I believe the answer to all the rhetorical questions is reference to Friedman’s view that the only moral obligation corporations have is to shareholder value.
It is certainly reductivist, and I’d be delighted if the current moment shattered the consensus about that.
Just last week I wanted to reduce my Comcast service. Naturally it is impossible to do online, and I had to call and eventually get an agent.
Cancelling one sports package was easy enough. When I tried to ask about lower tiers of channel packages, I faced four different attempts to upsell into bigger packages, with more internet speed, and a contract.
Every time I asked a question, I got put on hold, presumably to strategize how to get me not to do what I wanted.
Finally I told him to cancel ALL the cable TV, and give the the straight rack rate for internet. Which was cheaper than anything he was willing to offer.
That took 40 minutes of my life away.
They cannot die soon enough.
"American movie studios will often censor their own films to appease foreign governments, even if those governments don’t threaten legal action."
I think a case can be made that behavior is as much discretion as censorship. To be sure, it's a liminal space.
In a conversation on the atproto slack, I kind of decided that the labeler needs to anonymize reports after authentication of valid used as a retention policy. That would make it impossible to work backwards from the label to the reports without a forward going court order, for which there could be a canary.
I was also enlightened that the labeler probably needs to put a multiple report threshold in place to mitigate malicious reports, but that seems flawed. You'll always need a way to escalate to human.
Actually, it seems I'm speed-running a lot of the moderation problem contemplating this, which is not a good sign for my project, or, I think small labelers at all.
Doing third-party labelers on a small scale seems as much challenged by these issues as at large and ridiculous scale, but with fewer resources. Can a hobby labeler afford to get involved with even a single complaint that s.230 should get tossed? That suggests almost certainly a person doing it would want to set up a corporate shield to run the labeler without putting personal assets at risk.
This makes the price of entry non-trivial. Set up a company, get hardware resources, setup the Ozone or compatible server and onboard users.
(I get here by wanting to be able to share block lists on Quora)
I'm at the beginning stages of a project that may try to use atproto and bluesky labellers to provide shared moderation to a completely different social platform.
The extension would be able to report things to a labeller, and get things from the labeller and block them on that platform, using the labeller as the database of reports and blocks.
Now, it appears the labeller has to be spun up somewhere, and is therefore providing a service to the extension users.
Questions arise, which seem true for Bluesky as well.
Suppose somebody feels improperly cancelled or libel/slandered by being on a block list kept by the labeller. Who do they sue -- the operator of the labeller, whoever pushed the button that made the report, or whatever person/thing handled the report and put it on the block list?
I expect there -will- be litigation on this as soon as third-party labellers get used and the aggrieved blossom as a result.
If the S.230 argument made here holds, the labeller is off the hook, and the person who pushed the report button is liable. Since the labelling system requires credentials, that person is probably possible to locate, unlike a certain Cow.
This would have a chilling affect on attempts to report potentially problematic things that people might want blocked. "Joe Blow reported us for being Nazis and we're not Nazis and don't want to be on a list of Nazis, Waaah!". Yes, Streisand would apply, but Joe Blow doesn't want to have to deal with it, and might chose not to report "Storm88" as a result.
If the company wants to protect it's IP for some period of time, it can compensate the person for the length of that period.
In the UK, it's called "gardening leave".
If the company isn't willing to do this, or litigate, then the IP in question isn't worth much, and they should let it go.
Not sure why you can't code the emulator in WASM and run it in a full screen browser, pulling whatever you want from wherever you want.
Perhaps someone will illuminate me.
The fixation on apps seems annoying.
I'm not naive, and I'm still having a very hard time with the moderation concepts here, and in Bluesky generally. None of the intros I've seen so far do the job.
For example, the first time we see "labeler" here is in a block diagram, and in the second use we're expected to know what that means and does.
What is a label? Is it the same thing as a "tag" in the stop-hand muted-words-and-tags description? Why are they different words, "label" and "tag"?
Is there a shared dictionary of labels, or is it unique to each labeller?
Where is the control panel for selecting labellers and labels to filter?
I accept that all this is rolling out, and that different apps and the web interface are at different points of development.
At the same time, I don't think it's usable yet.
Dominion is a business, for which the coin is the coin.
It was not realistic for people to hope they would act in righteous public interest and insist on repeated public on-air apologies. At best, we can hope going forward there is some restraint on Fox about using "Dominion" in these contexts.
Perhaps the other suits will get more on the admission front.
Also, given the predictability of a track, it's plausible to discard certain pixels at the predicted positions during transit while recording. The satellites aren't going to be in the field of view for the length of the long exposure.
Yes, this would result in more complicated software for astronomical cameras.
While there may be occlusion to Hubble from satellites, that isn't clearly related to those lower than its orbit.
The data cited don't really explain how that would come to pass, and it's rather central to the thesis.
Nor is there any discussion of how streaks damage the scientific value of the collected images. To what extent do streaks invalidate everything?
It's certainly fair to be concerned about ground-based observation, but I'm not finding the case very clearly made.
seems to be what obligation to put on sites once they are informed of problematic content.
There is reasonable dissatisfaction with "some stuff" staying around once there has been notification.
Now, of course, the problems then come into defining lines for "problematic content" and "reasonable notification", and what a reasonable response would be.
In discussions I have with people, the blanket protection 230 offers is a hard sell, and they Want To Do Something about doxxing (for example) and whatever they consider to be obvious lies.
I suspect that Something is going to be Done, and the challenge may be in finding the least problematic Something. I'm personally short of ideas, but I'd really like to see some honest exploration.
Is good one-touch ability to locate people of interest and search effectively.
This was the problem that google "solved" for the web better than AltaVista and Yahoo.
I don't think most people need algorithm produced feeds at all. Using a timeline of followed and a "trending" seems to be quite workable.
Now, there is certainly secret-sauce algorithm in search results and in picking trends, and a certain amount of secrecy is needed to avoid SEO-gamers from screwing things up.
What does seem viable to me would be having the search and trending functions be provided by third-parties that have first class access to the distributed platform. The frontend app stays the same, and you select your search and trending from those who seem to give you the results you find most useful. They can have some space for advertising. Some of these can be free-by-advertising, and some can subscribe for better ad-free results.
"Much of it only exists because the government failed to do its job."
I'm not at all sure I have any idea what any particular agency of the many governments in the vicinity should have been doing.
And Mike is kind of short on specific things he thinks monolithic Government should have been done as well.
I think we're agreed we don't want the Government much involved in regulating speech.
Based on the contents of my inboxes, I'm hard pressed to believe anybody is "better" at maintaining the integrity of their email lists or controlling their consultant/contractors.
But we must Do Something
or at least be Seen to be Doing Something. Think of the Children!
Milton Friedman
I believe the answer to all the rhetorical questions is reference to Friedman’s view that the only moral obligation corporations have is to shareholder value. It is certainly reductivist, and I’d be delighted if the current moment shattered the consensus about that.
Yes, flee Comcast
Just last week I wanted to reduce my Comcast service. Naturally it is impossible to do online, and I had to call and eventually get an agent. Cancelling one sports package was easy enough. When I tried to ask about lower tiers of channel packages, I faced four different attempts to upsell into bigger packages, with more internet speed, and a contract. Every time I asked a question, I got put on hold, presumably to strategize how to get me not to do what I wanted. Finally I told him to cancel ALL the cable TV, and give the the straight rack rate for internet. Which was cheaper than anything he was willing to offer. That took 40 minutes of my life away. They cannot die soon enough.
Yeah, Reynolds too
The usual trash outlets today are carrying leaks that Reynolds was mean on the Deadpool set. Search for "Ryan Reynolds' treatment of TJ Miller" Yawn.
movies...
"American movie studios will often censor their own films to appease foreign governments, even if those governments don’t threaten legal action." I think a case can be made that behavior is as much discretion as censorship. To be sure, it's a liminal space.
labeller vs. reporter and small labelers
In a conversation on the atproto slack, I kind of decided that the labeler needs to anonymize reports after authentication of valid used as a retention policy. That would make it impossible to work backwards from the label to the reports without a forward going court order, for which there could be a canary. I was also enlightened that the labeler probably needs to put a multiple report threshold in place to mitigate malicious reports, but that seems flawed. You'll always need a way to escalate to human. Actually, it seems I'm speed-running a lot of the moderation problem contemplating this, which is not a good sign for my project, or, I think small labelers at all. Doing third-party labelers on a small scale seems as much challenged by these issues as at large and ridiculous scale, but with fewer resources. Can a hobby labeler afford to get involved with even a single complaint that s.230 should get tossed? That suggests almost certainly a person doing it would want to set up a corporate shield to run the labeler without putting personal assets at risk. This makes the price of entry non-trivial. Set up a company, get hardware resources, setup the Ozone or compatible server and onboard users. (I get here by wanting to be able to share block lists on Quora)
Presents an interesting angle for Bluesky labellers
I'm at the beginning stages of a project that may try to use atproto and bluesky labellers to provide shared moderation to a completely different social platform. The extension would be able to report things to a labeller, and get things from the labeller and block them on that platform, using the labeller as the database of reports and blocks. Now, it appears the labeller has to be spun up somewhere, and is therefore providing a service to the extension users. Questions arise, which seem true for Bluesky as well. Suppose somebody feels improperly cancelled or libel/slandered by being on a block list kept by the labeller. Who do they sue -- the operator of the labeller, whoever pushed the button that made the report, or whatever person/thing handled the report and put it on the block list? I expect there -will- be litigation on this as soon as third-party labellers get used and the aggrieved blossom as a result. If the S.230 argument made here holds, the labeller is off the hook, and the person who pushed the report button is liable. Since the labelling system requires credentials, that person is probably possible to locate, unlike a certain Cow. This would have a chilling affect on attempts to report potentially problematic things that people might want blocked. "Joe Blow reported us for being Nazis and we're not Nazis and don't want to be on a list of Nazis, Waaah!". Yes, Streisand would apply, but Joe Blow doesn't want to have to deal with it, and might chose not to report "Storm88" as a result.
Gardening leave
If the company wants to protect it's IP for some period of time, it can compensate the person for the length of that period. In the UK, it's called "gardening leave". If the company isn't willing to do this, or litigate, then the IP in question isn't worth much, and they should let it go.
why use an app?
Not sure why you can't code the emulator in WASM and run it in a full screen browser, pulling whatever you want from wherever you want. Perhaps someone will illuminate me. The fixation on apps seems annoying.
This is still pretty opaque
I'm not naive, and I'm still having a very hard time with the moderation concepts here, and in Bluesky generally. None of the intros I've seen so far do the job. For example, the first time we see "labeler" here is in a block diagram, and in the second use we're expected to know what that means and does. What is a label? Is it the same thing as a "tag" in the stop-hand muted-words-and-tags description? Why are they different words, "label" and "tag"? Is there a shared dictionary of labels, or is it unique to each labeller? Where is the control panel for selecting labellers and labels to filter? I accept that all this is rolling out, and that different apps and the web interface are at different points of development. At the same time, I don't think it's usable yet.
But still on the books
Unless repealed, I fully expect cops to arrest people for violating it when it suits their mood.
Naive to expect otherwise
Dominion is a business, for which the coin is the coin. It was not realistic for people to hope they would act in righteous public interest and insist on repeated public on-air apologies. At best, we can hope going forward there is some restraint on Fox about using "Dominion" in these contexts. Perhaps the other suits will get more on the admission front.
Amenable to signal processing
Also, given the predictability of a track, it's plausible to discard certain pixels at the predicted positions during transit while recording. The satellites aren't going to be in the field of view for the length of the long exposure. Yes, this would result in more complicated software for astronomical cameras.
How to satellites below Hubble orbit interfere with it?
While there may be occlusion to Hubble from satellites, that isn't clearly related to those lower than its orbit. The data cited don't really explain how that would come to pass, and it's rather central to the thesis. Nor is there any discussion of how streaks damage the scientific value of the collected images. To what extent do streaks invalidate everything? It's certainly fair to be concerned about ground-based observation, but I'm not finding the case very clearly made.
The problem to address
seems to be what obligation to put on sites once they are informed of problematic content. There is reasonable dissatisfaction with "some stuff" staying around once there has been notification. Now, of course, the problems then come into defining lines for "problematic content" and "reasonable notification", and what a reasonable response would be. In discussions I have with people, the blanket protection 230 offers is a hard sell, and they Want To Do Something about doxxing (for example) and whatever they consider to be obvious lies. I suspect that Something is going to be Done, and the challenge may be in finding the least problematic Something. I'm personally short of ideas, but I'd really like to see some honest exploration.
And to realize
that he did this on National Brotherhood Week!
What's missing from decentalized...
Is good one-touch ability to locate people of interest and search effectively. This was the problem that google "solved" for the web better than AltaVista and Yahoo. I don't think most people need algorithm produced feeds at all. Using a timeline of followed and a "trending" seems to be quite workable. Now, there is certainly secret-sauce algorithm in search results and in picking trends, and a certain amount of secrecy is needed to avoid SEO-gamers from screwing things up. What does seem viable to me would be having the search and trending functions be provided by third-parties that have first class access to the distributed platform. The frontend app stays the same, and you select your search and trending from those who seem to give you the results you find most useful. They can have some space for advertising. Some of these can be free-by-advertising, and some can subscribe for better ad-free results.
blame taker
Whoever gets that job will be well paid for taking that blame, and safely parachute away. Hell, I'd take it and bail after a couple of years.
What should "government" have done?
"Much of it only exists because the government failed to do its job." I'm not at all sure I have any idea what any particular agency of the many governments in the vicinity should have been doing. And Mike is kind of short on specific things he thinks monolithic Government should have been done as well. I think we're agreed we don't want the Government much involved in regulating speech.
Oh really?
Based on the contents of my inboxes, I'm hard pressed to believe anybody is "better" at maintaining the integrity of their email lists or controlling their consultant/contractors.