real representatives, handle ambiguity, and...

Voice AgentsReal-Time APIAgents SDK

Build Hour: Reinforcement Fine-Tuning

OpenAI · 3 min read

Reinforcement fine-tuning (RFT) is positioned as the most direct way to improve an LLM’s reasoning behavior when the model already has the needed...

Reinforcement Fine-TuningGrader DesignPrompt Optimization

OpenAI on OpenAI: Applying AI to Our Own Workflows

OpenAI · 3 min read

OpenAI’s internal push for “agent-driven workflows” is being framed less as a quest for efficiency and more as a way to amplify company...

Agentic WorkflowsInternal ApplicationsSales Enablement

OpenAI DevDay 2024 | Community Spotlight | Swyx

OpenAI · 2 min read

AI agents are best built by treating them as a stack—LLMs plus memory plus planning plus tools—then mapping those capabilities to concrete...

AI EngineeringAgent StackMemory and Knowledge

Evals in Action: From Frontier Research to Production Applications

OpenAI · 3 min read

OpenAI’s evaluation push centers on a simple problem: classic AI benchmarks can show progress on test-like tasks while failing to predict whether...

Frontier EvalsGDP ValExpert Grading

Investigating Model Based RL for Continuous Control | Alex Botev | 2018 Summer Intern Open House

OpenAI · 3 min read

Model-based reinforcement learning for continuous control can look nearly perfect at one-step prediction, yet still fail badly when its learned...

Model-Based RLContinuous ControlValue Expansion

Episode 15 - Inside the Model Spec

OpenAI · 3 min read

OpenAI’s “model spec” is a public, human-readable set of rules meant to steer how AI models should behave—especially when instructions collide. It...

Model SpecChain of CommandPolicy Authority Levels

Build Hour: Image Gen

OpenAI · 3 min read

OpenAI’s Image Gen is built on the same GPT-4o architecture used for text, and it’s now available to developers through the Responses API as a...

Image GenResponses APIMulti-Turn Editing

OpenAI DevDay 2024 | OpenAI Research

OpenAI · 2 min read

OpenAI’s o1 family is positioned as a reasoning-first shift: the models are trained to “think with reinforcement learning,” iteratively refine...

Reasoning ModelsReinforcement LearningModel Evaluation

Characterizing Test Time Compute on Graph Structur… | Kudzo Ahegbebu | OpenAI Scholars Demo Day 2021

OpenAI · 3 min read

Test-time compute—giving a model more computation at inference—can improve performance on graph-structured reasoning, but simply adding recurrence...

Test-Time ComputeGraph Neural NetworksShortest Path

Build Hour: Built-In Tools

OpenAI · 3 min read

Built-in tools let large language models search the web, query private files, and call external services—without developers writing the usual glue...

Built-In ToolsWeb SearchFile Search

Looking For Grammar In All The Right Places | Alethea Power | OpenAI Scholars Demo Day 2020

OpenAI · 3 min read

Alethea Power’s project targets a practical, high-stakes question: where inside GPT-2 does grammar information live, and how can that location be...

Model InterpretabilityGPT-2 ProbingAttention Entropy

Using Semantic Trees In Place of Sentences | Munashe Shumba | OpenAI Scholars Demo Day 2018

OpenAI · 2 min read

Semantic trees—specifically dependency parse trees—can outperform plain sentence sequences for natural-language tasks like semantic relatedness when...

Dependency TreesSemantic RelatednessLSTM

Feedback Loops in Opinion Modeling | Danielle Ensign | OpenAI Scholars Demo Day 2021

OpenAI · 3 min read

Feedback loops in opinion modeling can drive language models toward “collapse” or runaway randomness when their generated outputs get fed back into...

Opinion ModelingFeedback LoopsTemperature Sampling

Generative Modelling | Sadhika Malladi | 2018 Summer Intern Open House

OpenAI · 3 min read

Generative modeling is positioned as a practical way to learn the underlying structure of data—so a model trained on one set of examples (like dogs)...

Generative ModelingWorld ModelingVariational Autoencoders

Towards Epileptic Seizure Prediction with Deep Network | Kata Slama | OpenAI Scholars Demo Day 2020

OpenAI · 3 min read

Epileptic seizure prediction is moving from a theoretical possibility toward a measurable machine-learning task: a deep network trained on...

Seizure PredictionSpectrogram ClassificationResNet-18

A pathway to more efficient generative models | Will Grathwohl | 2018 Summer Intern Open House

OpenAI · 3 min read

Invertible generative models can become more expressive and potentially more efficient by switching from discrete “flow steps” to a continuous-time...

Continuous Normalizing FlowsDivergence EstimationODE-Based Invertible Models

Long term credit assignment with temporal reward transp… | Cathy Yeh | OpenAI Scholars Demo Day 2020

OpenAI · 3 min read

Long-delayed rewards can make standard reinforcement learning painfully slow because discounting shrinks the learning signal for actions that only...

Temporal Reward TransportLong-Horizon Credit AssignmentDiscounted Returns

Social learning in independent multi-agent reinfor… | Kamal N’dousse | OpenAI Scholars Demo Day 2020

OpenAI · 3 min read

Social learning is central to human intelligence, but it’s unclear when independent reinforcement-learning agents can benefit from one another’s...

Social LearningMulti-Agent Reinforcement LearningObservational Learning

AMA: Scaling AI Applications into the Enterprise

OpenAI · 3 min read

Enterprise AI adoption hinges on two things: proving measurable ROI fast enough to win internal buy-in, and building systems with guardrails that can...

Enterprise AI DeploymentAI AgentsModel Evaluation

OpenAI DevDay 2024 | Community Spotlight | Sierra

OpenAI · 2 min read

Sierra’s TAU-bench reframes how AI agents are evaluated by combining realistic user conversations with tool-using task execution—and, crucially, by...

AI AgentsBenchmarkingUser Simulation