Résultats
76 articlesTrier
10:57EN
Matthew Berman· 13 sept. 2024Testing OpenAI o1-preview: Tetris, Physical Reasoning and a North Pole ErrorMatthew Berman · Forward FutureIn September 2024, Matthew Berman runs the new reasoning-model preview through a personal rubric and examines where its answers succeed or fail.
21:21ENMatthew Berman· 12 sept. 2024OpenAI o1’s 2024 Launch: Reasoning Benchmarks and Early TestsMatthew Berman · Forward FutureMatthew Berman reviews o1-preview and o1-mini, OpenAI’s reported evaluations, and his own early coding experiments.
51:15EN
Sequoia Capital· 10 sept. 2024PodcastAGI Has One Problem Left: Spend a Million Times More on a TokenMagic founder Eric Steinberger tells Sequoia’s Training Data why he trains his own models, why a software engineer is the MVP of AGI, and why 96% still is not a colleague.
12:03ENMatthew Berman· 6 sept. 2024Reflection 70B: Self-Correction Claims and the Limits of a 2024 DemoMatthew Berman · Forward FutureMatthew Berman reviews a model trained to generate reflection steps. Its initial benchmark claims later proved difficult for independent evaluators to reproduce.
12:17ENMatthew Berman· 7 août 2024Testing Gemini 1.5 Pro on Code, Reasoning, Images and VideoMatthew Berman · Forward FutureMatthew Berman compares standard and experimental Gemini versions in an informal 2024 review, finding mixed results and repeated output errors.
14:15ENMatthew Berman· 16 juil. 2024OpenAI’s Reported Strawberry Project and a Five-Level AGI FrameworkMatthew Berman · Forward FutureMatthew Berman examines 2024 reporting on a reasoning project, long-horizon tasks and a proposed scale for AI capabilities.
1:56:03EN
Cognitive Revolution "How AI Changes Everything"· 13 juil. 2024PodcastHumans Near a Perfect Score. The Best AI Still Gets 39%Anime : Nathan LabenzZapier co-founder Mike Knoop and François Chollet put $1 million on ARC-AGI, a 2019 test of sample-efficient intelligence that language models still cannot crack under a one-cent-per-puzzle budget.
55:13ENSequoia Capital· 2 juil. 2024PodcastThe Best AI Scores 39% on This Test. Zapier’s Co-Founder Put $1 Million on Beating ItFrançois Chollet designed ARC to measure skill-acquisition efficiency; five years later humans still crush it and frontier models do not.
17:03ENMatthew Berman· 8 juin 2024Qwen2 0.5B and 72B: A Hands-On ComparisonMatthew Berman · Forward FutureMatthew Berman tests coding, arithmetic and reasoning prompts on the smallest and largest Qwen2 models in a 2024 demonstration.
1:03:12EN
20VC with Harry Stebbings· 5 juin 2024Podcast · 2 voixAravind Srinivas on Reasoning, Model Economics and Building PerplexityInvité : Aravind Srinivas · Anime : Harry StebbingsThe Perplexity co-founder discusses AI scaling, foundation-model competition, advertising and the challenges of building a search product.
17:12ENMatthew Berman· 30 avr. 2024Testing the Mystery 'gpt2-chatbot' Model in April 2024Matthew Berman · Forward FutureMatthew Berman tries an anonymous model on coding, arithmetic and reasoning tasks while its identity remains uncertain.
22:17ENMatthew Berman· 21 mars 2024Q* Rumors, Energy-Based Models, and Research on ReasoningMatthew Berman, Yann Lecun · Forward FutureMatthew Berman examines an anonymous, unverified account of Q* and compares its ideas with published research on teaching language models to reason.
14:04ENMatthew Berman· 15 déc. 2023Mistral Medium in 2023: API Pricing and Informal Model TestsMatthew Berman · Forward FutureMatthew Berman compares the original Mistral Medium prototype with Mixtral and GPT-4, then tests coding, writing and reasoning through the API.
1:22:59ENCognitive Revolution "How AI Changes Everything"· 2 déc. 2023PodcastIf Your Agent Is Right 80% of the Time, Ten Steps Leave You With 10%Anime : Nathan LabenzImbue CTO Josh Albrecht tells Nathan Labenz how a 27-person, pre-product research company is putting $200 million and 10,000 H100s toward agents that can actually finish a long job.
11:05ENMatthew Berman· 7 nov. 2023Testing Dolphin 2.2, a Mistral 7B Fine-TuneMatthew Berman · Forward FutureMatthew Berman tries a small open-weight model on coding, formatting and reasoning prompts, with mixed results.
9:46ENMatthew Berman· 17 juil. 2023Orca Mini V2 vs. OpenOrca: A 2023 Informal Model ComparisonMatthew Berman · Forward FutureMatthew Berman tests two Orca-inspired models on coding, writing and reasoning, with results shaped by prompt setup and subjective scoring.
2:00:19ENCognitive Revolution "How AI Changes Everything"· 6 juin 2023Podcast1.5 Million GPT-4 Fairy Tales Let a 33-Million-Parameter Model Out-Reason GPT-2Anime : Nathan LabenzTry this fill-in: Lily likes cats and dogs. She asked her mom for a dog. Mom said no, so she asked for a…
9:58ENMatthew Berman· 26 avr. 2023HuggingChat’s 2023 Debut: Testing Writing, Reasoning and Python CodeMatthew Berman · Forward FutureMatthew Berman tries the early OpenAssistant-powered HuggingChat on writing, arithmetic, logic and a Snake game, then compares a few outputs with GPT-3.5 and GPT-4. The results illustrate both promise and limitations in this small historical test.
1:00:36EN
No Priors: AI, Machine Learning, Tech, & Startups· 2 févr. 2023Podcast · 2 voixForty games of Diplomacy, and nobody spotted the bot: Noam Brown’s 100,000x lessonAnime : Sarah Guo, Elad GilMeta’s Noam Brown tells Sarah Guo and Elad Gil how Cicero negotiated in English for 40 games on a tiny web community — and why thinking at decision time beat three years of scaling.
1:26:21EN
Lex Fridman· 9 déc. 2018Podcast · 2 voixThe Machine Should Never Treat Its Objective as GospelInvité : Stuart Russell · Anime : Lex FridmanStuart Russell talks with Lex Fridman about metareasoning, self-driving reliability, and why uncertainty about human goals is the core of the control problem.
1:45:49ENCognitive Revolution "How AI Changes Everything"PodcastTrust at Scale: Elicit's Founders on Reasoning Languages, World Models and Automating the CompanyAnime : Nathan LabenzAndreas Stuhlmüller and Jungwon Byun explain how Elicit keeps frontier models checkable for high-stakes science, and where they think AI reasoning is heading.
15:01ENLex FridmanPodcast · 2 voixDeep Learning Is a System 1 Machine: Daniel Kahneman on What AI Still LacksInvité : Daniel Kahneman · Anime : Lex FridmanThe Nobel-winning psychologist tells Lex Fridman how his fast-and-slow thinking framework maps onto deep learning, and why driving among pedestrians may require understanding people.
10:16ENLex FridmanPodcast · 2 voixReasoning Without Logic: Yann LeCun on Teaching Neural Networks to ThinkInvité : Yann Lecun · Anime : Lex FridmanYann LeCun explains why he believes neural networks can reason, and what memory, recurrence and continuous representations would need to look like.
58:57ENCognitive Revolution "How AI Changes Everything"PodcastWhen the Model Learns to Look Around: Apollo Research on Testing o1 for SchemingAnime : Nathan LabenzApollo Research founder Marius Hobbhahn explains how his team probed OpenAI's o1 for deception, what it found, and why the trajectory worries him more than the model itself.Charger plus24 sur 76 affichés