Hi friends! To do my small part for the TTRPG fediverse, I resurrected an old RPG blog rss aggregator I ran about 10 years ago and refreshed it with some GPT-based summarization. You can find it right here:
It updates every morning from the following feed lists.
https://slyflourish.com/sly_flourish_rpg_blogroll.html
If there are any great TTRPG blogs I’m missing, please let me know!
Very cool.
I’m surprised you don’t have any columns from rpg.net on your blogroll, however. Any reason for that, or is it just one you haven’t seen before? I know its been around since, like, the 90s.
They have an RSS specifically for their “columns” section, which looks to me to have about one post a week or so.
https://www.rpg.net/columns/index.xml
I dont know how good they are these days, though, i used to follow them back in the mid 2000’s.
@eerongal @slyflourish The column pickings there have shriveled over the years in favor of their forums. The main column there that still matters is Designers & Dragons.
Oh this is great :)!
This is truly great! I really appreciate the summaries as well, so i don’t have to open tabs for everything to find out if I am even interested. I also realised that the fediverse reemergence goes very well with the reimagining of the web experience that the people working on arc browser are doing! Check it out, it is great for a workflow involving multiple sites bookmarks etc
Nice work, always handy to have another resource…and a good use of AI Chat doing the summaries.
Very cool! But I think the current summarization is a bit hit or miss. Would you be interested in community help (volunteer basis) refining the prompts for summarization? GPT-4 also is rolling out to paid users, I’m making a hunch here but it looks like it’s using GPT-3.5-turbo instead of GPT-4, we might be able to get more interesting summarizations using the newer model. But either way it seems like the prompts for summarizing could be tweaked a bit to more than just summarize the contents of hte article, but summarize the contents and extract what is interesting for the reader for the ttrpg DMing community. Just my 2c. Awesome work, hugely appreciate (another) Lazy DM resource, Mike!
Sure, lets give it a try. I am indeed using 3.5 turbo. 4 is a bit more expensive (like 10x)
Yeah, GPT-4 is way more costly. It’s rolled out to the paid tier of users AFAIK, though, so in your script it’s as easy as subbing out “3.5-turbo” with “4” – it may be worth just doing some trial runs and seeing if it’s markedly improved. I expect though there’s some room for tweaks with the 3.5-turbo model, though.
Temperature is something you can play with to get more predictable or more creative results; and is set between 0 and 1 with .1/.2 being more useful for every wrote data entry stuff and .7-.9 being more creative. That’s something to play with (I’d probably start around .6 or .7). But tweaking the prompt itself and the “role:” instructions will almost certainly return the most immediate benefits.
I’m assuming your code looks something like:
response = openai.ChatCompletion.create( model="gpt-3.5-turbo", temperature=0.2, messages=[ { "role": "user", "content": f"Please summarize this blog post in 3-5 bullet points: {blog}" }
You can add a “role”: “system” instruction to make clear that, e.g., the model is supposed to assume that they are summarizing this work for a dm/ttrpg audience, and reiterate that in your user content message as well:
response = openai.ChatCompletion.create( model="gpt-3.5-turbo-16k", temperature=0.6, messages=[ { "role": "system", "content": "You are an assistant for a TTRPG blog designed to help gamemasters and dungeonmasters. You have been tasked with reading blog posts from other game masters and identifying the creative and interesting takeaways and summarizing them for a DM/GM audience. The factual summarizing of the structure of the blog is less important than the relevant actionable advice for running interesting and engaging TTRPG sessions" }, { "role": "user", "content": f"Please summarize this ttrpg blog post in 3-5 bullet points for an audience of gamemasters and ttrpg enthusiasts, identifying the most interesting, actionable and creative take-aways: {blog}" }
Go crazy with those instructions, though. That’s the real secret to getting good results from GPT chat/completion models. I also swapped out the vanilla 3.5-turbo model with the only barely more expensive ‘turbo-16k’ model, which has a higher token limit, and bumped the temperature up a bit.
Great stuff! I’ll experiment with it and see how I can get better results. I’ve already been using the turbo-16k so I can feed large blog articles back to it. I also need to find a way to limit the total tokens coming out to ensure the summaries don’t get to long sometimes.