Results
97 articlesSort
37:01EN
No Priors: AI, Machine Learning, Tech, & Startups· Sep 28, 2023Podcast · 2 speakersDNA Is Code, but You Compile It with a PipetteHost 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.
46:37EN
20VC with Harry Stebbings· Sep 22, 2023Podcast · 2 speakersChristian Kleinerman on AI Models, Enterprise Data and Product DecisionsGuest Christian Kleinerman · Host Harry StebbingsSnowflake’s product leader discusses model choice, enterprise adoption and the trade-offs between speed, reliability and simplicity.
2:15:11EN
Cognitive Revolution "How AI Changes Everything"· Sep 22, 2023PodcastTwenty Gibberish Tokens Jailbroke ChatGPT and BardHost Nathan LabenzCarnegie Mellon’s Zico Kolter, Andy Zou, and Asher Trockman walk through a universal jailbreak that transfers from open models to commercial APIs, and a training-free way to steal some of pretraining’s benefit.
9:59EN
Matthew Berman· Sep 7, 2023Open Interpreter: A 2023 Local Code-Execution WalkthroughMatthew Berman · Forward FutureMatthew Berman installs Open Interpreter, analyzes local files with GPT-4 and tries a local Code Llama setup, with mixed results.
4:48ENMatthew Berman· Jul 12, 2023PoisonGPT: A Demonstration of Model Supply-Chain RiskMatthew Berman · Forward FutureMatthew Berman discusses Mithril Security’s 2023 demonstration of an edited GPT-J model and a misleading repository name. The experiment illustrates a supply-chain risk, not a breach of the original model provider.
54:00EN20VC with Harry Stebbings· Jun 30, 2023Podcast · 2 speakersDouwe Kiela on Data, Retrieval and Enterprise AIGuest Douwe Kiela · Host Harry StebbingsThe Contextual AI CEO discusses language model trade-offs, evaluation, open models and the hurdles to enterprise use.
51:22ENNo Priors: AI, Machine Learning, Tech, & Startups· Jun 29, 2023Podcast · 2 speakersThree IPOs Later, Frank Slootman Still Thinks the Job Is UrgencyHost Sarah Guo, Elad GilSnowflake's CEO tells No Priors how he picked the right elevator, turned a 15-person backup startup into an EMC acquisition, and why a data lake is a landfill you should not train on.
6:49ENMatthew Berman· Jun 21, 2023Reddit’s API Changes, AI Training Data and the Open InternetMatthew Berman · Forward FutureMatthew Berman argues that the 2023 Reddit API dispute exposed tensions between AI data demand, third-party developers and the people who create online content.
3:15:07EN
Lex Fridman· Jun 18, 2023Podcast · 2 speakersWikipedia Worked Once Writing Stopped Feeling Like Grad SchoolGuest Jimmy Wales · Host Lex FridmanJimmy Wales tells Lex Fridman how Nupedia's seven-stage review died, why Wikipedia still refuses ads and governments, and what large language models should actually do with an encyclopedia.
59:28ENCognitive Revolution "How AI Changes Everything"· Jun 16, 2023PodcastFrom $500,000 to $50,000: MosaicML Made Training Your Own Model a Priced ProductHost Nathan LabenzJonathan Frankle and Abhi Venigalla explain who is training custom LLMs, why extraction beats chat, and how ALiBi stretches a 65k context window.
2:00:19ENCognitive Revolution "How AI Changes Everything"· Jun 6, 2023Podcast1.5 Million GPT-4 Fairy Tales Let a 33-Million-Parameter Model Out-Reason GPT-2Host 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…
10:16ENMatthew Berman· May 21, 2023DarkBERT: Adapting a Language Model to Dark-Web TextMatthew Berman · Forward FutureMatthew Berman reviews research on a RoBERTa-based model and its cybersecurity evaluations.
1:13:52ENMatthew Berman· Apr 15, 2023Andriy Mulyar on GPT4All, Data Quality and Open Local ModelsMatthew Berman, Andriy Mulyar · Forward FutureThe Nomic AI co-founder explains how GPT4All emerged, why training data matters, and the tradeoffs between local and centralized AI.
8:02ENMatthew Berman· Apr 12, 2023Meta’s Segment Anything: Promptable Image Masks and the SA-1B DatasetMatthew Berman · Forward FutureIn April 2023, Matthew Berman reviews Meta’s Segment Anything project and demonstrates selecting image regions with clicks, boxes and automatic mask generation. The model segments pixels; it does not by itself name every object or guarantee accurate results on every image.
6:24ENMatthew Berman· Apr 6, 2023BloombergGPT: Mixing Financial and General Data for a Language ModelMatthew Berman · Forward FutureMatthew Berman examines the 2023 finance-focused research model, its training corpus, evaluation and decision not to release weights.
5:08ENMatthew Berman· Apr 3, 2023Chatting with a PDF in Colab Using LangChain and a Vector IndexMatthew Berman · Forward FutureMatthew Berman demonstrates a 2023 PDF question-answering notebook; its retrieval workflow does not fine-tune model weights and its revenue example is incorrect.
1:29:31ENCognitive Revolution "How AI Changes Everything"· Mar 30, 2023Podcast130,000 Teleop Demos: Google's GPT Playbook for RobotsHost Nathan LabenzKeerthana Gopalakrishnan of Google Robotics tells Nathan Labenz how RT-1, SayCan, and PaLM-E turn language-model scaling into kitchen-ready mobile manipulators.
1:03:29ZH → English
跟李沐学AI· Jan 12, 2023A Multimodal Paper Review: BLIP, CoCa, and BEiT-3From cleaning image–text data to unified pretraining, a guide to the architectures, objectives, and connections behind three vision–language models.
2:31:48ENLex Fridman· Dec 28, 2021Podcast · 2 speakersA Car That Cannot Drive Itself Will Be About as Useful as a HorseGuest Elon Musk · Host Lex FridmanIn a Lex Fridman conversation built around Tesla Autopilot, Elon Musk argues that every driver input is an error, that today’s hardware is already enough for full self-driving, and that watching the human will soon matter less than making the system safer than a person.
1:28:46ENLex Fridman· Jan 4, 2021Podcast · 2 speakersThe Speech Company That Gets Paid to Label Its Own Training DataGuest Dan Kokotov · Host Lex FridmanRev VP of engineering Dan Kokotov tells Lex Fridman why a $1.25-a-minute human transcript, a 14% machine word-error rate, and a two-sided freelancer market are the same flywheel.
1:12:03ENLex Fridman· May 5, 2020Podcast · 2 speakersMice Don't Get Alzheimer's: Daphne Koller on Disease in a DishGuest Daphne Koller · Host Lex FridmanThe Coursera co-founder and insitro CEO on why most diseases are not one disease, how to turn cells into data, and why overconfident neural nets can kill.
1:17:29ENLex Fridman· Sep 23, 2019Podcast · 2 speakersThe Algorithm Is Already Better. The Clinic Still Uses 1967.Guest Regina Barzilay · Host Lex FridmanMIT's Regina Barzilay on why deep learning already beats the breast-cancer 'density' score — and why data, liability, and incentives still will not move.
57:44ENLex Fridman· Aug 1, 2019Podcast · 2 speakersKevin Scott: Treat AI Like a Hash Table, and Measure Platforms by What Others EarnGuest Kevin Scott · Host Lex FridmanMicrosoft's CTO talks radical markets, data dignity, Dunbar's number, and why a platform fails if only a handful of companies decide what AI is for.
32:45ENLex Fridman· Apr 12, 2019Podcast · 2 speakersElon Musk: Every Driver Input Is an Error Tesla Autopilot Must EliminateGuest Elon Musk · Host Lex FridmanAfter MIT published a study of functional vigilance on Autopilot, Lex Fridman sat down with Musk to argue about cameras on drivers, fleet data, and whether humans still belong in the loop.Load moreShowing 24 of 97