hence the running joke: VPNs are an important privacy and security tool that can benefit every internet user, except for the hour or so a day I'm doing the spamfighting, during which period they are a scourge upon the planet and must be destroyed

And, to confirm a point that you made in your article today, Mike: if it's been too long without a false positive, it means I'm not being aggressive enough with the accounts that give me the bad vibes, and my spamfighting other half will have a lot of them turn up as false negatives in a week.

[Untimely death - Steve Sobol]

I’ve just learned of the untimely death of longtime #spam-fighter and web/app developer Steve Sobol. I've known him in the #spamfighting context since the 90s.

The only info I have is that and his 2 socials — FB and LI —which I was not connected to because I […]

Airtel flags 8 billion spam calls, 800 million spam SMSes
yespunjab.com?p=71143

#BhartiAirtel #SpamCalls #SpamSMS #AIPoweredSpamDetection #TRAI #TelecomRegulations #SpamFighting

The opacity around a lot of this came from spamfighting (as in, don’t reveal that you’ve caught the spammers and are limiting their reach). How would you navigate that balance?

Apropos of nothing, the thing about rate limits for spamfighting is, if you don’t also effectively stop bulk account signups, it sort of doesn’t matter what the rate limit is — the activity will just disperse across more accounts, each posting or DMing or whatever right up to the limit.

I'll also note that automated spamfighting is as much algorithmic as the promotion type stuff most people focus on, and platforms would simply be unusable without it (indeed, look at what's happening as Twitter actively gets worse at it)

We did ban these accounts. Tens of thousands of them. (I did a lot of this myself, because I’m a weirdo and find spamfighting relaxing.) They’re exceedingly persistent, because the incentives on their end ($$$ for selling accounts) are significant.

wait, sipped a decimal, 0.002%. That's individual artisinal handcrafted spam account removal, zero automated detection assistance. (My spamfighting other half does the human confirmation on the autodetects.)

I have been doing social media spamfighting for about 23 years now so I'm really experienced at spotting it, but I always encourage people to check a block list by loading, eg, 1 in every 10 accounts and seeing if they agree they should be on the list before subscribing!

That said, dealing with DM spam was never Twitter’s strongest suit. We operated under a… very strict… interpretation of when it was acceptable to look at the content of DMs (answer: hardly ever), which is good for privacy but bad for spamfighting.