Rantburg

Today's Front Page   View All of Sun 09/27/2026 View Sat 09/26/2026 View Fri 09/25/2026 View Thu 09/24/2026 View Wed 09/23/2026 View Tue 09/22/2026 View Mon 09/21/2026
2026-09-27 Science & Technology
The Culture of Executioners
Direct Translation via Google Translate. Edited

Text taken from the Telegram channel of https://t.me/darpaandcia

Commentary by Russian military journalist Boris Rozhin ins in italics.

[ColonelCassad] SKYNET and Palantir: A Succession of Executioners Former NSA and CIA Director Michael Hayden once said, "We kill people based on metadata."

This phrase was uttered in April 2014 at a public debate at Johns Hopkins University. Hayden's opponent was Georgetown University law professor David Cole. During the debate, Cole quoted former NSA General Counsel Stuart Baker: "Metadata will tell you absolutely everything about someone's life. If you have enough metadata, you don't really need the content." Hayden agreed, calling the statement "absolutely true," and added his own cynical and cruel remark. This statement directly echoes the secret NSA program called SKYNET, revealed in documents leaked by Edward Snowden.

The program revolved around data collection: the NSA collected metadata (who called whom, when, and from where) from 55 million mobile phone users in Pakistan. This data was then fed to a machine learning algorithm that looked for patterns similar to the behavior of known "couriers" for terrorist groups—people who hand-deliver messages to avoid interception. The program was supposed to generate a probabilistic estimate of how similar each of the 55 million people was to a courier. If the algorithm was wrong, then anyone could be targeted by a drone strike.

Since 2004, drone strikes have killed between 2,500 and 4,000 people in Pakistan. And according to NSA documents, the system mistakenly classified ordinary people as terrorists 0.18% of the time. With a sample size of 55 million people, this means tens of thousands of innocent people caught in the crosshairs as potential targets. The most striking example of the program's failure was the highest threat rating given to Ahmad Zaidan, Al Jazeera's Islamabad bureau chief.

In an NSA presentation, his photo was tagged as an "al-Qaeda member." This was because Zaidan, as a journalist, frequently traveled to dangerous regions and interacted with various people, which the algorithm mistakenly identified as a courier. Much time has passed since then. Now the Americans have Palantir and powerful neural networks that allow them to even more accurately identify potential "targets" for strikes. But judging by recent events—targeted killings in Iran and the Gaza Strip—civilian casualties are not decreasing. It is not for nothing that even American human rights activists see in this a “culture of executioners”, which has taken root at the top of the US intelligence community.

Palantir is the direct heir to the SKYNET logic: while the NSA collected metadata and transmitted it to the military, Palantir commercialized this model and sold it to Israel as a finished product. A private contractor now performs a function that was previously the preserve of state intelligence—and does so without the public reporting formally required of government agencies. This is why the "culture of executioners" isn't a thing of the past, but is being commercially continued.

Of course, modern neural network technologies will primarily be used for surveillance and control over people, as well as for the assassination of others.To think that it will be any different is as naive as believing that Starlink was launched to provide affordable internet for ordinary people.

# id) ?>
Posted by badanov 2026-09-27 00:00|| E-Mail|| Front Page|| ||Comments [55 views ]  Top

1 Human vs. ChatGPT NFL Week 3 betting picks: can AI recover from its ugly 2-8 start to the season?
Posted by Skidmark 2026-09-27 10:27||   2026-09-27 10:27|| Front Page || Comments   Top

15:48 NoMoreBS
15:48 no mo uro
15:43 trailing wife
15:23 trailing wife
14:27 Chantry
14:14 Mercutio
14:09 Mercutio
13:33 NN2N1
12:48 Besoeker
12:12 Lionel the Lesser1066
12:07 Lionel the Lesser1066
12:06 Remoteman
12:06 MikeKozlowski
11:51 Herb Panda7448
11:37 Frank G
11:26 Difar Dave
11:18 Penguin of the Best Desert
11:14  chris
11:10 Skidmark
10:59 Skidmark
10:53 Skidmark
10:51 Skidmark
10:34 Skidmark
10:31 trailing wife
Merry-Go-Blog










Paypal:
Google
Search WWW Search rantburg.com