OpenAI cuts out the middleman; HPE gets Cray Cray
Today: OpenAI would rather ChatGPT users spend more time using its tool than other "copilots," HPE rolls out a new supercomputer design, and the quote of the week.
Today: why some of the biggest names in enterprise tech will back an open-source fork of Redis, Databricks jumps into the open LLM game, and the latest moves in enterprise tech.
Welcome to Runtime! Today: why some of the biggest names in enterprise tech will back an open-source fork of Redis, Databricks jumps into the open LLM game, and the latest moves in enterprise tech.
(Was this email forwarded to you? Sign up here to get Runtime each week.)
Last week after Redis became the latest enterprise tech company to change the licensing terms behind a popular open-source project that powered its growth, several developers quickly announced plans to create a fork that would live on as a permissively licensed project. This week, that group received a big vote of confidence from some deep-pocketed backers.
The Linux Foundation announced Thursday that it will host Valkey, "an open source alternative to the Redis in-memory, NoSQL data store," it said in a press release. Led by Madelyn Olson of AWS, Valkey ("like a Valkyrie") will keep the BSD 3-clause license that Redis (the company) plans to drop in favor of source-available licenses as it releases new versions of Redis (the project).
That's because AWS, Google Cloud, Oracle, Ericsson, and Snap announced that they will contribute time and money toward establishing Valkey as an open-source competitor to the Redis project.
The Linux Foundation also supported OpenTofu, a fork of HashiCorp's Terraform project released in response to its decision last year to restrict how others could use Terraform. But this feels different.
It will take months if not years for Valkey to make an impact, but a lot of startups, investors, and Big Cloud companies will be watching its progress.
Databricks has long been one of the behind-the-scenes engines of the generative AI boom, but this week it jumped directly into the fray. On Wednesday it released DBRX, an open large-language model that compares favorably with other open models like Meta's Llama as well as older closed models like OpenAI's GPT-3.5.
DBRX was built by the MosaicML team that Databricks acquired last year for $1.3 billion, and that team was very careful to describe DBRX as an "open model," given that Databricks will require companies with more than 700 million monthly users to purchase a license to use DBRX and prohibit anyone from using it to improve LLMs that compete with DBRX. However, Databricks' corporate blog, Wired's breathless and exclusive behind-the-scenes report, and countless other stories all described DBRX as "open source," which it is not thanks to those restrictions.
There's no doubt that models like DBRX and Llama are more open than the GPT family, but the AI community is trampling all over the established definition of "open source" in a rush to associate itself with the goodwill that the term still carries in the tech community. You'd hope companies developing predictive text engines would have a little more respect for precision in language, but here we are.
Matt Renner is the new president, global field organization for Google Cloud, taking over one of the top roles at the company less than a year after joining and following the departure of Adaire Fox-Martin to Equinix.
Steven Chung is the new president of Starburst, joining the data lakehouse company after serving as president of Delphix.
Tamar Yehoshua is the new president of product and technology at Glean, following four years as chief product officer at Slack.
Rick Scurfield is the new chief revenue officer at VAST Data, after spending 20 years in various sales leadership roles at NetApp.
Dennis Dayman is the new chief information security officer at Code42, joining the data security company after serving as CISO of Proofpoint.
Gurdeep Dhillon is the new chief marketing officer at Contentstack, following several years in a similar role at Zuora.
Amazon forked over the remaining $2.75 billion it was expected to invest in Anthrophic, after announcing plans to invest up to $4 billion in the foundation model company last year.
Something weird is happening at NIST, which has fallen way behind when it comes to updating a crucial software vulnerability database that security professionals rely upon to patch their systems.
Microsoft introduced new safety measures in Azure AI Studio to fight against prompt injections that can hijack LLMs to produce bizarre output or access unauthorized systems.
Walmart's tech infrastructure teams have been scrambling to deal with a rash of "major incidents" causing downtime that impacted revenue, according to Business Insider.
The University of Washington is also continuing to struggle with a years-long Workday implementation that has delayed the release of millions of dollars in research grants.
Thanks for reading — Runtime is off for the holiday weekend, see you Tuesday!