MLST is a media production company based in London focussing on producing science-based content. Our associated AI podcast is the top-rated globally on Spotify. Each week we have hard-hitting discussion with the leading thinkers in the AI space. Please get in touch if you are interested in sponsoring the show in some way, or need AI advisory work.
External link for Machine Learning Street Talk (MLST)
London, GB
Machine Learning Street Talk (MLST) reposted this
More #words, more fascinating conversation and yet more marvellously eloquent reflection on #language, #technology and - directly and/or indirectly, inevitably - of #humanity. What is missing here? #consciousness Thanks, Machine Learning Street Talk (MLST). Great conversation.
Aidan Gomez will free us all from tyranny of the word "delve" in language model outputs! This is why Cohere doesn't use any output from OpenAI for training their models. He also reveals how they're tackling AI hallucinations and improving reasoning this year. Interview up now on MLST!
Machine Learning Street Talk (MLST) reposted this
Speaking with Machine Learning Street Talk (MLST), Cohere Co-Founder Nick Frosst offers a fresh and practical take on evaluating Large Language Models. Nick emphasizes that while benchmarks provide a standardized way to measure LLM performance, they may not capture the specific needs and requirements of enterprises. Cohere focuses on creating models that excel in real-world business tasks, ensuring practical and tangible value for businesses.
We interview the current winners of the ARC Challenge—Jack Cole, Mohamed Osman Abdelghany and their collaborator Michael Hodel. They discuss how they tackled the ARC (Abstraction and Reasoning Corpus) Challenge using language models and neural networks. We also discuss the new "50%" approach announced today from Redwood Research. A big recurring theme is "where does the reasoning happen?" and if it's not efficient - the system ain't intelligent!
Cohere co-founder Nick Frosst gives an extremely refreshing take on how we shouldn't distract ourselves with AGI fantasies and get on with the job at hand, which is to say -- solving real business problems with LLMs. Bravo to that! "Transformers are nowhere close to the human mind" - Nick Frosst
Machine Learning Street Talk (MLST) reposted this
#UofTEngineering Engineering Science alumnus Aman Bhargava and Electrical & Computer Engineering at U of T MASc student Cameron Witkowski were recently featured on the Machine Learning Street Talk (MLST) podcast! They sat down with Tim Scarfe to discuss their work mapping out the “reachable space” of a language model using control theory. Listen now: uofteng.ca/1g6rhb
Machine Learning Street Talk (MLST) reposted this
Love to make strong partnerships with startups like XRAI Glass! They are on a mission to help people see, translate, and understand conversations in real time. Such a moving story this one ❤️ Check out this video where we caught up with CTO & Co-Founder, Tim Scarfe (also Founder of Machine Learning Street Talk (MLST)), and Chief Brand Ambassador, Jacqueline Press, to understand how they leverage AWS to provide real time transcription and translation. Also on the AWS Youtube channel https://lnkd.in/efiSgKhV Giuseppe Battista, Tricia Troth, Dan Scarfe, Lara Scarfe, Jack Bloomfield, Filming/editing credit goes to Rick Bronks Satureyes Media
Prof. Murray Shanahan on whether we can think of LLMs as being a form of "conscious exotica". See his paper "Simulacra as Conscious Exotica" https://lnkd.in/euCg2Udm
Machine Learning Street Talk (MLST) reposted this
Cohere hosted our first-ever Build Day in four cities across the globe! London, Toronto, San Francisco, and New York opened their doors to developers and business professionals eager to build with Cohere. The energy and excitement were electric as attendees rolled up their sleeves to bring their ideas to life. For those who couldn’t join us, stay updated on future events at Cohere.com/events to join us at the next one.
We had a fantastic time at the Cohere London Build Day with a huge group of developers hacking on their new Command R+ models! With Aidan Gomez & Anna Bialas