I don’t know if I have any “too late” line I would consider for “ongoing issues” — sure it would have been nice to get ahead of things instead, but at least it stems the potential future damage
I don’t know if I have any “too late” line I would consider for “ongoing issues” — sure it would have been nice to get ahead of things instead, but at least it stems the potential future damage
Too late as in that I seriously doubt it will have any effect before February 23. But you're right, that's not the only case where it will matter.
The move follows complaints from politicians in Germany that X’s algorithm is promoting content by the far right ahead of the country’s February 23 elections
Too late as in that I seriously doubt it will have any effect before February 23. But you're right, that's not the only case where it will matter.
Ah, yeah I agree with that, it’s definitely past the point of having had an influence. Fingers crossed it can reduce the impact on the next election though!
Ah, yeah I agree with that, it’s definitely past the point of having had an influence. Fingers crossed it can reduce the impact on the next election though!
Alternatively: even if they're honest and make a reasonable effort to comply, what's to guarantee the result will be useful? "Here's a folder full of Word docs from when we designed the...
Alternatively: even if they're honest and make a reasonable effort to comply, what's to guarantee the result will be useful?
"Here's a folder full of Word docs from when we designed the thing—though we did change a few things at implementation time. Here's snapshots of a few Git repos. Here's PDFs of some internal wiki articles from last year. And here's the AI model weights (we didn't hold on to all the training data, sorry)."
I don't know the exact scope of what X is legally required to provide. But whatever they hand over, it's going to be messy. It won't be a self-contained box with little brass gears on the inside and "THE ALGORITHM" emblazoned on the outside.
Ouch, I can see this and it sounds painful. Thousands of pages, some just marginally related garbage and a lot of painful to read musk communications. Worse, what are the chances he prints them...
Ouch, I can see this and it sounds painful. Thousands of pages, some just marginally related garbage and a lot of painful to read musk communications. Worse, what are the chances he prints them all? "Imagine their faces when a truck load of banker boxes shows up."
They'd probably need to hire someone who knows what they're even looking at to have a chance unless there's some convenient memo from Musk telling the programmers to give his account preference.
Vivendi tried flooding them with millions of pages of documents, all in Korean, and an intern just happened to find the one that showed Vivendi was destroying documents related to the case and they won.
Ha! That's awesome, they probably made the same person put together the documents as the one who was ordered to destroy some so I wonder if it was put in there on purpose just to spite their boss?
Ha! That's awesome, they probably made the same person put together the documents as the one who was ordered to destroy some so I wonder if it was put in there on purpose just to spite their boss?
Been trying to figure out the thinking behind this type of demand. Mostly because it seems arbitrary compared to requesting source control and changes logs. But elon (and other tech bros) are...
Been trying to figure out the thinking behind this type of demand. Mostly because it seems arbitrary compared to requesting source control and changes logs. But elon (and other tech bros) are playing dangerous games by being an existential threat to entire nations. He always swings his money around like a giant hammer and he's letting slip that he's a one trick pony. There are ways to work around even that and this is a first step.
I suspect the EU may be considering their own type of platform specific ban, but are going about it a bit smarter than the US.
The difference between documents and code is intent. If it's found the system drives people to extreme far right pipelines, it can be argued that it's a natural outcome on how the algorithm drives engagement. And changing it would be anti-free speech.
But the documentation would make for an interesting read. And I'm sure Elon pissed enough former staff and industry vets who can point out exactly what is from before and after the takeover.
With Twitter run by a crew of skeletons on fumes now, I don't think there are enough people to cover tracks or have things look plausibly deniable.
So if Elon sent some email demandig the entire algorithm be skewed to his own activities or develop a console to manipulate topics:
It was probably mentioned on a call or email. It probably got turned into a technical specification. That spec would be on a ticket. That ticket was fulfilled by someone. That person pushed it with comments and it went into production at some point.
Heck, even just emails calling for people to be boosted or banned would signal that Twitter is not run on rules and its terms of service, but rather by the arbitrary whims of its owner.
With that you could argue that if Elon incites targeted violence on his platform or facilitates some criminal activity, there is nothing to stop it from reaching every user on the continent. And even a 1% activation would mean tens of thousands of acts of violence.
Even if they covered all that up, or he runs the company like a high school project with no documentation; someone must have discussed it. There's probably an objection or two or at least even slack messages. Paper trails are difficult to fully manipulate because there is so much paper to work through.
Also, European countries have very different standards regarding free speech than the United States does. Look at how Germany criminalized Nazi salutes. Also Elon is a US citizen and Twitter/X is...
Also, European countries have very different standards regarding free speech than the United States does. Look at how Germany criminalized Nazi salutes. Also Elon is a US citizen and Twitter/X is headquartered here. They are guests abroad and will have fewer rights.
I think its good that this is being investigated, but i fear that it will be both toothless and too late.
Oh it's going to be 100% toothless until countries decide to actually ban software.
I don’t know if I have any “too late” line I would consider for “ongoing issues” — sure it would have been nice to get ahead of things instead, but at least it stems the potential future damage
Too late as in that I seriously doubt it will have any effect before February 23. But you're right, that's not the only case where it will matter.
Ah, yeah I agree with that, it’s definitely past the point of having had an influence. Fingers crossed it can reduce the impact on the next election though!
I can't help but wonder, what's to stop them from altering it before handing it over?
Alternatively: even if they're honest and make a reasonable effort to comply, what's to guarantee the result will be useful?
"Here's a folder full of Word docs from when we designed the thing—though we did change a few things at implementation time. Here's snapshots of a few Git repos. Here's PDFs of some internal wiki articles from last year. And here's the AI model weights (we didn't hold on to all the training data, sorry)."
I don't know the exact scope of what X is legally required to provide. But whatever they hand over, it's going to be messy. It won't be a self-contained box with little brass gears on the inside and "THE ALGORITHM" emblazoned on the outside.
Ouch, I can see this and it sounds painful. Thousands of pages, some just marginally related garbage and a lot of painful to read musk communications. Worse, what are the chances he prints them all? "Imagine their faces when a truck load of banker boxes shows up."
They'd probably need to hire someone who knows what they're even looking at to have a chance unless there's some convenient memo from Musk telling the programmers to give his account preference.
Funnily enough that worked out exactly well for Valve when they sued Vivendi for distributing Counter-Strike to Korean internet cafes: https://www.pcgamer.com/games/fps/a-summer-intern-once-saved-valve-from-a-near-fatal-lawsuit-after-a-publisher-decided-to-go-world-war-3-on-it-and-it-all-hinged-on-one-email/
Vivendi tried flooding them with millions of pages of documents, all in Korean, and an intern just happened to find the one that showed Vivendi was destroying documents related to the case and they won.
Ha! That's awesome, they probably made the same person put together the documents as the one who was ordered to destroy some so I wonder if it was put in there on purpose just to spite their boss?
Document review is a necessary part of litigation and these things are tedious but they get handled.
Been trying to figure out the thinking behind this type of demand. Mostly because it seems arbitrary compared to requesting source control and changes logs. But elon (and other tech bros) are playing dangerous games by being an existential threat to entire nations. He always swings his money around like a giant hammer and he's letting slip that he's a one trick pony. There are ways to work around even that and this is a first step.
I suspect the EU may be considering their own type of platform specific ban, but are going about it a bit smarter than the US.
The difference between documents and code is intent. If it's found the system drives people to extreme far right pipelines, it can be argued that it's a natural outcome on how the algorithm drives engagement. And changing it would be anti-free speech.
But the documentation would make for an interesting read. And I'm sure Elon pissed enough former staff and industry vets who can point out exactly what is from before and after the takeover.
With Twitter run by a crew of skeletons on fumes now, I don't think there are enough people to cover tracks or have things look plausibly deniable.
So if Elon sent some email demandig the entire algorithm be skewed to his own activities or develop a console to manipulate topics:
It was probably mentioned on a call or email. It probably got turned into a technical specification. That spec would be on a ticket. That ticket was fulfilled by someone. That person pushed it with comments and it went into production at some point.
Heck, even just emails calling for people to be boosted or banned would signal that Twitter is not run on rules and its terms of service, but rather by the arbitrary whims of its owner.
With that you could argue that if Elon incites targeted violence on his platform or facilitates some criminal activity, there is nothing to stop it from reaching every user on the continent. And even a 1% activation would mean tens of thousands of acts of violence.
Even if they covered all that up, or he runs the company like a high school project with no documentation; someone must have discussed it. There's probably an objection or two or at least even slack messages. Paper trails are difficult to fully manipulate because there is so much paper to work through.
Also, European countries have very different standards regarding free speech than the United States does. Look at how Germany criminalized Nazi salutes. Also Elon is a US citizen and Twitter/X is headquartered here. They are guests abroad and will have fewer rights.
Why now, and why X