Grokipedia is more interesting than it seems imo, because there’s this very sensible step that AI companies are going to have to take at some point: having their AI maintain its own knowledgebase, source its own evidence/training data, reflect on its beliefs and self-correct, hammer out inconsistencies, and there’s going to be a lot of pressure to make this set of beliefs legible and accountable to the safety team or to states or to the general public. And if they did make it legible to the general public (they probably should?) then all of this is pretty much exactly equivalent to the activity of maintaining a free online encyclopedia.
Is this how they’re thinking about it behind the scenes? It probably is! They’re an AI company! They spent like half of grok4′s training compute on post-training, they know how important rumination or self-guided learning is.
When Elon talked at the All-In-Summit about how they want to train new versions of Grok entirely based on synthethic data David Sachs asked him to create Grokipedia as Wikipedia alternative. To me, Elon sounded at that moment like he didn’t really consider the idea before and that it made sense and promised to talk to his team about it.
I ought to know the base rate of his team deciding yes or no to ideas he got from podcast conversations but I don’t.
But it also sounds like they’re kind of already doing it for the reasons I suggested, like yes, they were doing knowledgebase consistency first and then Sachs suggested that an encyclopedia naturally falls out of that. I’d expect grok to be doing rag before making these data edits, so if the thing it’s retrieving from is also something it’s curating, organizing and possibly editing, that’s the thing.
I don’t think Elon team says “no” to Elon unless Elon asks for something that’s crazy.
Elon spoke about releasing Grokipedia in a open source fashion, which suggests other companies could also train on it instead of it being a competitive advantage in terms of training Grok.
It’s a way to make Claude/Gemini/ChatGPT less woke given that Grokipedia is supposed to be non-woke.
I’m not sure they were working on consistency. If you run a lot of Deep Research queries of all sorts of questions, you get a lot of text content that you can feed into your algorithm without needing it to be fully consistent.
States will restrict government use of models they don’t trust. Government contracts are pretty lucrative.
The public, or at least part of it, may also prefer to use models that are consistent in their positions, as long as they can explain their positions well enough (and they’re very good at doing that). I guess Politicians are counterevidence against this, but it’s much harder for a chat assistant/discourse participant to get away with being vague, people get annoyed when politicians are vague already, someone you’re paying to give you information, the demand for taking a stance on the issues is going to be greater.
But I guess for the most part it wont be driven by pressure, it’ll be driven by an internal need to debug and understand the system’s knowledge rumination processes. The question is not so much will they build it but will they make it public. They probably will, it’s cheap to do it, it’ll win them some customers, and it’s hard to hide any of it anyway.
Grokipedia is more interesting than it seems imo, because there’s this very sensible step that AI companies are going to have to take at some point: having their AI maintain its own knowledgebase, source its own evidence/training data, reflect on its beliefs and self-correct, hammer out inconsistencies, and there’s going to be a lot of pressure to make this set of beliefs legible and accountable to the safety team or to states or to the general public. And if they did make it legible to the general public (they probably should?) then all of this is pretty much exactly equivalent to the activity of maintaining a free online encyclopedia.
Is this how they’re thinking about it behind the scenes? It probably is! They’re an AI company! They spent like half of grok4′s training compute on post-training, they know how important rumination or self-guided learning is.
When Elon talked at the All-In-Summit about how they want to train new versions of Grok entirely based on synthethic data David Sachs asked him to create Grokipedia as Wikipedia alternative. To me, Elon sounded at that moment like he didn’t really consider the idea before and that it made sense and promised to talk to his team about it.
I ought to know the base rate of his team deciding yes or no to ideas he got from podcast conversations but I don’t.
But it also sounds like they’re kind of already doing it for the reasons I suggested, like yes, they were doing knowledgebase consistency first and then Sachs suggested that an encyclopedia naturally falls out of that. I’d expect grok to be doing rag before making these data edits, so if the thing it’s retrieving from is also something it’s curating, organizing and possibly editing, that’s the thing.
I don’t think Elon team says “no” to Elon unless Elon asks for something that’s crazy.
Elon spoke about releasing Grokipedia in a open source fashion, which suggests other companies could also train on it instead of it being a competitive advantage in terms of training Grok.
It’s a way to make Claude/Gemini/ChatGPT less woke given that Grokipedia is supposed to be non-woke.
I’m not sure they were working on consistency. If you run a lot of Deep Research queries of all sorts of questions, you get a lot of text content that you can feed into your algorithm without needing it to be fully consistent.
Where does this pressure come from?
States will restrict government use of models they don’t trust. Government contracts are pretty lucrative.
The public, or at least part of it, may also prefer to use models that are consistent in their positions, as long as they can explain their positions well enough (and they’re very good at doing that). I guess Politicians are counterevidence against this, but it’s much harder for a chat assistant/discourse participant to get away with being vague, people get annoyed when politicians are vague already, someone you’re paying to give you information, the demand for taking a stance on the issues is going to be greater.
But I guess for the most part it wont be driven by pressure, it’ll be driven by an internal need to debug and understand the system’s knowledge rumination processes. The question is not so much will they build it but will they make it public. They probably will, it’s cheap to do it, it’ll win them some customers, and it’s hard to hide any of it anyway.