What is the point of Grokipedia. It's an interesting experiment I guess (Really is it just a bunch of pre-rendered prompts I could ask Grok instead?), curious how much of wikipedia is in the training set. I would think if you wanted an alternate encyclopedia you would want something that AI can train against, so Grok itself can't probably bet too much value out of it. (
I guess people can choose their truth now? I suppose the US Government could require grokipedia to be chosen over wikipedia for use in schools?
I mean I guess I'll check it out for the lols but I don't see myself actually using it.
>... Musk announced xAI was building a new AI-generated online encyclopedia, to be called Grokipedia, in the midst of his criticisms of Wikipedia's ideological biases. The project was suggested and named by White House AI and crypto czar David O. Sacks at the All-In podcast conference earlier that month. According to Musk's announcement, it would be an AI-powered knowledge base designed to rival Wikipedia by addressing its perceived biases, errors, and ideological slants.
I suggest that anyone interested compare the content of Wikipedia and Grokpedia articles on topics that interest them, as well as the differences in sources between these two projects. Of course, only if someone finds this research interesting.
I have tried briefly checking two pages about Russo-Ukrainian War. First of all, hilariously Elonopedia starts from 1917-1921 war and goes on about it for multiple paragraphs, then suddenly switches to the 2014 invasion. And no, it's not in the "history" section, it's a main starting section.
Then actual description of the war is much more biased in the Elonopedia. In every case possible the invasion is presented as "both sides are guilty". I wouldn't list the examples, anyone can do it. Too much effort imo.
Then I checked Russo-Georgian War articles, this time at least the century and war was correct in Elonopedia. But again, right from the start it is incredibly biased towards Russia. Elonopedia completely omits the initial attack make bu Russian forces at 01 Aug 2008, skip a week and presents war as if it was initiated by Georgians, following Kremlin propaganda line. Didn't both reading full article.
All in all it is 100% as I have expected reading the news about this supposedly "unbiased" encyclopedia - it's a LLM-generated slop, with no human fact checking (mixing two different century separated wars into one article is telling), and it is essentially a far-right propaganda outlet. It will follow Goebbels rule of mixing 60% or truth with 40% of lies, to prime up unsophisticated readers towards Elon's and rightwing crowd goals.
I compared Grokipedia's entry on the band "American Football" [^0] to Wikipedia's [^1] and they are _almost_ the same. While Grok does attribute Wikipedia in the footer, they added this to their article:
> On July 2, 2025, the band released their first live album, American Football (Live in Los Angeles), recorded during the anniversary shows at the El Rey Theatre in Los Angeles with guest appearances by Ethel Cain and M.A.G.S., accompanied by a concert film documenting the performance.
If you go to the source [^4] for this claim, you'll see that:
- They dropped a film of the same name alongside the album release.
- The "guest appearances" are actually interviews in the film.
- The entry excluded the female artist that was cited in the source.
I, then, compared Grok's entry on United Airlines [^2] against Wikipedia's [^3]. Grok's seemed to be autogenerated this time.
I skipped to the section on MileagePlus since I know a bit about how that program works. It has a few inaccuracies:
- It only lists the four published MileagePlus tiers: Silver, Gold, Platinum and 1K and omits the two unpublished, but well-known, tiers above 1K: Global Services and Chairman's Circle.
- The 2025 premier qualifying point (PQP) redemptions are actually from 2024.
- Some of the language it uses wouldn't meet Wikipedia's editorial standards, like the nebulous "priority everything" benefit from obtaining 1K status (whose source is unclear, as neither of the two sources cited use this phrase).
- "The current logo features a stylized "U" incorporating a world map outline, symbolizing global connectivity" That's United's old logo. They absorbed Continental's logo when they merged.
- The article opens with the claim that United has 1018 aircraft in its fleet as of APR 2025, then, later, states that it has 1,001 active aircraft as of OCT 2025. The source for the 1,001 figure states 1,055 on the page with 1,003 in revenue service.
So I wouldn't use Grokipedia as a source for anything, just like Wikipedia, though I'm sure some will try.
By "apartheid apologetics" do you mean that it is factually wrong or merely that you dislike the framing? I think there is a huge difference between those two accusations.
I've spotted surprising amounts of confidently-stated nonsense even in fairly neutral articles where Elon / xAI is unlikely to have a particular political slant.
Many of the most glaring errors are linked to references which either directly contradict Grokipedia's assertion or don't mention the supposed fact one way or the other.
I guess this is down to LLM hallucinations? I've not used Grok before, but the problems I spotted in 15 mins of casual browsing made it feel like the output of SoA models 2-3 years ago.
Has this been done on the cheap? I suspect that xAI should probably have prioritised quality over quantity for the initial launch.
Some time ago, there was a project called Citizendium that aimed for quality over quantity, with articles written and peer-reviewed by subject matter experts who had to use their real names and working email addresses, among other requirements. I always thought that was interesting, since the main critique of Wikipedia is its open editing model.
Citizendium is still around, though they've loosened some of the requirements in order to encourage more contributions, which seems self-defeating to me. I think they should have tried to cooperate with Wikipedia instead. The edits and opinions of subject matter experts could be a special layer on top of existing Wikipedia articles. Maybe there could be a link for various experts with highlights of sections they have peer-reviewed and a diff of what they would change about the article if those changes haven't been accepted. There could also be labels for how much expert consensus and trust there is on a given snapshot of an article or how frozen the article should be based on consensus and evidence provided by the experts. This would help users delineate whether an article contains a lot of common knowledge or whether it's more speculative or controversial.
When I was a kid you could subscribe to an encyclopedia. They were too expensive to be bought in one go and they were too expensive to make in one go. The solution was to sell them in installments and mail you the monthly addition that you could add. Obviously the marketing ploy was that if you were halfway through the 'A' that you would buy the rest of the A and once you had the 'A' then you'd buy the rest of the book.
Regardless, the business was there. Wikipedia killed all that. So if you want to create an expertly created encyclopedia anno 2025 you have a real problem: you will need to pay experts for their time somehow otherwise why would they compete with the million monkeys, but your source of revenue has been strangled by those very same monkeys, who it turns out produce content that is orders of magnitude better than anything I've ever read in a for-pay encyclopedia from before Wikipedia.
I'm no apartheid apologist, but I have lived here (ZA) all my life.
Whilst I haven't read the entire article, the first paragraph is actually on-point: apartheid was shit in a lot of respects, but the schools, especially in rural areas, have dramatically declined since 1994, as have most government-run companies (with the exceptions like Eskom being bailed out every year).
You don't have to like the facts, but that's what they are.
As a Green Line enjoyer, I'm not enough of an expert to spot the factual errors. The article does seem a bit much though. I've noticed a lot of the Grokipedia articles just go on. If I wanted to know that much about the Green Line I could probably just buy a book on it.
Looking up the Republican Party "controversies" vs the Democratic Party "controversies" should let you know exactly what this projects intentions are.
That being said, my biggest issue with it is how Grok is writing everything. It's like it is trying REALLY hard to be neutral but it's conversational training slips up and starts "spicing" things up a little. For example on Elon's article:
"...at age 12 in 1983, developing a space-themed video game called Blastar, which he sold to PC and Office Technology magazine for approximately $500. *This early entrepreneurial act foreshadowed Musk's later pursuits in technology and business*."
Sentences like that are designed to subtly bring emotion to certain topics.
I guess this poses an interesting question: if Wikipedia was being created today, would it be a human- edited encyclopedia or would they just resort to AI because it’s easier? It makes me wonder if people will shy away from hard problems and just take the easy path, resulting in a shallower and less useful product to society.
I have to wonder, obviously the grok version exists just to push Elons politics and take control over the “truth”. But it seems like an LLM could take over. Considering Wikipedia is not meant to contain any original facts, just a collection of references to external information.
If you’ve tried OpenAI’s Deep Research or similar tools, you’ll know they pull far more info than Wikipedia. But if you’re an expert, you’ll quickly spot errors since the breadth is huge but the depth and accuracy are only so-so.
For non-experts just exploring new topics, it’s still perfectly useful. Grokipedia probably uses a similar search, verify, summarize workflow, so it naturally inherits mistakes from the internet, which isn’t really an LLM problem.
Grok is just the first to make it public, and other AI companies could easily build their own synthetic data Wikipedias, and some probably already have.
Wikipedia’s coverage looks broad, but it still can’t keep up with how fast knowledge grows. And the gaps are even more severe in non-English versions of Wikipedia.
I tried it briefly. Its an ok initial start with significant flaws - while it counters Wikipedia's (editorial demographic) bias on cultural topics it seems to do so by assuming the style of a bland, unanalytical reporter, accepting the self-framing of the subject and relaying it at turgid length.
An example: the classical liberal writer Douglas Murray is one of the many targets on Wikipedia of ludicrous "far right" style categorizations; nevertheless its correct to attempt to draw out his own alignments and biases especially where he writes provocatively in areas with cultural tensions.
Grokipedia seems to smooth over those tensions almost in denial while Wikipedia stirs them up via exaggeration. I don't think either are helpful or honest.
patiently waiting for grokipedia’s article on grokipedia. it seems to not be available at the moment. i’m interested from a philosophical perspective: on the completeness of self-description. for example, here’s wikipedia on wikipedia[0]
Well, they use AI to create "content". So I guess it's fair to use AI to read it.
This is what Claude says about "Grokipedia"'s article about the Gaza war, compared to the original:
Major Differences Between Wikipedia and Grokipedia's Gaza War Articles
1. Framing and Perspective
Wikipedia: Presents the conflict with multiple perspectives, acknowledging disputed narratives. Uses neutral language like "armed conflict" and presents genocide allegations as claims made by "many human rights organizations and scholars."
Grokipedia: Frames the conflict almost entirely from an Israeli perspective. Hamas is consistently portrayed as the aggressor and sole source of civilian suffering, with Israeli actions defended as necessary self-defense.
2. Casualty Figures and Reporting
Wikipedia: Reports over 79,000 Palestinians killed in Gaza as reported figures, noting they come from the Gaza Health Ministry but presenting them as the available data.
Grokipedia: Systematically questions and undermines Palestinian casualty figures, dedicating entire sections to "Verification Challenges and Inflated Figures" and "Combatants Versus Civilians." It emphasizes that figures are "Hamas-administered" and suggests deliberate fabrication, claiming the ministry has "incentives for propagandistic reporting."
3. Treatment of Genocide Allegations
Wikipedia: States that "many human rights organizations and scholars of genocide studies and international law, including an independent UN commission, say that Israel is committing genocide in Gaza, though some dispute this".
Grokipedia: Dismisses genocide allegations as part of "double standards in scrutiny" and frames them as politically motivated attacks on Israel's legitimate self-defense. The word "genocide" appears primarily in sections criticizing those who make such claims.
4. Hamas's Responsibility for Civilian Harm
Wikipedia: Mentions Hamas's use of civilian infrastructure but doesn't make it the primary explanation for Palestinian casualties.
Grokipedia: Contains extensive sections titled "Impact of Hamas Tactics on Civilian Suffering" arguing that Hamas's embedding of military assets in civilian areas is the primary cause of Palestinian civilian deaths, stating "Hamas's operational choices to a disproportionate share of Palestinian suffering, independent of Israeli response intensity."
5. Aid and Humanitarian Crisis
Wikipedia: Describes Israel's blockade cutting off necessities and causing famine, with Israel's actions as a key factor.
Grokipedia: Features a section on "Aid Distribution Failures and Diversion by Hamas," emphasizing that "Hamas diverts up to 25% of incoming aid supplies" and that aid failures stem primarily from Hamas's control and diversion rather than Israeli restrictions.
6. Language and Terminology
Wikipedia: Uses terms like "Israeli invasion," "Israeli offensive," and "Israeli strikes" in a descriptive manner.
Grokipedia: Uses emotionally charged language like Hamas's "systematic atrocities," "barbarism," and describes October 7 as involving "mass killings, sexual violence, and arson" while Israeli actions are described as "targeted operations," "precision strikes," and "necessary self-defense."
7. International Law and War Crimes
Wikipedia: Notes that "experts and human rights organizations have stated that Israel and Hamas have committed war crimes", treating both sides' alleged violations seriously.
Grokipedia: Has separate sections for "Hamas Violations" and "Israeli Actions Under International Law," with the Hamas section focusing on terrorism...
You're right, that's definitely a mistake. Though to be fair, the same article gets it right if you scroll down to the Tesla section. The article on Tesla also gets it right.
51 comments
[ 4.8 ms ] story [ 75.3 ms ] threadI guess people can choose their truth now? I suppose the US Government could require grokipedia to be chosen over wikipedia for use in schools?
I mean I guess I'll check it out for the lols but I don't see myself actually using it.
(source Wikipedia)
Nevertheless this reminds me of an old Curtis Yarvin post on his proposal for a meta-wikipedia. "Uberfact". He's not everyone's cup of tea but I quite enjoy this article of his - https://www.unqualified-reservations.org/2007/08/uberfact-ul...
Then actual description of the war is much more biased in the Elonopedia. In every case possible the invasion is presented as "both sides are guilty". I wouldn't list the examples, anyone can do it. Too much effort imo.
Then I checked Russo-Georgian War articles, this time at least the century and war was correct in Elonopedia. But again, right from the start it is incredibly biased towards Russia. Elonopedia completely omits the initial attack make bu Russian forces at 01 Aug 2008, skip a week and presents war as if it was initiated by Georgians, following Kremlin propaganda line. Didn't both reading full article.
All in all it is 100% as I have expected reading the news about this supposedly "unbiased" encyclopedia - it's a LLM-generated slop, with no human fact checking (mixing two different century separated wars into one article is telling), and it is essentially a far-right propaganda outlet. It will follow Goebbels rule of mixing 60% or truth with 40% of lies, to prime up unsophisticated readers towards Elon's and rightwing crowd goals.
> On July 2, 2025, the band released their first live album, American Football (Live in Los Angeles), recorded during the anniversary shows at the El Rey Theatre in Los Angeles with guest appearances by Ethel Cain and M.A.G.S., accompanied by a concert film documenting the performance.
If you go to the source [^4] for this claim, you'll see that:
- They dropped a film of the same name alongside the album release.
- The "guest appearances" are actually interviews in the film.
- The entry excluded the female artist that was cited in the source.
I, then, compared Grok's entry on United Airlines [^2] against Wikipedia's [^3]. Grok's seemed to be autogenerated this time.
I skipped to the section on MileagePlus since I know a bit about how that program works. It has a few inaccuracies:
- It only lists the four published MileagePlus tiers: Silver, Gold, Platinum and 1K and omits the two unpublished, but well-known, tiers above 1K: Global Services and Chairman's Circle.
- The 2025 premier qualifying point (PQP) redemptions are actually from 2024.
- Some of the language it uses wouldn't meet Wikipedia's editorial standards, like the nebulous "priority everything" benefit from obtaining 1K status (whose source is unclear, as neither of the two sources cited use this phrase).
- "The current logo features a stylized "U" incorporating a world map outline, symbolizing global connectivity" That's United's old logo. They absorbed Continental's logo when they merged.
- The article opens with the claim that United has 1018 aircraft in its fleet as of APR 2025, then, later, states that it has 1,001 active aircraft as of OCT 2025. The source for the 1,001 figure states 1,055 on the page with 1,003 in revenue service.
So I wouldn't use Grokipedia as a source for anything, just like Wikipedia, though I'm sure some will try.
[^0]: https://archive.is/twkBP (might not be available yet; it's still getting archived)
[^1]: https://archive.ph/lOkdT
[^2]: https://archive.ph/EnN2T
[^3]: https://archive.ph/uooNW
[^4]: https://pitchfork.com/news/american-football-to-share-new-li...
Reflections on Trusting Trust is mentioned in Ken Thompson's page but when I searched for it he wasn't part of the results.
And no surprise, apartheid apologetics: https://grokipedia.com/page/Apartheid#debunking-prevailing-n...
Hilarious factual errors in https://grokipedia.com/page/Green_Line_(CTA)
Many of the most glaring errors are linked to references which either directly contradict Grokipedia's assertion or don't mention the supposed fact one way or the other.
I guess this is down to LLM hallucinations? I've not used Grok before, but the problems I spotted in 15 mins of casual browsing made it feel like the output of SoA models 2-3 years ago.
Has this been done on the cheap? I suspect that xAI should probably have prioritised quality over quantity for the initial launch.
Citizendium is still around, though they've loosened some of the requirements in order to encourage more contributions, which seems self-defeating to me. I think they should have tried to cooperate with Wikipedia instead. The edits and opinions of subject matter experts could be a special layer on top of existing Wikipedia articles. Maybe there could be a link for various experts with highlights of sections they have peer-reviewed and a diff of what they would change about the article if those changes haven't been accepted. There could also be labels for how much expert consensus and trust there is on a given snapshot of an article or how frozen the article should be based on consensus and evidence provided by the experts. This would help users delineate whether an article contains a lot of common knowledge or whether it's more speculative or controversial.
Regardless, the business was there. Wikipedia killed all that. So if you want to create an expertly created encyclopedia anno 2025 you have a real problem: you will need to pay experts for their time somehow otherwise why would they compete with the million monkeys, but your source of revenue has been strangled by those very same monkeys, who it turns out produce content that is orders of magnitude better than anything I've ever read in a for-pay encyclopedia from before Wikipedia.
The bar to entry is insanely high.
Weird that it displaying some other web site's embed/shortcodes:
> ![Cottage Grove-bound Green Line train approaching Roosevelt station][float-right] The Green Line utilizes primarily 5000-series railcars
Whilst I haven't read the entire article, the first paragraph is actually on-point: apartheid was shit in a lot of respects, but the schools, especially in rural areas, have dramatically declined since 1994, as have most government-run companies (with the exceptions like Eskom being bailed out every year).
You don't have to like the facts, but that's what they are.
Impossible! That article was "Fact checked by Grok yesterday!"
I'm glad we've solved the LLM hallucination problem by fact-checking with LLMs. No way that could go wrong.
That being said, my biggest issue with it is how Grok is writing everything. It's like it is trying REALLY hard to be neutral but it's conversational training slips up and starts "spicing" things up a little. For example on Elon's article:
"...at age 12 in 1983, developing a space-themed video game called Blastar, which he sold to PC and Office Technology magazine for approximately $500. *This early entrepreneurial act foreshadowed Musk's later pursuits in technology and business*."
Sentences like that are designed to subtly bring emotion to certain topics.
For non-experts just exploring new topics, it’s still perfectly useful. Grokipedia probably uses a similar search, verify, summarize workflow, so it naturally inherits mistakes from the internet, which isn’t really an LLM problem.
Grok is just the first to make it public, and other AI companies could easily build their own synthetic data Wikipedias, and some probably already have.
An example: the classical liberal writer Douglas Murray is one of the many targets on Wikipedia of ludicrous "far right" style categorizations; nevertheless its correct to attempt to draw out his own alignments and biases especially where he writes provocatively in areas with cultural tensions.
Grokipedia seems to smooth over those tensions almost in denial while Wikipedia stirs them up via exaggeration. I don't think either are helpful or honest.
https://grokipedia.com/page/Douglas_Murray_(author)
https://en.wikipedia.org/wiki/Douglas_Murray_(author)
[0]: https://en.wikipedia.org/wiki/Wikipedia
Major Differences Between Wikipedia and Grokipedia's Gaza War Articles
1. Framing and Perspective
Wikipedia: Presents the conflict with multiple perspectives, acknowledging disputed narratives. Uses neutral language like "armed conflict" and presents genocide allegations as claims made by "many human rights organizations and scholars."
Grokipedia: Frames the conflict almost entirely from an Israeli perspective. Hamas is consistently portrayed as the aggressor and sole source of civilian suffering, with Israeli actions defended as necessary self-defense.
2. Casualty Figures and Reporting
Wikipedia: Reports over 79,000 Palestinians killed in Gaza as reported figures, noting they come from the Gaza Health Ministry but presenting them as the available data.
Grokipedia: Systematically questions and undermines Palestinian casualty figures, dedicating entire sections to "Verification Challenges and Inflated Figures" and "Combatants Versus Civilians." It emphasizes that figures are "Hamas-administered" and suggests deliberate fabrication, claiming the ministry has "incentives for propagandistic reporting."
3. Treatment of Genocide Allegations
Wikipedia: States that "many human rights organizations and scholars of genocide studies and international law, including an independent UN commission, say that Israel is committing genocide in Gaza, though some dispute this".
Grokipedia: Dismisses genocide allegations as part of "double standards in scrutiny" and frames them as politically motivated attacks on Israel's legitimate self-defense. The word "genocide" appears primarily in sections criticizing those who make such claims.
4. Hamas's Responsibility for Civilian Harm
Wikipedia: Mentions Hamas's use of civilian infrastructure but doesn't make it the primary explanation for Palestinian casualties.
Grokipedia: Contains extensive sections titled "Impact of Hamas Tactics on Civilian Suffering" arguing that Hamas's embedding of military assets in civilian areas is the primary cause of Palestinian civilian deaths, stating "Hamas's operational choices to a disproportionate share of Palestinian suffering, independent of Israeli response intensity."
5. Aid and Humanitarian Crisis
Wikipedia: Describes Israel's blockade cutting off necessities and causing famine, with Israel's actions as a key factor.
Grokipedia: Features a section on "Aid Distribution Failures and Diversion by Hamas," emphasizing that "Hamas diverts up to 25% of incoming aid supplies" and that aid failures stem primarily from Hamas's control and diversion rather than Israeli restrictions.
6. Language and Terminology
Wikipedia: Uses terms like "Israeli invasion," "Israeli offensive," and "Israeli strikes" in a descriptive manner.
Grokipedia: Uses emotionally charged language like Hamas's "systematic atrocities," "barbarism," and describes October 7 as involving "mass killings, sexual violence, and arson" while Israeli actions are described as "targeted operations," "precision strikes," and "necessary self-defense."
7. International Law and War Crimes
Wikipedia: Notes that "experts and human rights organizations have stated that Israel and Hamas have committed war crimes", treating both sides' alleged violations seriously.
Grokipedia: Has separate sections for "Hamas Violations" and "Israeli Actions Under International Law," with the Hamas section focusing on terrorism...
Does anyone know why replies are disabled in that other submission?
>..Musk founded SpaceX in 2002 as CEO and chief engineer, Tesla in 2003 ... (grok)
>Musk joined the automaker Tesla as an early investor in 2004 and became its CEO ... (wikipedia)
I think Wikipedia is more accurate on that one.