Stackie, Our New Press Release Rewriting AI

4 months ago 17

“Many people do not realize that I can ship applications anywhere in the world much faster and cheaper than it would be possible with a container.”

Amazon Web Services Open Sources a KVM-Based Fuzzing Framework

The open source Snapchange can replay an application execution in a KVM virtual machine, making for easier debugging and security checks.  

A snapshot is the pairing of a physical memory dump and a register state. Fuzzing with a snapshot allows for granular execution in order to reach code blocks that are traditionally difficult to fuzz without the complexities of managing state within the target, said AWS Security Engineer Cory Duplantis, in an interview with The New Stack.

Stackie, Our New Press Release Rewriting AI

At The New Stack, we get dozens of press releases each day from various companies large and small in the cloud native computing space. A press release summarizes a recent announcement that some organization has made, such as an update to a new product. Because we get far more press releases than we have reporters to write them up, we choose just those announcements to write up that would be of most interest and have the widest repercussions for our readers. And these take time to produce because we want to bring additional context and reporting to the story not provided in the press release. 

But what about all those other discarded press releases? Many leftovers are clearly not really relevant to us, so we can easily toss those. However, there are also a few that may not be “hot news” per se, but have a bit of interesting information for readers. These are the press releases we’d otherwise discard entirely. 

Until now! Enter our newest “reporter,” codenamed “Stackie.” Stackie is our first attempt at working with ChatGPT and GenerativeAI in general. 

For The New Stack, Stackie, a fictional non-de-plume, will post short news summaries generated from press releases that have a tidbit of useful info, but not quite enough to warrant a proper story. Our first one, “Tech Byte: Meta Backs the OpenJS Foundation” ran Thursday.

We realize we are asking the trust of our readers as we move in this highly debatable direction of AI itself. So we kept Stackie on a short leash. Here are the ground rules for Stackie’s Stories:

  • At the top of each AI-generated article will be a disclaimer (“The following is an AI-generated post.”)
  • At the bottom of the post will be a link to the source press release. 
  • Also at the bottom will be the actual prompt the editor used to create the story. For the Meta news, we used this ChatGPT prompt: "Write a 500-word news story based only on the press release pasted below. Minimize any marketing jargon while emphasizing technical details." (The prompts, we hope, will evolve over time.)
  • In order to prevent AI hallucinations, the AI service will be directed to ONLY use material from the press release itself.
  • Each resulting story will be reviewed by an editor and treated like any other TNS story, with the usual changes made for clarification, concision, and accuracy. 
  • Each AI-generated story will get the tag “Tech-Byte” (and “Tech Byte” in the headline) to further distinguish it as a computer-generated article.

The format we are going for here will be for short, succinct summaries. Stackie will not analyze or interpret the material in any way, nor offer a “narrative” of any sort. The article will simply be a collection of potentially interesting tidbits, offered in an “in case you missed it”-styled brief.

To be clear, this is not a cost-cutting move. Our writers will continue to do the heavy lifting in providing analysis, explanation, and further reports on the news in their own stories, using their own voices. Their talents are our true Secret Sauce. And we’ll keep these Stackie articles to a handful each week, as they still incur costs for editing, copy editing and social media promotion.

Instead, we’re thinking that ChatGPT will allow us to cover a bit more of the cloud native ecosystem. 

Honestly, we don’t know if this will provide value to the reader or not. So please do let us know if you’d find these robo-blurbs useful, or just annoying. One thing is clear though, Generative AI is here to stay, and The New Stack, like everyone else, needs to find a way to work with it, rather than be rolled over by it. Stackie is our first attempt at finding this path. 

Tableau, Informatica, ThoughtSpot Tout Generative AI

All three data-focused vendors kick off their respective conferences this week and launch new GPT-powered capabilities to their platforms. Data market intelligence expert Andrew Brust, the principal of Blue Badge Insights, gives us the lowdown here.

The New JavaScript Features Coming in ECMAScript 2023

The next JavaScript update brings smaller additions familiar from other languages, but there are more significant developments in the wings.

Why Developers Are Flocking to LLaMA, Meta’s Open Source LLM

Developers are diving into the open source LLaMA, but what benefits does it have over GPT? We spoke to Sebastian Raschka from Lightning AI to learn more.

Mirantis Updates k0s Lightweight Kubernetes Distro

Kubernetes 1.27 compatibility, bug fixes, and support for containerD WASM and gVisor container sandboxes, these are a few of our favorite things that the new version of k0s brings us.

Want more from The New Stack?

Sign up for the TNS Daily Newsletter to get a roundup of all the most recent articles in your inbox each weekday. Register now, never miss a story, and always stay in-the-know.
Read Entire Article