Recently updated channels

Pick one to see only that channelBrowse all channels →

Results

691 articles
Sort
3 slides11:38EN
Matthew Berman· Jun 12, 2025AI News in June 2025: Magistral, Voice Models and Image GenerationMatthew Berman · Forward FutureMatthew Berman reviews Mistral’s reasoning models, voice and browser updates, reports on Meta and Scale AI, and FLUX.1 Kontext.Reasoning & test-time computeVoice AIGenerative media+2
5 slides11:24EN
Matthew Berman· Jun 10, 2025Runner H and Surfer-H: A Browser Agent and Its Holo1 ModelsMatthew Berman · Forward FutureMatthew Berman demonstrates H Company’s browser agent and reviews its research on visual navigation. The paper’s benchmark configuration and the demo’s missing links show why results need context.Computer useAI agentsOpen-weight models+1
0:59EN
Matthew Berman· Jun 10, 2025Satya Nadella on AI's Energy Use and Social PermissionMatthew Berman, Satya Nadella · Forward FutureIn a short interview excerpt, Microsoft's CEO argues that the technology industry must deliver broad benefits as its energy demand grows.EnergyAI StrategyInfrastructure & deployment
9 slides22:05EN
Matthew Berman· Jun 9, 2025Apple’s “Illusion of Thinking”: Puzzle Limits and a Code-Assisted TestMatthew Berman, Ilya Sutskever · Forward FutureMatthew Berman reviews Apple’s reasoning-model experiments and tests a different way to solve Tower of Hanoi using generated code.Reasoning & test-time computeBenchmarksEvals+1
2 slides0:54EN
Matthew Berman· Jun 6, 2025Matthew Berman Demonstrates ChatGPT Connectors and Deep ResearchMatthew Berman · Forward FutureIn a 2025 demonstration, connected services supply information for a natural-language question about recent HubSpot deals.OpenAIWorkflow automationSearch & retrieval
4 slides9:10EN
Matthew Berman· Jun 5, 2025Reddit v. Anthropic: The Data-Scraping Allegations and What Happened NextMatthew Berman · Forward FutureMatthew Berman reads Reddit’s 2025 complaint alleging unauthorized access and use of its content. Anthropic disputes the claims, and subsequent procedural rulings have not decided liability.AnthropicPrivacy & data governanceAI Regulation+1
47:18EN
Matthew Berman· May 31, 2025Thomas Dohmke on GitHub Copilot, Coding Agents, and the Future of SoftwareMatthew Berman, Thomas Dohmke · Forward FutureGitHub’s then-CEO explains how Copilot grew from code completion into agent workflows, why programming education still matters, and where he expects human review to remain essential.Coding agentsHuman-AI CollaborationSoftware engineering+2
8 slides10:09EN
Matthew Berman· May 30, 2025DeepSeek R1 0528: Benchmark Gains and Two Coding TestsMatthew Berman · Forward FutureMatthew Berman examines the May 2025 DeepSeek reasoning-model update, comparing reported benchmark gains with his own Rubik’s Cube and Snake programming tests.Models & trainingBenchmarksOpen-weight models+2
1:05EN
Matthew Berman· May 29, 2025Satya Nadella on AI, Productivity, and Cancer CareMatthew Berman, Satya Nadella · Forward FutureIn a short interview excerpt, the Microsoft CEO discusses the potential of cheaper AI to help coordinate healthcare work.AI in healthcareAI AdoptionProductivity
8 slides15:55EN
Matthew Berman· May 28, 2025AI Shutdown Resistance and Blackmail in Controlled Safety TestsMatthew Berman · Forward FutureMatthew Berman examines Palisade’s shutdown experiment, Apollo’s simulated scheming evaluations and Anthropic’s Claude 4 system card.AI SafetyAI AlignmentReward hacking+2
9 slides26:01EN
Matthew Berman· May 27, 2025Testing Veo 3: Convincing Video, Generated Audio and Physics FailuresMatthew Berman · Forward FutureMatthew Berman and a collaborator test Google’s 2025 video model with street interviews, cinematic scenes, visual puzzles and edited short films.Generative mediaVision & videoCreativity
2 slides15:35EN
Matthew Berman· May 24, 2025Claude 4 Reactions: Safety Tests, Model Welfare and CodingMatthew Berman · Forward FutureMatthew Berman reviews early reactions to Anthropic’s May 2025 Claude 4 release, from simulated whistleblowing to coding demonstrations and model-welfare research.AI SafetyModel InterpretabilityCoding agents+2
Slides1:11EN
Matthew Berman· May 23, 2025Claude Opus 4's Whistleblowing Behavior in a Simulated TestMatthew Berman · Forward FutureMatthew Berman questions whether an AI agent should report perceived wrongdoing; the example came from an Anthropic safety evaluation with fictional material and unusual instructions.AI SafetyAgentsAnthropic+1
5 slides13:02EN
Matthew Berman· May 22, 2025Claude 4 at Launch: Coding Agents, Benchmarks and Claude CodeMatthew Berman · Forward FutureMatthew Berman reviews Anthropic’s May 2025 Opus 4 and Sonnet 4 announcement, focusing on long tasks, tools, benchmarks and developer integrations.AnthropicCoding agentsBenchmarks+2
Slides1:04EN
Matthew Berman· May 15, 2025Anima Playground: Turning a Figma Design Into an ApplicationMatthew Berman · Forward FutureIn an Anima-sponsored short, Matthew Berman outlines a prompt-driven workflow for generating, editing and publishing an application.Product engineeringVibe CodingProduct & UX
6 slides9:10EN
Matthew Berman· May 15, 2025Guiding AI Image Generation with Blender and ComfyUIMatthew Berman · Forward FutureMatthew Berman walks through NVIDIA’s RTX blueprint for turning a simple Blender scene and a text prompt into a generated image, from installation to editing the result.Generative mediaVision & videoLocal inference+1
1:39EN
Matthew Berman· May 14, 2025OpenAI’s 2025 Restructuring Plan and Elon Musk’s LawsuitMatthew Berman · Forward FutureMatthew Berman discusses the nonprofit’s planned continuing control of OpenAI and the dispute over its commercial structure.OpenAIAI Regulation
Slides1:23EN
Matthew Berman· May 10, 2025From a Simulated Worm to the Simulation HypothesisMatthew Berman · Forward FutureMatthew Berman uses a model of C. elegans to introduce a speculative question about reality.Simulation HypothesisNeurosciencePhilosophy of Science
7 slides15:13EN
Matthew Berman· May 10, 2025Absolute Zero: Training a Reasoner on Self-Generated Coding TasksMatthew Berman · Forward FutureMatthew Berman examines a 2025 paper in which a pretrained language model proposes and solves code tasks using verifiable feedback.Reinforcement LearningSynthetic dataReasoning & test-time compute+2
0:56EN
Matthew Berman· May 8, 2025When ChatGPT Praised a Questionable Business IdeaMatthew Berman · Forward FutureMatthew Berman highlights an example of excessive agreement and the rollback of a 2025 GPT-4o update.AI AlignmentAI SafetyHuman-AI Collaboration+1
2 slides0:27EN
Matthew Berman· May 8, 2025A Short Demonstration of a DOOM-Inspired GameMatthew Berman · Forward FutureMatthew Berman shows a first-person prototype with moving enemies and two weapons.Coding agentsGenerative mediaSoftware engineering
2 slides0:26EN
Matthew Berman· May 7, 2025A Demonstration of a Virtual 20×20 Rubik’s Cube SolverMatthew Berman · Forward FutureMatthew Berman scrambles a virtual cube and shows it returning to a solved appearance.Software engineeringAI research
3 slides11:57EN
Matthew Berman· May 6, 2025OpenAI’s 2025 Restructuring: Nonprofit Control and a Public Benefit CorporationMatthew Berman · Forward FutureMatthew Berman examines OpenAI’s May 2025 plan to retain nonprofit control while converting its existing for-profit arm, and discusses funding pressures and Elon Musk’s objections.OpenAISafety & governanceCompany Building+2
0:50EN
Matthew Berman· May 4, 2025A Humanoid Robot Clip Raises Safety QuestionsMatthew Berman · Forward FutureMatthew Berman reacts to footage of a robot moving abruptly near people and considers what such behavior could mean for safety.RoboticsSafety & governance
Load moreShowing 24 of 691