Discovered not Designed: Hidden Capabilities of Large Language Models with Dolly

Register here: https://bit.ly/4193uDw

We are only now discovering the immense power, latent capability, and emergent behaviors of Large Language Models (LLMs) and Generative AI. If software is eating the world, then AI is eating software itself and LLMs are only accelerating the revolution. In a matter of days and weeks, all the rules of the technology industry that we take for granted have become uncertain, and maybe even dangerously wrong.

Foucault said that a singularity is “when things are no longer perceived, described, expressed, characterized, classified, and known in the same way." By that definition, like it or not, we are in a singularity. Perhaps not in the Kurzweilian sense, but certainly in the way Foucault meant it: a discontinuity, beyond which things are unknown or unknowable.

Join The Hive Think Tank, Mike Conover and Sam Shah from Databricks for a wide-ranging conversation moderated by Alistair Croll, which will be about philosophy as much as it is about technology. They'll explore metaphors for understanding this transformative time in software, consider the impact of human augmentation, and discuss the moral obligation to democratize and govern this unprecedented technological breakthrough. Of course, they’ll also cover Databricks’ recent Dolly innovations and why Databricks open sourced the Dolly 2.0 model, training code, dataset, and weights so any organization can create, own, and customize powerful LLMs that can talk to people.

Register on Zoom using the [bit.ly](https://us02web.zoom.us/webinar/register/WN_EXMKnVZQRKekEgb-dbjL3Q#/registration) link!

About the Speakers:
Mike Conover is the co-creator of Dolly and works on Applied AI at Databricks. Mike's work has been featured in the New York Times, the Wall Street Journal, and on NPR.

Sam Shah is VP of Engineering at Databricks and head of the company’s data strategy. Dr. Shah has a diverse set of experience as a data executive and founder of an AI startup.

Alistair Croll (moderator) has also launched and chaired some of the world’s leading conferences on emerging technology, including Startupfest, Strata, Cloud Connect, FWD50, Bitnorth, Scaletech, and more. Alistair is the author of four books on technology and entrepreneurship, including the best-selling Lean Analytics, which has been translated into eight languages and is in its tenth printing in China. He speaks internationally on topics such as data science, innovation, scaling startups, digital government, AI, and applying critical thinking to technology.

Dolly 2.0 is the first open source, instruction-following LLM, fine-tuned on a human-generated instruction dataset licensed for research and commercial use. Dolly 2.0 is a 12B parameter language model based on the EleutherAI pythia model family and fine-tuned exclusively on a new, high-quality, human-generated, instruction-following dataset, crowdsourced among Databricks employees. It comes on the heels of Databricks’ initial Dolly discovery, showing that anyone can take a dated off-the-shelf open source LLM and give it magical ChatGPT-like instruction following ability by training it in 30 minutes on one machine, using high-quality training data. Databricks believes that models like Dolly will help democratize LLMs, transforming them from something very few companies can afford into a commodity every company can own and customize to improve their products.