I’m so excited to finally announce one of the amazing things I’ve been working on at Snowflake! Imagine what you could do with a single, unified platform for your transactional and analytical needs! Watch the keynote to see this in action today.
Concurrency and cost matter more than the speed of a single query. Real AI workloads are mixed: large scans and aggregations, point lookups, search, and transactional queries all running together. I'm sure that Snowflake leads, or is highly competitive, on concurrency, scalability, and is the easiest to use platform. That translates to less cost for customers at scale than solutions like ClickHouse. I'd bet on it.
"The AI bear case for @Snowflake revolves around differences in human vs. agent preferences for accessing data and the continued march of infrastructure that prices to one paradigm becoming obsolete as the world advances. In particular, while Snowflake's query engine works very well at human speeds (loading a dashboard, running a complex SQL query) upstarts like @ClickHouseDB ... argue that agents have very different preferences and prefer lightning fast queries that would be very expensive on Snowflake." 👏
The era of bespoke data pipelines is ending. 🛠️
VP of Product Chris Child joins Tristan Handy on The Analytics Engineering Podcast to discuss the Snowflake AI Data Cloud roadmap—from governing agents with Cortex to why we're betting on Apache Iceberg. 👉 https://t.co/4qypOxcyBz
Snowflake AI Research is releasing optimizations for limited resource fine-tuning of massive LLMs like Llama 3.1 405B making them more accessible to the broader AI community.
Our optimization stack can fine-tune Llama 3.1 405B over 20M tokens in just 4 hours with two 8 x H100-80GB nodes, and in roughly 24 hours on a single node.
Announcing: Snowflake Performance Index (SPI) to provide you with increased transparency and metrics around Snowflake’s ongoing performance improvements #SnowflakeSummit
Snowflake CIO Sunny Bedi shares how the #DataCloud company is powering employee experience and rationalizing costs with Snowflake Connector for @ServiceNow on the latest episode of #DataCloudNow: https://t.co/mvdJc3FnTW
On August 5th I left my home after I was swatted. I thought I would be safe at a hotel until I was doxxed. Things are escalating, and now I am in hiding. I need everyone to share this as widely as possible. I am scared of what is going to come next.
https://t.co/R0mtuHaAAD
@mim_djo@felipehoffa @iamontheinet There’s a talk that @macklu is giving on Wednesday from 1-1:45 where is will talk about this as part of our cost governance investments.
Get ready for a lot of cool ML announcements next week — in the meantime I called Tal Shaked to ask him about his thoughts on ML and life since joining @SnowflakeDB.
Btw, did you know that Tal was already doing Machine Learning at @Google back in 2004? https://t.co/4lLBGP9jXX ❄️
@cdixon Maybe a critical part of customers purchasing tokens it not to achieve significant returns on funding but instead to help their favorite company exist and continue to serve them.
@cdixon Really insightful thread. The question that I can’t answer is will the broader customer base purchasing tokens be more, less, or equally successful at picking or enabling winners than venture funding has been?
#PoweredBySnowflake is a brand new program announced today at the #SnowflakeSummit
Companies building consumer-facing apps on Snowflake can apply to join today
Check out some new interviews https://t.co/0aWI2GkgZ6
@SnowflakeDB