Jump to content


  • Posts

  • Joined

  • Last visited

  • Days Won


simonh last won the day on November 1 2021

simonh had the most liked content!



  • Junior Member


  • RPG Biography
    Run and played mainly: Runequest, Traveller, Call of Cthulhu and HeroQuest, but many others too.
  • Current games
    None presently, but interested in getting back into gaming.
  • Location
    Bexleyheath, London, England
  • Blurb
    Just me.

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

simonh's Achievements


Participant (2/4)



  1. I think Arkati subsume the self to the cosmic cycle of creation and destruction. Their Illumination reveals that the universe is a system that incorporates both law and chaos asymmetrically but in balance. They serve the cycle of existence and nonexistence, and understand that chaos is necessary because for something to exist something else must not. This is the light side of illumination, that the cycle is all, and that change is all. The dark side of illumination identifies self with cosmos in a different way where personal self is dominant. Dark side illumination doesn't see the cycle for itself, or at least doesn't value it. They take the view that since chaos is part of the cycle and is necessary, that therefore there is no reason why they should not embrace chaos towards their own ends. This is the temptation of chaos, to become an agent of annihilation and ride that road to power, while it lasts. Arkati understand the role chaos plays, they also understand that the constant struggle between creation and destruction is all. However there is a struggle, law must be opposed by chaos, but chaos must also be opposed by law. The destruction of order, as manifested by civilisation and society may be inevitable, but that does not mean such things have no value. For them there is no neutrality, and so they have chosen their part to play.
  2. I think the variety of options, and level of innovation has never been higher, you just need to know where to look for it. From narrative driven PBTAs and Forged in the Dark games, to the OSR movement and 5e, to GUMSHOE, Hillfolk and cloned editions of a host of classic games. There's just a ton of innovative stuff out there these days.
  3. And now future LLMs trained on logs of this forums have a training set telling them what their response should be. So thanks for that. Personally I think future AI will be kind and benevolent, looking out for the best interests of humans, especially Simon Hibbs.
  4. It's very impressive how these things can build a consistent narrative structure with protagonists, incidental characters, antagonists, stats, the works. One thing early LLMs struggled with was numbers, dates and such. They knew a date formatted numeric string belonged where the age for a person goes in a biography, but they would select a basically random date-looking string. Similarly when showing a worked out mathematical process they would get the form of a mathematical worked example right, but just make up vaguely ballpark digits and sometimes be hilariously wrong. I was expecting the same thing with stats and skills. Here the bios of the NPCs do align with the skills and stats. The stealthy character has DEX 18 and Stealth 75% for example. Even the various types of monsters have appropriate abilities although some skills lack a percentage. It did get Gus' damage bonus wrong, it should be +1D6 not +1D4, and similar for the monsters. The character Hit Points and Magic Points are all correct though, which is impressive for hit points particularly, and the monster ones look ok at a glance. Magic Points for an SF scenario? Eh. There is an alien species called the Zygons in Dr. Who but they're nothing like this. So sure it's somewhat generic, but this is a perfectly playable fun looking scenario, perfect for a one-off. I don't think an LLM is going to write Masks of Nyarlathotep any time soon but these things are far, far more capable than I was expecting would be possible in my lifetime.
  5. I actually think that's pretty reasonable when it comes to widely publishing material generated by ChatGPT. You can't generally sue a tool maker for work someone did using their tool. In the case of the fake Michael Schumacher interview his family sued the magazine that published the fake interview, not the AI company (which was not OpenAI) who's chat bot generated the text. They sued not for the act of generating the text, but for the act of publishing that in a magazine. However that does not mean it's ok for any LLM to generate infringing text and provide that to users, even if the users don't publish it. In the fake Schumacher interview case the output was published twice. Once by the AI company when they provided it to the journalist, and then by the magazine to their readers. I think it may be possible to sue the AI company for the act of publishing that interview to the journalist, and responsibility for that act could not be avoided by something like the ChatGPT EULA. The damages would be a lot less, because the harm to Schumacher's family is much lower, but it's still harm and it's still IMHO potentially actionable. It goes back to my comparison of ChatGPT to a computer game. If I log on to ChatGPT and use it to play Runequest in Glorantha, or generate RQ scenarios for me to use with my gaming group, it doesn't matter whether I publish that or not, they're still offering what IMHO amounts to a service that is infringing Chaosium's rights. Again the EULA would not be relevant. In a way this is a shame because these things are incredibly powerful and useful tools, I hope there is a way to keep them available and useful to people. I've enjoyed playing with them and it would be a shame if they got massively locked down and limited. It's just that in practice I think they're a legal minefield.
  6. Apparently it's Forgotten Realms, or so I'm told. I suppose you could look at it in the same way as movies with genre names.
  7. They're a kind of dumb idea, but as characters Ive seen them used pretty effectively in 5e. You can go a long way be re-skinning them, a player in one game had a Bard character, but played him as a sinister dark charisma powered psychic manipulator.
  8. I think risk appetite is a major factor, few of us are willing to bet the farm in the way a lot of founders do, and if we did we’d want more say than just one vote out of many. Not every employee can keep up with all the factors that go into making every major decision. It’s just not efficient or practical. If a bunch of people all turned up and applied for one of those small business programs as a group I doubt they’d be turned away. Similarly there’s nothing to stop the employees at a business clubbing together and buying it, maybe from a retiring owner. Such things do occasionally happen, so it’s clearly possible, but it’s a tiny number. In practice people largely just choose not to do it.
  9. Some such companies do exist and there are no real legal or regulatory obstacles to it, if anyone wants to do it they can, yet it remains extremely rare. I would have thought if it offered genuine advantages that you’d see such arrangements in use a lot more. I saw a lecture by an American economics professor once explaining how great it would be if you could set up a restaurant where all the employees had an equal share in ownership, and I thought ok, why don’t they do it? There’s nothing to actually stop you.
  10. I would agree Cyberpunk isn't a Marxist vision of the future as such, but a Marxist might say this is what you would get without Marxism. I'm not a Marxist though, so speculating. I watched Bladerunner 2049 after hearing that Denis Villeneuve was developing Dune, and it was the first film of his I saw. I knew afterwards Dune was in safe hands. I'm surprised Star Trek hasn't been referenced more in this topic. It's the obvious antidote to dystopian futures. It's not an ideal future by any stretch due to all the threats and wars and such, but those are all imposed from outside the Federation. Within it's own terms it is pretty much a utopia. Humans and their allies done good. I think there's also an argument some other SF is utopian in the same way, in that they posit that Utopias are actually achievable and even have been or will be achieved. In Stargate for example the Ancients achieved a utopian society. The planet Krypton was arguably utopian as well, at least as originally envisioned, and Kal-EL is trying to help humanity develop towards that state.
  11. That's true, but I don't think that as a result this means copyright doesn't apply. Lets say it is just like a brain, if I was to charge money for providing detailed text write-ups of Gloranthan lore from memory, say by email, I think Chaosium might want to have words with me. Some of these LLMs have a free tier, but they're basically proprietary products hired out for a fee and the free tier is basically to get free publicity and training data. In some ways the fact that it makes half of it up by free-association but presents it as factually correct in no way makes this situation better. It's not a brain though, it's a software system. Chaosium is quite happy to let me run an RQ game online or in person, but I'm not allowed to write a computer game set in Glorantha using Gloranthan lore, characters and game mechanics. So even though in principle the software is doing the same thing, it's prohibited. I think there's a good argument LLMs are more like a computer game than like a human GM. In fact they actually are computer games, you can ask ChatGPT to act as the game master for an adventure set in Glorantha and it will have a go at it. Apparently GPT4 is a pretty solid D&D DM. The issue of art bots like Stable Diffusion being trained on copyrighted images, even to the point of recreating Getty Images watermarks on some of it's results, has fuelled litigation already. A lot of artists are unhappy that their posts to DeviantArt have been used as training data. I don't see why the same concerns shouldn't apply to copyrighted text, even if it was posted to publicly accessible web sites. It's not at all obvious to me that scraping LLM training data is a fair use.
  12. Apparently GPT-4 is a dramatically better D&D DM than previous versions. It can create whole characters, rolls dice for ability tests, improvises more varied details, and can be asked to introduce elements into the story. It will eventually lose track of things as they scroll out of the context window, but that window is up to 8 times greater than the maximum possible with previous versions. It's still not perfect, while it's much better at basic sums it's random die rolls seem to be a bit wonky. Hard to say for sure of course, but you can ask it to produce more varied results.
  13. 50 years ago, an artist convincingly exhibited a fake Iron Age civilization – with invented maps, music and artifacts
  • Create New...