Will Elon Musk open source the twitter Algorithm? (in 2023)
88
346
1.8K
resolved Apr 1
Resolved
YES

Will resolve to yes if the Twitter recommendation algorithm is available on github at any point before the end of 2023.

Here is what Elon said before buying twitter:

“One of the things that I believe Twitter should do is open source the algorithm, and make any changes to people’s tweets, if they are emphasized or de-emphasized, that action should be made apparent so anyone can see that that action has been taken. So there’s no sort of behind the scenes manipulation, either algorithmically or manually.”

https://youtu.be/cdZZpaB2kDM

Feb 18, 7:12pm: Will Elon Musk open source the Twitter Algorithm? (in 2023) → Will Elon Musk open source the twitter Algorithm? (in 2023)

Get Ṁ1,000 play money

🏅 Top traders

#NameTotal profit
1Ṁ806
2Ṁ571
3Ṁ411
4Ṁ380
5Ṁ193
Sort by:
predicted YES

code has been published to github.
its a lot of code.
I'm a coder but not fluent in scala and starlark so im gonna probably look at general consensus on this topic to resolve this market.
If the open code is missing critical parts of user recommendation this market will stay open.
Feel free to let me know what you think on this topic.

sold Ṁ4 of NO

@Kormann I'm not sure what your criteria are, but if you're using the criteria of this market, people are discussing it in the most recent comments. Based on this comment of yours, it may just come down to whether you think the missing parts are "critical," which is very subjective.

bought Ṁ5 of NO

It seems released https://twitter.com/elonmusk/status/1641874582473695246

But it can be an April Fool!

Not sure what “the” twitter algo is, they have a lot

bought Ṁ140 of YES

Per George Hotz quitting—impossible to ship anything at twtr

Elon probably knows some anon will write a better rec engine over the weekend and then he can make twtr use it

bought Ṁ180 of YES

It’s basically half decade old simple ml, in before

bought Ṁ30 of NO
bought Ṁ200 of YES
bought Ṁ30 of NO

There is no singular "Twitter algorithm". Like Google search or Netflix recommendations or anything else, the decisions are being made by an incomprehensibly massive web of interacting systems. There's no coherent single thing to open source.

@jonsimon I agree. This is a weird market that, if interpreted in a technically rigorous way, I don't think could plausibly resolve YES. Would a market with a resolution criterion like, "Twitter makes publicly available at least 100 lines of code that it says represent its recommendation system," be better?

bought Ṁ100 of YES

I see the point that the question resolution is somewhat open to interpretation. I'm not sure how one could ask the question better about a code base that is not openly available atm.

@jonsimon I'm also not quite sure what you mean by it beeing an incomprehensibly massive web of systems. Isn't software always?

However 100 or even 10.000 lines of code will probably not nearly be enough to count as open source.
Clarification on the criterion from my point of view: For this to resolve to yes there needs to be a consensus that the released code base is making all the automated decisions about who gets to see what tweets. If there sometimes is a human in the loop for some situations that is still possible as long as the algorithm makes decisions 99% of the times when no human is involved

predicted YES

interesting ppl buying no ... not trusting in big E's words?

bought Ṁ50 of YES

@jcb but to price it at around 50% when he litterally says next week?

bought Ṁ15 of YES

@jcb Selection bias. He says a shit-ton and usually does what he says, even when crazy.

predicted YES

@jcb thank you I love those markets you linked although I dont come to the same conclusion as you

bought Ṁ3 of NO

@Dominik yes, even though he literally says "next week". I'd still put the true probability below 10%, maybe even below 5% (but not willing to risk enough Ṁ to push the market there).

Compare Fauci files. On Jan 1 those were promised "later this week", and they still haven't appeared. And releasing some emails or whatever was actually a plausible thing to happen. Publishing the relevant part of Twitter's codebase sounds (a) annoying, (b) potentially embarrassing (if the reported "power user multiplier" thing is true), and (c) of little to no value for Elon, Twitter, or anyone else. It's not like Elon cares about keeping his word for its own sake.

@Birger re selection bias: I just dug scrolled back through Elon's tweets until I spotted a concrete statement about the future with a time frame attached (god help me, why am I spending time on this). Here's the first one I found: https://twitter.com/elonmusk/status/1624009630778298370 I'm not an expert on the Twitter product but I can't find any indication that long tweet format changes were rolled out in the Feb 12-15 timeframe.

predicted YES

@jcb Not all tweets are a press statement, and timing is not that important. It is the intent of the tweet that is important, not a literal interpretation. This is why you thought his funding secret joke was serious. Different ways of communicating, I guess. But I see your point.

predicted NO

@Birger He recently testified in court that the "funding secured" tweet was "absolutely truthful"! Was that a joke too? Under penalty of perjury? Who knows, maybe so! But then why think that "our algorithm ... open source next week" is any different?

predicted YES

@jcb did not know that. an we will see on this i guess. and still, i do think there is a selection bias goin on here :D

predicted YES

@jcb "god help me why am I spending time on this?" I think we are spending time on trying to improve our collective information processing in a highly complex world. I believe there is something deeply relevant to this stuff.
Furthermore believe you are not quite sure about your opinion yourself otherwise you would spend more game money, but you want to be sure about this since this man has literally so much influence on all of our lives.
I know for sure I want to be more competent in understanding this stuff and there is no understanding if you cant predict in my view. I am somewhat of an elon fan but deeply scared of being too naive at the same time

predicted YES

@Dominik if I loose this market I will loose a great deal of trust. this stuff actually matters in my view. I dont care about next week but if he says next week and its not there by 2024 I have to rethink my worldview 4 sure

predicted NO

@Dominik Yeah agreed that there's value in improving our understanding and judgement about the world! That said I don't think the time I spent scrolling Elon's twitter timeline was particularly fruitful: it didn't do much to update my view, and it doesn't seem to have updated your views, and I don't think my view of Elon is particularly important in my overall understanding of the world. 🤷

For what it's worth, my assessment of Elon Musk is roughly:

  • says a lot of ideas out loud without much consideration or meaning behind it

  • follows through on a small portion of those (like, I dunno, 3%?)

  • some fraction of those are brilliant and executed effectively

  • others aren't

  • the good ideas bring in enough $ that Elon can basically shrug off the consequences of the bad ones

This "open source the algorithm" idea has been around long enough that maybe it's more than a passing impulse. But still, the overall truth rate of Elon's statements is low enough that I don't expect it to happen.


It's true that I'm not highly confident in my <10% estimate. But also there's counterparty risk when the market creator is on YES and doesn't have much of a Manifold resolution track record. (I don't have reason to believe you're dishonest! I just don't have much to go on here.)

OK, that's all the words I want to spend on Elon Musk today 😅

bought Ṁ20 of YES
bought Ṁ55 of YES

@Birger very interesting

predicted YES

hope its ok if i resolve to yes even if the algorithm isnt getting open sourced to github in directly since after open sourcing it it will be available on github or can easily be made available on github by anyone

bought Ṁ20 of YES

@Dominik i asumed as much