Thank you for having me to share our story past 6 yrs. #CrossTheChasm is the most important for a new tech product.

Over the past year, collaborating with multiple pre-trained LLM teams has been rewarding and insightful.

It seems we're on the right path. Yet, a new era of technology has just begun, and who knows what the future holds? The release of the Transformer paper is akin to Google's…

#LLMTraining in #MultiCloud , performance at scale, cost effeciency

help users of BentoML - AI Inference Platform in your cloud deploy large models faster.🚀 Thanks yetone for sharing his practice.

Helpful. Chrome is back to my default browser. Elmo is free

#JuiceFS design for this ;) pinecone.io/blog/the-new-a… QUOTE: Cloud centrality is becoming a key success factor. Negotiating between data gravity and model gravity requires new cloud native services that are very dynamic and optimized in their use of compute, storage, and…

If you use Colaboratory and dev env togethher, juicefs is a easy way to share your data. Run Chroma on the juicefs is a nice try

AI robot will sit on my seat

youtube.com/watch?v=lXLBTB… at 5:20, after the worst elevator pitch, the ex-boss introduced first money to Huang who is #NVIDIA CEO. The thing he learned you can make up a great interview, or even a bad interview, but you can't run away from your past, so you have a good past, so try…

Advancing networks have unlocked the potential of large-scale GPU clusters, and storage systems inevitably evolve in tandem. Yet, a balance must be struck between performance, scale, reliability, cost, maintainability, and ecosystem. Will the mythical 'six-sided warrior' emerge,…

in past 15yrs, public clouds support web & api services by its IaaS (specific CPU visualization) and PaaS (like DB, MQ, cache). In AI era, inference and training service is a new platform, advantage of clouds are not competitive. NEW AI CLOUD is coming, Lepton AI BentoML - Run Any AI Models in the Cloud…

performance at scale in #LLM training

98% GPU Utilization Achieved in 1K GPU-Scale AI Training Using Distributed Cache #MLPerf dzone.com/articles/98-gp…

BentoML: Reducing LLM Loading Time from 20+ to a Few Minutes with a Distributed File System by yetone shared in Mandarin, and convert into English by HeyGen
youtube.com/watch?v=bh31rh…

现在浏览器的 AI 助手插件五花八门，但是都会在网页上插入一堆按钮

过年的时候给 Lepton 做了个内部工具，一个老实待在侧边栏的插件，点一下闪电 logo 就会总结当前页面

公司内大家玩的还是挺开心的，于是最近我们把它放到了 Chrome 应用商店，欢迎大家来玩（不收费也不用注册）

下载地址在下一条

Hi, we've just launched our product on Product Hunt, Aili aims to help our users with a better AI native bookmark solution which boost your reading productivity, and yet with a cost-effective subscription plan. #producthunt #launchday

producthunt.com/posts/aili-swi…

#NAVER , top1 search engine of Korea, he selects JuiceFS as the storage engine in his AI platform