Lightnews — Scholar-powered news

ads

@adrienne.radioactive.dog

66 followers 30 following 560 posts

cool, cute, and collected! (they/them)
technical game designer
adrienne(at)radioactive(dot)dog

Posts Replies Media Videos

ads

@adrienne.radioactive.dog

I wanna do a ftp test in a couple hours but my brain is done sleeping 😭😭 not even 4am yet bro

December 1, 2025 at 11:07 AM

ads

@adrienne.radioactive.dog

yay!

December 1, 2025 at 1:56 AM

Reposted by ads

Alanna Smith

@alannamode.bsky.social

Anyway, here are my posts about DnA Creative being "Narrative QA" for games, film, television, and creatives + IPs at large.

Feel free to reach out if you'd actually like us to work on your projects. We do terrific stuff! 😁

Site + Contact: www.dnacreative.ca
Case Studies: dnacreative.notion.site

A post of mine from LinkedIn:

What if your narrative team had its own dedicated QA support?

A discussion on Bluesky about creative writing (and the related incompetence of AI) brought up a great thing that DnA Creative Inc. does that I realize I don't talk about enough - specifically narrative support.

Coming from film and into games, Delaney and I have built up our expertise particularly in roles like script supervisor, script coordinator, data wrangler, producer and more. So now as part of our lore and knowledge management work we:

- compile & help ratify scattered details into a comprehensive source
- write up internal wiki articles, even summarizing content that may not have been documented before
- identify plot holes or continuity errors the team can address
- build and maintain character, location & item databases for reference
- create choice & consequence breakdowns for storylines and conversations

Our role is to *support* the writers, flag issues effectively, and provide them with the tools they need to do their best work. We learn the content, becoming their reliable archivists and making connections that present the whole story as a more legible and tightly-woven tapestry.

Continuation of the first post:

On Code 8 we made a labelling system to denote canon info vs. promising yet unconfirmed ideas. For visual novel Love Shore we took 80+ page scripts and wrote full choice + consequence breakdowns (see below!) As a bonus, it became a base for the walkthrough we made on the public Fandom wiki for players

Tracking, documenting and maintaining it all is a huge advantage, that’s why we offer to work with studios throughout development!

(The post shows links to more case studies and to our contact info, as well as an image of a story lore page.)

A post of mine from LinkedIn:

An opportunity to dunk on GenAI *and* uplift the value of DnA Creative's work? Don't mind if I do! 😁 Strap in for a long post my friends 💁🏾‍♀️

This article covering GenAI policies from Embracer's annual report caught my eye. Particularly where they claim that "as AI models become more powerful, we can leverage their capacity also in the creative process, for example, by identifying inconsistencies in scripts and storytelling." Ha!

This is, objectively, not something GenAI/LLMs are truly capable of. I know because being proficient at it is part of my job and the business we've built.

I've mentioned how as Lore Managers we sometimes fulfill a sort of 'narrative QA' role for teams, identifying continuity errors while building out their source of truth. Something like ChatGPT can scrape documents and tell you info based on the content - but it doesn't "know" anything. It can't "learn" like a person, it can only consume, memorize, combine, and regurgitate based on a set of your instructions (prompts).

Continuity and inconsistencies are not so simple (ask any script supervisor!). Sometimes they involve an absence of information; seemingly correct info that unfortunately conflicts in certain contexts; a minor error that branches into larger additional issues; or worse, something MEANT to be conflicting or mysterious in the story that a bot wouldn't understand. GenAI can't tell right from wrong, adjudicate information, or seek answers from key knowledge holders, and there's no evidence it'll be able to.

A continuation of the second post:

I've told folks to pay attention to the use of conjectural language when it comes to things that techbros are pushing. "As AI becomes more powerful", overuse of words like eventually, might, suspect, could - it's all intentional obfuscation. Embracer themselves even note that GenAI may produce "unethical, biased, discriminatory or completely wrong results" - but hey forget all that logic and focus on how capable it'll be in the undefined future, based on no evidence at all! (No thanks)

If you're interested in actually uplifting the creative process and removing barriers for your team to do their best work through documentation and knowledge management, you know where to find us. Until then, stay alert when it comes to those empty AI promises!

(The post then shows links to our other work and ways to contact DnA Creative.)

November 30, 2025 at 4:27 PM

ads

@adrienne.radioactive.dog

i was supposed to test the weird autumn update but i honestly had 0 time or motivation to do that cause my career took off at that point

November 30, 2025 at 3:06 AM

ads

@adrienne.radioactive.dog

huge idea guy energy to announce you are implementing a storefront into your emulation launcher when the pics people are sending me of your current builds look like this

November 30, 2025 at 1:53 AM

ads

@adrienne.radioactive.dog

yay!!!

November 29, 2025 at 10:51 PM

Reposted by ads

noé

@celechii.itch.io

here's some that i have the Biggest Opinions About wrt the new designs and the software. i can make them feel like what using the programs feel like to me, and it feels sooooo good. i'll prolly draw some more and write up a blog post about my thoughts about each of them cause it was really fun :)