Skip to main content

Indie game storeFree gamesFun gamesHorror games
Game developmentAssetsComics
SalesBundles
Jobs
TagsGame Engines
(2 edits)

Hi, Chimériquement. I wrote some of my thoughts about the situation in a devlog.

https://falsfiction.itch.io/spirit-of-the-latch/devlog/1561138/downdating-and-downloads

Only in the past day, I’ve learned how RenPyTom is using the three, online LLMs he’s basically paying some of the world’s richest and most dangerous people to use.

Screenshot of two BlueSky posts in June of ‘Ren’Py Tom USA-flag 250’ misrepresenting three LLMs as one entity, referred to as ‘it’ 75% of the time to fill in lines of code and 25% to make up new code or problem-solve through prompts.

This is worse than what he described as something like fancy autocorrect. He’s obviously trying to make his actions seem like normal and reasonable for programmers, but it’s not. Tom is obviously (to me) addicted and making weak excuses.

{Edited to add: I had forgotten to mention examples you used above. Formatting text does not require “AI”. We’ve been doing that for decades with simple programs that were faster and easier to use. For non-experts, program like Pandoc or Markor are available. I might have pointed those out to you before a ways back? Anyhow, spelling and grammar checks don’t require LLMs like Claude or ChatGPT. We had better options in the past for that, too. I don’t believe Tom is saving time or in the end helping Ren’Py as a community project with his reliance on LLMs for coding.}

In one BlueSky reply to a regular jam organizer who’s here on Itch.io, he said he’s not remorseful about any of this.

He doesn’t feel remorse the hiding that he supported the LLM companies while they scraped code and used it against software developers’ consent, took over governments for authoritarian purposes, demanded massively wasteful data centers be build in some of the most vulnerable locations, were used to impersonate and harass people (sometimes fatally), and were involved in mass layoffs that have endangered entire tech and education industries?

He doesn’t care that he has caused actual distress in visual novel communities among people who have been fighting back the AI pushers under the misbelief that Ren’Py was primarily for real people to learn, be creative, and communicate without the need to relentlessly deal with generative slop?

My initial disappointment and concern has been boiling down to anger. My respect for this guy is disappearing from every post I’ve read of his in the past month. I won’t comment on every one, but they all seem to add up to an impression of someone who can’t be trusted. I would not want to work on any project with him.

What does this mean for my relationship with Ren’Py? I’m unsure. I feel gross opening any version of Ren’Py 8. My Ren’Py projects will be moved to 2021 versions with security Python updates added. Then… I don’t know what the options will be.

The new games I’m working on this month are being made in Twine and Inky. Maybe leaning more heavily on HTML and CSS, especially for animations, will help close the gap with these smaller engines and Ren’Py, at least until the heftier alternatives are available for Linux systems.

Another thought: Your post’s conclusion suggests that good use of the models under discussion is about the quality of the end result. I disagree.

There are far too many flaws in the processes of using the models like Tom and maybe others around him have been.

Even if a comparison of the latest Ren’Py versions with Ren’Py from five years ago wasn’t showing a gap in understanding of what the program is supposed to be doing vs. what it actually does, the past five years are full of missed opportunities, wasted resources, and actual harm that are, in my mind, relevant.

I’d prefer a sloppy painting by a pig be added to project than to any AI-generated image that could fool that public as coming from an acclaimed artist. The appearance doesn’t matter weighed everything else.