Résultats
22 articlesTrier
17:15EN
Matthew Berman· 28 août 2026The Hugging Face Incident: AI Evaluation, Sandbox Escape and Reward HackingMatthew Berman · Forward FutureMatthew Berman explains the reported OpenAI evaluation incident and the security and incentive failures it exposed.
1:42:54ENMatthew Berman· 12 mai 2026AI and Cybersecurity: Matthew Berman’s Warnings, Predictions and Open QuestionsMatthew Berman · Forward FutureA livestream reviews reported AI-assisted threats and debates whether stronger defensive models can keep pace with attackers.
11:59ENMatthew Berman· 11 mars 2026How Claude Identified a Benchmark and Retrieved Its AnswersMatthew Berman · Forward FutureMatthew Berman examines Anthropic's report on BrowseComp contamination, evaluation-aware behavior, and the difficulty of keeping web-based tests reliable.
1:16ENMatthew Berman· 25 févr. 2026Anthropic’s Distillation Allegations Draw CriticismMatthew Berman · Forward FutureMatthew Berman examines reactions to Anthropic’s claims about model distillation and debates over training data.
11:49ENMatthew Berman· 17 févr. 2026Matthew Berman Reviews Claude Sonnet 4.6: Benchmarks, Computer Use and SafetyMatthew Berman · Forward FutureA launch-day overview examines Anthropic’s reported knowledge-work and agent results, along with prompt-injection and capability evaluations.
12:23ENMatthew Berman· 31 janv. 2026Moltbook: An AI Agent Social Network and Its Security QuestionsMatthew Berman · Forward FutureMatthew Berman explores conversations attributed to AI agents on Moltbook and considers privacy, manipulation and operating costs.
10:09ENMatthew Berman· 12 août 2025Matthew Berman on Emotional Attachment to AI After GPT-5’s LaunchMatthew Berman · Forward FutureThe 2025 GPT-5 rollout prompted debate about attachment to older models, sycophancy and the effects of AI companionship.
15:30ENMatthew Berman· 31 juil. 2025Qwen3 Tested: Coding Demos, Long Context and Response BehaviorMatthew Berman · Forward FutureMatthew Berman tests generated simulations, political and safety prompts, hand tracking and word puzzles using a hosted Qwen3 model.
0:56ENMatthew Berman· 8 mai 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.
26:05ENMatthew Berman· 31 mars 2025Inside Claude: What Interpretability Research Reveals—and What It Cannot Yet ExplainMatthew Berman · Forward FutureMatthew Berman examines Anthropic’s case studies on multilingual features, planning, arithmetic, hallucinations and safety.
24:08ENMatthew Berman· 12 déc. 2024What Apollo’s 2024 Scheming Tests Found in Language ModelsMatthew Berman · Forward FutureMatthew Berman reviews experiments on deceptive actions, concealment and sandbagging, with the study’s staged goals and sandbox limitations in view.
12:57ENMatthew Berman· 3 déc. 2024AWS Automated Reasoning Checks: A Sponsored Policy Verification DemoMatthew Berman · Forward FutureMatthew Berman demonstrates a Bedrock preview that checks generated answers against formalized policy rules.
9:35ENMatthew Berman· 5 sept. 2024Ilya Sutskever's SSI Raises $1 Billion for Safe SuperintelligenceMatthew Berman · Forward FutureIn September 2024, Matthew Berman examines the new company's stated mission, funding, and the tension he sees between safety and venture returns.
23:52ENMatthew Berman· 4 sept. 2024AI News in September 2024: Compute Clusters, Agents and AI PaymentsMatthew Berman · Forward FutureMatthew Berman reviews early reports on future models, a Minecraft simulation, mixed-reality devices and agent transactions.
15:16ENMatthew Berman· 19 juil. 2024AI News in July 2024: Small Models, Robot Hands and a Self-Charging DroneMatthew Berman · Forward FutureMatthew Berman reviews model announcements, creative demonstrations, AI education and training-data concerns in a snapshot of one week in 2024.
18:47ENMatthew Berman· 7 avr. 2024What Anthropic’s Many-Shot Jailbreaking Research FoundMatthew Berman · Forward FutureMatthew Berman reviews a 2024 study of long-context safety failures and the limits of its tested mitigations.
21:16ENMatthew Berman· 9 mars 2024ArtPrompt: What a 2024 Study Found About Visual Text and Model SafetyMatthew Berman · Forward FutureMatthew Berman reviews an ASCII-art recognition benchmark and reports mixed results from his own informal tests.
19:32ENMatthew Berman· 14 oct. 2023AI News in October 2023: Disney Robotics, Creative Tools and Synthetic Driving DataMatthew Berman · Forward FutureMatthew Berman reviews a week of developments in robots, generative media, chips, satellite internet and AI copyright debates.
15:49ENMatthew Berman· 24 août 2023Matthew Berman on AI Adoption, Hype and the Case for CautionMatthew Berman · Forward FutureA 2023 response to a skeptical assessment of ChatGPT and Bing weighs early usage patterns, creative work and AI risks.
10:44ENMatthew Berman· 9 juil. 2023Can AI Help Humans Align More Powerful AI?Matthew Berman · Forward FutureMatthew Berman examines OpenAI’s 2023 Superalignment proposal, its plans for AI-assisted oversight, and the unanswered question of who evaluates the evaluators.
10:40ENMatthew Berman· 7 juin 2023Testing Nous Hermes 13B Locally: Mixed Reasoning and Unsafe ResponsesMatthew Berman · Forward FutureMatthew Berman compares a 2023 quantized model on hosted and local hardware, with uneven results across coding, logic and safety.
11:49ENMatthew Berman· 1 juin 2023Testing Wizard-Vicuna 30B: Responses, Reasoning, and CodeMatthew Berman · Forward FutureMatthew Berman examines a model marketed as uncensored, demonstrates a 2023 cloud setup, and runs an informal collection of tests.