Résultats
347 articlesTrier
2:30:06EN
Cognitive Revolution "How AI Changes Everything"· 22 déc. 2023PodcastTrained on a Thirtieth of GPT-4's Tokens, Mamba Beat Transformers on Home TurfAnime : Nathan LabenzIn a solo emergency episode, Nathan Labenz argues that Albert Gu and Tri Dao's selective state space model is the first architecture since 2017 that deserves to share the Transformer’s throne.
14:04EN
Matthew 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.
20:49ENMatthew Berman· 12 déc. 2023Mixtral 8x7B: Mixture-of-Experts Architecture, Setup and Early TestsMatthew Berman · Forward FutureMatthew Berman reviews Mistral AI’s December 2023 open-weight model, demonstrates a cloud setup and tests its coding and reasoning responses.
1:19:24ENCognitive Revolution "How AI Changes Everything"· 17 nov. 2023PodcastA Toy Paper Did Not Prove LLMs Cannot Generalize. The Discourse Still Ran With It.Anime : Nathan LabenzNathan Labenz and Erik Torenberg unpack a viral DeepMind result, a weekend reproduction, and a premature war over voluntary AI commitments.
1:21:50ENCognitive Revolution "How AI Changes Everything"· 14 nov. 2023PodcastA 16K Context Window Holds About 350 Rows. That's Why Tables Need an Agent.Anime : Nathan LabenzGretel AI founder Alex Watson tells Nathan Labenz why most company data is still a spreadsheet, why GPT-4 is the wrong tool for generating it, and how differential privacy lets a model learn a disease without learning a patient.
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.
41:58EN
No Priors: AI, Machine Learning, Tech, & Startups· 2 nov. 2023Podcast · 2 voixReliability is the bottleneck. Autonomy is what would make AI feel alive.Anime : Sarah Guo, Elad GilIlya Sutskever talks with Sarah Guo and Elad Gil about AlexNet, OpenAI’s capped-profit bet, why smaller models mainly lose reliability, and the line he draws for digital life.
1:19:52ENCognitive Revolution "How AI Changes Everything"· 18 oct. 2023Podcast1,000 H100s Take 115 Days; 22,000 Take FiveAnime : Nathan LabenzNathan Labenz and Trey Kollmer walk through cluster math that turns GPT-4-scale training into a weekly experiment, then four papers on search-augmented models, recursive code improvers, spatial maps inside Llama 2, and what networks do when they have no idea.
13:11ENMatthew Berman· 5 oct. 2023Testing Mistral 7B OpenOrca: A 2023 Cloud Setup and Prompt TrialMatthew Berman · Forward FutureMatthew Berman runs a quantized OpenOrca variant on a rented GPU and reviews coding, writing and reasoning examples.
37:01ENNo Priors: AI, Machine Learning, Tech, & Startups· 28 sept. 2023Podcast · 2 voixDNA Is Code, but You Compile It with a PipetteAnime : Sarah Guo, Elad GilGinkgo Bioworks CEO Jason Kelly tells No Priors why cells are worth the magic, why biology has no AWS, and why a pandemic should look like cybersecurity.
12:49ENMatthew Berman· 14 sept. 2023Falcon 180B: A 2023 Review of TII’s Open-Weight ModelMatthew Berman · Forward FutureMatthew Berman tests coding, reasoning and writing prompts while examining the model’s training scale and license.
4:35ENMatthew Berman· 5 sept. 2023Running WizardCoder on a Cloud GPU: A 2023 RunPod WalkthroughMatthew Berman · Forward FutureMatthew Berman demonstrates downloading a quantized 13B coding model into a hosted text-generation interface, then explains how to stop the pod and avoid ongoing charges.
36:33EN
20VC with Harry Stebbings· 31 août 2023Podcast · 2 voixNoam Shazeer on Character.AI, the Cost of Training Models and the Uses of General AIInvité : Noam Shazeer · Anime : Harry StebbingsThe Character.AI co-founder explains his consumer strategy, language-model training and why he expects users to invent applications the company cannot predict.
55:08ENCognitive Revolution "How AI Changes Everything"· 29 août 2023PodcastOne Model Inside 38 Products: Paige Bailey on How Google Trains and Ships PaLM 2Anime : Nathan LabenzAt Google I/O in May 2023, PaLM 2 arrived already running in more than two dozen products — and its lead product manager walks through what that actually took.
6:28ENMatthew Berman· 29 août 2023A Historical GPT-3.5 Fine-Tuning WalkthroughMatthew Berman · Forward FutureMatthew Berman demonstrates generating synthetic examples in Colab and training a model to use a distinctive tone.
52:48ENCognitive Revolution "How AI Changes Everything"· 25 août 2023PodcastOpenAI Cut Fine-Tuning 90 Percent, and Founders Crossed Llama 2 Off the ListAnime : Nathan LabenzNathan Labenz and Erik Torenberg walk through GPT-3.5 Turbo fine-tuning, why self-hosting still loses on total cost, and what is actually working in code, school, and medicine.
35:24ENNo Priors: AI, Machine Learning, Tech, & Startups· 24 août 2023Podcast · 2 voixThe Author of Attention Is All You Need Now Compiles mRNA Like BytecodeAnime : Sarah Guo, Elad GilJakob Uszkoreit on why Transformers won, why models spend compute on the wrong things, and why drug discovery has to become generation rather than screening.
7:09ENMatthew Berman· 28 juil. 2023AI News in July 2023: SDXL, Worldcoin, and Biological ComputingMatthew Berman · Forward FutureMatthew Berman reviews a week of AI developments, from Google’s Gemini project and new generative models to biometric identity and experiments with cultured neurons.
2:13:07ENCognitive Revolution "How AI Changes Everything"· 27 juil. 2023PodcastHalf an Hour to a Circuit in GPT-2 Small — and a Quadratic Wall Before GPT-3Anime : Nathan LabenzArthur Conmy walks Nathan Labenz through ACDC, the algorithm that automates the most tedious step of mechanistic interpretability: finding which transformer components actually do a task.
4:59ENMatthew Berman· 25 juil. 2023Testing Airoboros L2 70B on Coding, Writing and ReasoningMatthew Berman · Forward FutureMatthew Berman runs a small set of prompts against a Llama 2 fine-tune, finding useful structured output alongside failures on code and logic.
9:40ENMatthew Berman· 20 juil. 2023Testing Llama 2 70B on Coding, Writing and ReasoningMatthew Berman · Forward FutureMatthew Berman tries Meta’s 70-billion-parameter chat model through a hosted interface, highlighting both useful outputs and inconsistent reasoning.
5:53ENMatthew Berman· 18 juil. 2023Petals: Distributing Large Language Model Inference Across DevicesMatthew Berman · Forward FutureIn 2023, Matthew Berman explains Petals as a way to use large models through a network of participating computers rather than loading every layer on one device.
8:04ENMatthew Berman· 11 juil. 2023WizardLM 13B v1.1 Tested on Coding, Writing and LogicMatthew Berman · Forward FutureMatthew Berman tries an instruction-tuned model on a small set of prompts, with mixed results and several subjective scoring choices.
12:20ENMatthew Berman· 10 juil. 2023Testing UltraLM 13B: A Small Model’s Strengths and Reasoning ErrorsMatthew Berman · Forward FutureA 2023 local test finds useful writing and simple coding alongside failures in games, arithmetic and logic.Charger plus24 sur 347 affichés