The Future of the Transformer Part 2 with Trey Kollmer

Published: Oct. 20, 2023, 7 a.m.

Trey Kollmer returns to discuss the latest AI research revelations with Nathan Labenz. They explore how new techniques will shave 10% off global compute needs, how analogical prompting beats few-shot prompting, and how compressive historical records can increase LLM memory and retention abilities. If you need an ERP platform, check out our sponsor NetSuite: http://netsuite.com/cognitive.\n\nSPONSORS: Shopify | NetSuite | Omneky\nShopify is the global commerce platform that helps you sell at every stage of your business. Shopify powers 10% of ALL eCommerce in the US. And Shopify's the global force behind Allbirds, Rothy's, and Brooklinen, and 1,000,000s of other entrepreneurs across 175 countries.From their all-in-one e-commerce platform, to their in-person POS system \u2013 wherever and whatever you're selling, Shopify's got you covered. With free Shopify Magic, sell more with less effort by whipping up captivating content that converts \u2013 from blog posts to product descriptions using AI. Sign up for $1/month trial period: https://shopify.com/cognitive\n\nNetSuite has 25 years of providing financial software for all your business needs. More than 36,000 businesses have already upgraded to NetSuite by Oracle, gaining visibility and control over their financials, inventory, HR, eCommerce, and more. If you're looking for an ERP platform \u2705 head to NetSuite: http://netsuite.com/cognitive and download your own customized KPI checklist.\n\nOmneky is an omnichannel creative generation platform that lets you launch hundreds of thousands of ad iterations that actually work customized across all platforms, with a click of a button. Omneky combines generative AI and real-time advertising data. Mention "Cog Rev" for 10% off.\n\nRECOMMENDED PODCAST:\xa0\nEvery week investor and writer of the popular newsletter The Diff, Byrne Hobart, and co-host Erik Torenberg discuss today\u2019s major inflection points in technology, business, and markets \u2013 and help listeners build a diversified portfolio of trends and ideas for the future. Subscribe to \u201cThe Riff\u201d with Byrne Hobart and Erik Torenberg: https://link.chtbl.com/theriff\n\n\nTIMESTAMPS:\n(00:00:00) - Episode Preview\n(00:01:11) - Paper: Think Before You Speak\n(00:03:13) - Multimodal models for combining vision and language\n(00:04:19) - Backspace Paper\n(00:06:25) - Chain of thought prompting for step-by-step reasoning\n(00:09:14) - Backspacing in language models to correct mistakes\n(00:12:05) - Attention sinks for expanding context length\n(0012:41) - Paper: Large Language Models as Analogical Reasoners\n(00:15:24) - Pause tokens for language models to "think"\n(00:18:23) - Analogical prompting to recall relevant examples\n(00:20:52) - Long context windows for language models\n(00:23:20) - Markdown works best for OpenAI\n(00:24:23) - Ring attention to break memory constraints\n(00:26:15) - Paper: StreamingLLMs\n(00:27:46) - Potential for superhuman performance with longer contexts\n(00:31:01) - Dynamic context window adjustment at runtime\n(00:33:53) - Retention and memory capabilities for transformers\n(00:37:12) - Planning algorithms combined with memory and scale\n(00:39:49) - Paper: Ring Attention\n(00:42:35) - Executive assistant prompting and critique\n(00:45:23) - Self-RAG for language models to find own examples\n(00:48:02) - Timelines and predictions for future capabilities\n(00:50:37) - Applications like analyzing long texts and scripts\n(00:53:15) - Local versus global attention in transformers\n(00:55:59) - Architectural changes versus just training adjustments\n(00:58:41) - Pre-training strategies like random start points\n\nThis show is produced by Turpentine: a network of podcasts, newsletters, and more, covering technology, business, and culture \u2014 all from the perspective of industry insiders and experts. We\u2019re launching new shows every week, and we\u2019re looking for industry-leading sponsors \u2014 if you think that might be you and your company, email us at erik@turpentine.co.