This evening's hackernews contribution: https://hai.stanford.edu/news/ais-ostensible-emergent-abilities-are-mirage
AI’s Ostensible Emergent Abilities Are a Mirage According to Stanford researchers, large language models are not greater than the sum of their parts. Which is a gloss on https://arxiv.org/abs/2304.15004: Recent work claims that large language models display emergent abilities, > abilities not present in smaller-scale models that are present in > larger-scale models. What makes emergent abilities intriguing is two-fold: > their sharpness, transitioning seemingly instantaneously from not present > to present, and their unpredictability, appearing at seemingly > unforeseeable model scales. Here, we present an alternative explanation for > emergent abilities: that for a particular task and model family, when > analyzing fixed model outputs, one can choose a metric which leads to the > inference of an emergent ability or another metric which does not. Thus, > our alternative suggests that existing claims of emergent abilities are > creations of the researcher's analyses, not fundamental changes in model > behavior on specific tasks with scale. -- rec -- On Mon, May 8, 2023 at 12:33 AM Pieter Steenekamp < piet...@randcontrols.co.za> wrote: > Sorry, I forgot to add the url of the interview. > (580) The Current State of Artificial Intelligence with James Wang - > YouTube <https://www.youtube.com/watch?v=V6WL4X6pmCY> > > > On Mon, 8 May 2023 at 08:27, Pieter Steenekamp <piet...@randcontrols.co.za> > wrote: > >> I am very excited about the basic idea that neither Google nor any Big >> Tech company has the Moat as per the hackernews reference above. >> Very interesting around this is the interview with James Wang of Cerebras >> James Wang about this where he makes a strong case (in my view in any case) >> that in future open source large language models are going to be much more >> prominent than those of Big Tech. >> >> I quote from the description on Toutube about the interview: >> "Scaling laws are as important to artificial intelligence (AI) as the law >> of gravity is in the world around us. AI is the empirical science of this >> decade, and Cerebras is a company dedicated to turning state-of-the-art >> research on large language models (LLMs) into open-source data that can be >> reproduced by developers across the world. In this episode, James Wang, an >> ARK alum and product marketing specialist at Cerebras, joins us for a >> discussion centered around the past and the future of LLM development and >> why the generative pre-trained transformer (GPT) innovation taking place in >> this field is like nothing that has ever come before it (and has seemingly >> limitless possibilities). He also explains the motivation behind Cerebras’ >> unique approach and the benefits that their architecture and models are >> providing to developers." >> >> On Mon, 8 May 2023 at 01:09, Merle Lefkoff <merlelefk...@gmail.com> >> wrote: >> >>> Thank you Roger. >>> -. --- - / ...- .- .-.. .. -.. / -- --- .-. ... . / -.-. --- -.. . >>> FRIAM Applied Complexity Group listserv >>> Fridays 9a-12p Friday St. Johns Cafe / Thursdays 9a-12p Zoom >>> https://bit.ly/virtualfriam >>> to (un)subscribe http://redfish.com/mailman/listinfo/friam_redfish.com >>> FRIAM-COMIC http://friam-comic.blogspot.com/ >>> archives: 5/2017 thru present >>> https://redfish.com/pipermail/friam_redfish.com/ >>> 1/2003 thru 6/2021 http://friam.383.s1.nabble.com/ >>> >> -. --- - / ...- .- .-.. .. -.. / -- --- .-. ... . / -.-. --- -.. . > FRIAM Applied Complexity Group listserv > Fridays 9a-12p Friday St. Johns Cafe / Thursdays 9a-12p Zoom > https://bit.ly/virtualfriam > to (un)subscribe http://redfish.com/mailman/listinfo/friam_redfish.com > FRIAM-COMIC http://friam-comic.blogspot.com/ > archives: 5/2017 thru present > https://redfish.com/pipermail/friam_redfish.com/ > 1/2003 thru 6/2021 http://friam.383.s1.nabble.com/ >
-. --- - / ...- .- .-.. .. -.. / -- --- .-. ... . / -.-. --- -.. . FRIAM Applied Complexity Group listserv Fridays 9a-12p Friday St. Johns Cafe / Thursdays 9a-12p Zoom https://bit.ly/virtualfriam to (un)subscribe http://redfish.com/mailman/listinfo/friam_redfish.com FRIAM-COMIC http://friam-comic.blogspot.com/ archives: 5/2017 thru present https://redfish.com/pipermail/friam_redfish.com/ 1/2003 thru 6/2021 http://friam.383.s1.nabble.com/
