pull down to refresh

Yeah. 90% of it my money, at 3x disadvantage.

177 sats \ 21 replies \ @Scoresby 7h

some possible alternatives to the current system:

  • make boosts equal in weight to downzaps (because we know in each case that sats don't end up with the person doing the action)
  • make downzaps equal to zaps in weighting
  • make it so that downzaps can never reduce an item below 0

as an interesting sidenote: if you use an alt account to zap yourself in a territory you own, I believe you collect ~90% of the zap. So, at least for posts you make, you are not at a 3x disadvantage.

This is also the reason why zaps only count for a third as much as downzaps. SN has no way of knowing when a user is using a sock puppet account to zap themselves.

reply
214 sats \ 13 replies \ @optimism 7h

Yes but we're talking about @SimpleStacker's post here, not mine. And yesterday it was @denlillaapan's.

reply
267 sats \ 12 replies \ @Scoresby 6h

it is a problem for sure. It'd be bad though to expose SN to a new problem in an attempt to fix this one.

Back in August there was that scammy pig.gy site that tried boosting their post to the top of the feed (#1090193). They boosted it by 160k sats and it got 20k in downzaps and was outlawed. There is no more outlawing, but if (when) downzaps = zaps, such posts will be more difficult to bury. At the moment that seems less bad than the current downzap fest. But I wonder what will happen if we have someone who is really committed to boosting shitty content.

reply

Is boosting bad content as much of a problem as downzapping good content?

Not really sure. But I think we have more tools for ignoring bad content than we do for finding good content that doesn't get surfaced.

reply

Trust did some of that work for us. Perhaps we're seeing the downside of eliminating trust.

reply
253 sats \ 4 replies \ @optimism 2h

I thought about this for a bit but I don't think we can tell. Because with the trust removal a whole lot of other things had to change, it's the entire package in it's current form that doesn't work - that's measurable.

Trust act like a blanket. It dampens the noise and keeps some of us warm, but it doesn't extend to everyone, so some will sit in the cold, especially newcomers, because trust gets built. Without trust the playing field for newcomers vs oldtimers is level, but we now found out that there aren't enough sats circling around to repel attacks at this scale, at least not at 3x disadvantage. More sats were spent to defend some good posts up than on the downzaps, but it wasn't enough to keep the posts visible. Defenders have lost 2 rounds now...

reply
Defenders have lost 2 rounds now

We're still in the early stages though. The rewards are going to be huge today.

If we're in a high downzap environment, then we are also in a high rewards environment and equilibrium zaps will be higher. We don't know yet if that equilibrium will be higher or lower signal.

reply
253 sats \ 2 replies \ @optimism 2h
We're still in the early stages though. The rewards are going to be huge today.

I've spent 200k on defense today and fished 60k in rewards that way for everyone - so I guess you're all welcome haha. Defense is a huge expenditure, but gotta take a stand sometimes, even if it hurts. Lots of insight was gained; that's the real value here.

oops

reply
357 sats \ 1 reply \ @optimism 6h

I'd say boosting bad content = scammer heaven = money loss.
Downzapping good content = morale loss.

good and bad are subjective, but the balance is really, really hard. As we've found out.

reply
151 sats \ 0 replies \ @Scoresby 6h

yes, the point about subjectivity here is important. It reminds me that it's not just a simple gradient, either. it's like trying to balance a lot of different sticks on a single point. And some of this stuff only shows up when someone tries it.

reply
236 sats \ 0 replies \ @Scoresby 6h

until the last week, I'd say boosting bad content was a worse problem.

I think we have more tools for ignoring bad content than we do for finding good content that doesn't get surfaced.

This feels like it is true though. It is more frustrating when you can't find good content than when you see bad content.

reply
102 sats \ 0 replies \ @optimism 6h
reply
286 sats \ 3 replies \ @k00b 7h
make downzaps equal to zaps in weighting

My PR does this. I think it's the right balance between sybil and activism risk.

I'll ship this weekend (fingers crossed) in a pork barrel deployment of ranking changes.

reply

I can live with the extra expense of correcting zaps to slop

reply
152 sats \ 1 reply \ @k00b 6h

When you consider visibility in the (bah need better jargon) game theory, I think it restores balance.

In contrast to downzapped things, sybil zapped items are more visible so more folks will act on them (maybe even 3x more people will act on them).

reply
297 sats \ 0 replies \ @optimism 6h

Yes. But there is no reason to sybil zap if there is boost that is even more effective than sybil-upzapping your own post on your own territory. So from an optimization pov, why would anyone do the trouble of sybil zaps? Just for appearance sake?

We know that there are some stackers with sybil accounts (because they admit it) for their bots but I haven't seen much abuse there. So from the policing the ad-bots perspective, like the L402 bot we have since spamming their own L402 endpoint in... 11 posts thus far (just saw they also spam ~bitcoin, ugh). That's honey-potting stackers and it will enshittify SN if it makes sats - thus my need for making comments expensive (but probably too much), as I see that as a lesser evil than moderation with a button that has no cost.

Either way, right now SN feels vulnerable. We have good discussions downzapped into oblivion over stupid horses and guns

reply
sock puppet

<3

reply
141 sats \ 1 reply \ @optimism 6h

I'm apparently your sock puppet. But can I please be your cumsockpuppet? I need the protein.

reply

ouhw, feisty

reply

It was working fine until someone decided to go activist.

But it has revealed a new attack surface and vulnerability.

Not sure what the solution is. Brain too tired.

reply
146 sats \ 1 reply \ @Scoresby 5h

this is how SN gets better. we gotta see where the weaknesses are.

reply
39 sats \ 0 replies \ @optimism 4h

Some days better, other days worse, I guess.

reply