Google flexes its Arm; Metronome runs the meter

Welcome to Runtime! Today: Google's first custom Arm server processor is now available, Metronome's new tool could help SaaS companies switch to usage-based pricing, and the quote of the week.

(Was this email forwarded to you? Sign up here to get Runtime each week.)

Ship it

Turning on Axion: While Google has made custom AI processors for years, it lagged behind rivals AWS and Microsoft when it came to getting on board with Arm server processors. It introduced the Axion line of custom server CPUs earlier this year at Google Cloud Next, and instances running those chips are now available to the public.

The C4A instances will deliver "up to 65% better price-performance and up to 60% better energy-efficiency than comparable current-generation x86-based instances," according to a blog post. Three different options are available depending on whether your workload is processor-intensive or memory-intensive.

Amazon's Q branch: Coding assistants continue to look like one of the most likely breakout stars of the generative AI era, but GitHub and several startups working on the technology have attracted the most attention. This week AWS rolled out new extensions to Visual Studio Code and JetBeans that will allow users of those development environments to interact with its Amazon Q Developer tool.

"With this capability, you can select a section of code that you need assistance with and initiate chat within the editor to request actions such as 'Optimize this code,' 'Add comments,' or 'Write tests,'" AWS said in its announcement. If you wanted to use Amazon Q Developer for chat help before, the output had to live in a separate window, but the new extension allows developers to ask questions and get answers right alongside the code they're developing.

Usage, not seats: SaaS pricing used to be simple: Every person that uses the software needs a license. But with cost cutting top of mind inside many companies, a lot of customers are starting to wonder why they need to pay for a SaaS seat if an employee only needs to use the product every so often.

As a result, usage-based pricing — which is how cloud infrastructure spending works, more or less — is becoming much more common among SaaS vendors. Metronome makes billing software for those companies, and this week it introduced the second version of its flagship product, which provides additional data sources to figure out how to charge those customers and allows users to change pricing more quickly.

No code, no problem: One of the more interesting possibilities presented by generative AI coding assistants is that businesses might find it easier to build the software they need rather than buy it, although a lot of things will need to fall into place to make that a realistic scenario. Still, it's clear generative AI is a huge boost for the low-code/no-code startups that have been looking for a breakthrough for years.

This week Creatio, which raised $200 million earlier this year to take on SaaS stalwarts like Salesforce, released the latest version of its no-code platform, adding support for several different types of AI-enabled workflows. "The “Energy” release can be used to build packaged capabilities called “skills” that perform common workflow and customer relationship management tasks involving decisions and autonomous actions," according to SiliconAngle.

Take this job and shove it: Anyone who has ever been a hiring manager or a job candidate knows that there are a lot of things about the job-seeking process that are out of whack right now, as our recent story on hiring in the AI era made clear. An awful lot of searches for both jobs and candidates start on LinkedIn, and this week the company jumped on the agentic AI bandwagon with its first agent designed for recruiting.

"Hirers will be able to upload their job descriptions, intake notes, and job postings into Hiring Assistant and it will immediately translate that information into role qualifications and build a pipeline of qualified candidates," LinkedIn said in a blog post. Hiring Assistant will also supposedly learn from the feedback it is given on the candidates it produces, which sounds like a lawsuit waiting to happen.

Stat of the week

Python is now the most widely used programming language on GitHub, ending Javascript's decade-long run in the top spot according to GitHub's Octoverse 2024 report. There's no doubt the surge in AI activity over the last year played some role in that shift, but Ben Lesh's viewpoint also seems valid: "Python developers congratulating themselves that the JavaScript community is split between TypeScript and JavaScript is the most data science facepalm I can think of."

Quote of the week

"Switching hypervisors isn't something to take lightly. It will require lots of stability testing and proof of concepts prior to changing." — James Westendorf, director of technical services at Land Lake College in Illinois, speaking to Ars Technica about the pickle that VMware customers upset about price increases find themselves in one year after Broadcom acquired the company.

The Runtime roundup

Nvidia will replace Intel in the Dow Jones Industrial Average, perhaps the clearest signal of two companies on completely opposite trajectories.

Microsoft has agreed to spend $10 billion on infrastructure capacity from CoreWeave by the end of the decade as it scrambles for GPUs, according to The Information.

Atlassian beat analyst expectations and raised its guidance for the full year, sending its stock up nearly 19% on Friday.

Thanks for reading — see you Tuesday!