Facebook alone takes down 1B pieces of (1st Amend protected) spam a quarter. The political content moderation decisions are outnumbered by those necessary to make platforms usable by at least 100:1.

The “1st Amendment for Platforms” people never have relevant experience.
Zero-marginal-cost communications to millions of people wasn’t exactly foreseeable in 1791.
2nd Amendment arguments often revolve around the fact that firearms are vastly more dangerous now, but at least modern rifles can exist within the conceptual frame of 18th century science (with amazing materials).

Modern communication technologies are effectively witchcraft.
Show up to Independence Hall with an M4: “What an amazing and curious rifled musket you have there sir!”

Demo Twitter running on an iPhone: “Perhaps the good gentlemen from Massachusetts had the right idea with the burnings.”

• • •

Missing some Tweet in this thread? You can try to force a refresh

Keep Current with Alex Stamos

Alex Stamos Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!


Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @alexstamos

24 Apr
I'm looking at a wall of con badges and realizing that most of them have a @dakami story attached. He wasn't only brilliant, but an incredibly generous colleague and friend. No matter what news he was making, he would have an encouraging word for and talk up other researchers.
"Larger than life" is a term that is thrown around too much, but the guy had a massive impact on others and it's difficult to imagine the last twenty years of infosec without him.
Thanks to the way he treated others, it was hard not to love Dan. In summer 2008, everybody wanted a picture (no selfies yet) with Fake Dan Kaminsky. The next summer, he faced incredible adversity but still stood tall and got a hug from half of Vegas.

Read 5 tweets
23 Mar
If Congress wants useful hearings they should call the:

1) Product VP of Integrity/Trust and Safety
2) VP of Platform Policy

...from Facebook, Twitter, YouTube and one upstart (maybe TikTok). Those are the people at the trade-off coal face.
Instead we get hours of trying to score imaginary C-Span points by tripping up a CEO on decisions made three levels down.

To make it really spicy, ask those eight witnesses to come with prevalence, precision and recall numbers on their efforts to stop four major types of abuse.
My fantasy hearing: "I'm looking at the confusion matrix for your COVID misinfo classifier and it feels like you made a politically motivated decision to back off of recall to not anger certain political groups. Did I get that right?"
Read 5 tweets
3 Mar
We have published the final report of the Election Integrity Partnership!

I am so proud of the team that came together to track election disinformation in real-time, archive examples and metadata, mitigate harm via partnerships and, finally, to provide this report.
With 283 pages, there is something for everyone. You can read about the structure of the partnership in Ch 1, a statistical overview in Ch 2, the evolution of narratives in Ch 3, the structure and actors of disinfo in Ch 4-5, policy analysis in Ch 6 and recommendations in Ch 7.
Make sure to join us tomorrow for a webcast discussing our findings and lessons for government, media and platforms.

Read 4 tweets
25 Feb
This is a spectacularly bad article and @thedailybeast should be ashamed. It completely lacks an understanding of automated versus manual reporting of CSAM and creates disincentives for more companies to aggressively scan.

The problem isn't the companies with billions of users and millions of reports, but those will billions of users and dozens or hundreds of reports. Those tiny numbers indicate providers who are not proactively scanning using known CSAM hash banks.
There is a big battle right now inside of companies to try to find a way to balance child safety protections with privacy. Creating disincentives for voluntary reporting pushes us in the wrong direction.
Read 7 tweets
16 Feb
In case you missed it on Friday evening, our team at @stanfordio published an analysis of how Clubhouse handles audio streams, finding serious risks related to data and the People's Republic of China.


Simplified Chinese:
We found Chinese servers being used even for conversations that only involved Americans.

At this time, I can't recommend that individuals who might find themselves adverse to the security services of the PRC to use Clubhouse for sensitive conversations.
Clubhouse provided us with the following statement promising changes. We found that the use of Shanghai-based Agora is fundamental to the function of the app and building logical and technical controls between the US and PRC infrastructure will be extremely complicated.
Read 4 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!