Thank you, Carlo, for the clear and direct questions. I'll try to answer them as honestly and transparently as I can, and again feel free to ask follow-ups if I miss anything.

> Why the stark earnings crash has not recovered and is still in full swing as late as today (Oct 25), and why does Medium keep saying it was fixed in the first week of October?

Here's a bit more detail into the timeline of the fraud incident and the fraud prevention measures that I hope will clear up some of this. Note, however, that some of this information is sensitive and has to be obscured to avoid giving fraudulent actors information that could help them get around current measures. But I will try my best to work around that and say when that is the case.

Mid-Sept through first week of Oct: We worked to identify, mitigate, and address the primary circle of fraudulent activity. After the first week of Oct, these accounts were no longer draining money from the writer earnings pool, so that's why we have been mentioning that date.

Mid-Sept, on-going: Part of the response to the fraud incident is that we updated our fraud prevention system to catch the loophole that they had exploited. A very high-level description of how this works is to make sure that any particular member's engagement can't result in author earnings that greatly exceed the cost of their membership. For example, if a member account reads, claps, follows, highlights, and replies to (hypothetically) 1,000 stories a day for 30 days (more than is generally humanly possible without a script or bot), we are taking measures to make sure that a $5 membership can't result in hundreds or thousands of dollars in author earnings. The engagements will still be seen by the writer, but those engagements will translate to significantly lower earnings, and the writer will still earn at the standard rate from all other members. We don't cap author earnings in any way.

As we investigated the fraud activity, we identified this very small percentage of membership accounts that were engaging with the identified fraudulent accounts' posts. Separately, they were also engaging with non-fraudulent accounts. When we reduced the impact of their engagements on earnings, this also ended up impacting other writers around the site for no fault of their own. Anyone who has seen spammy or AI replies or other engagement has likely had earnings that came from this small number of inhumanly-active accounts, and as we tuned this during the month of October, this has led to lowered earnings for writers that have done nothing wrong.

We're continuing to take steps to ensure that regular writers like yourself are not negatively impacted by this kind of thing as much as possible, by identifying this kind of behavior earlier. It's on-going, and the existence of this batch of inhumanly-active accounts is why some people have seen earnings dip through no fault of their own.

We released an update to this this week that should significantly reduce the impact of this behavior, and according to our metrics things do seem to be improved. Let us know if that is the case for you or not.

---

Okay, that was a lot, hopefully it helps to show how we are not only aware of the problem, but very concerned about the problem, and actively working to mitigate it as much as possible. We absolutely do not want fraudulent actors to cause problems for everyone else, and take this problem extremely seriously.

There should be very very few cases where earnings dropped 90% across the board for the same amount of activity... if this is happening I would be interested in looking into specific cases. Are you (or anyone else reading this) willing to volunteer as an account that we could talk about publicly and dig into specific stories, earnings, etc. I've looked into many specific accounts and instances of this and have found that this is far from being a widespread pattern, but every situation is different, and in some cases new information has been discovered.