Palantir is often called a data broker, a data miner, or a giant database of personal information. In reality, it’s none of these—but even former employees struggle to explain it.

Palantir sends its employees to work inside client organizations essentially as consultants, helping to customize their data pipelines, troubleshoot problems, and fix bugs. It calls these workers “forward deployed software engineers,” a term that appears to be inspired by the concept of forward-deployed troops, who are stationed in adversarial regions to deter nearby enemies from attacking.

Crucially, Palantir doesn’t reorganize a company's bins and pipes, so to speak, meaning it doesn’t change how data is collected or how it moves through the guts of an organization. Instead, its software sits on top of a customer’s messy systems and allows them to integrate and analyze data without needing to fix the underlying architecture. In some ways, it’s a technical band-aid. In theory, this makes Palantir particularly well suited for government agencies that may use state-of-the-art software cobbled together with programming languages dating back to the 1960s.

Palantir’s software is designed with nontechnical users in mind. Rather than relying on specialized technical teams to parse and analyze data, Palantir allows people across an organization to get insights, sometimes without writing a single line of code. All they need to do is log into one of Palantir’s two primary platforms: Foundry, for commercial users, or Gotham, for law enforcement and government users.

Foundry focuses on helping businesses use data to do things like manage inventory, monitor factory lines, and track orders. Gotham, meanwhile, is an investigative tool specifically for police and government clients, designed to connect people, places, and events of interest to law enforcement. There’s also Apollo, which is like a control panel for shipping automatic software updates to Foundry or Gotham, and the Artificial Intelligence Platform, a suite of AI-powered tools that can be integrated into Gotham or Foundry.

Foundry and Gotham are similar: Both ingest data and give people a neat platform to work with it. The main difference between them is what data they’re ingesting. Gotham takes any data that government or law enforcement customers may have, including things like crime reports, booking logs, or information they collected by subpoenaing a social media company. Gotham then extracts every person, place, and detail that might be relevant. Customers need to already have the data they want to work with—Palantir itself does not provide any.

Foundry and Gotham are similar: Both ingest data and give people a neat platform to work with it. The main difference between them is what data they’re ingesting. Gotham takes any data that government or law enforcement customers may have, including things like crime reports, booking logs, or information they collected by subpoenaing a social media company. Gotham then extracts every person, place, and detail that might be relevant. Customers need to already have the data they want to work with—Palantir itself does not provide any.

Since leaving Palantir, Pinto says he’s spent a lot of time reflecting on the company’s ability to parse and connect vast amounts of data. He’s now deeply worried that an authoritarian state could use this power to “tell any narrative they want” about, say, immigrants or dissidents it may be seeking to arrest or deport. He says that software like Palantir’s doesn’t eliminate human bias.

People are the ones that choose how to work with data, what questions to ask about it, and what conclusions to draw. Their choices could have positive outcomes, like ensuring enough Covid-19 vaccines are delivered to vulnerable areas. They could also have devastating ones, like launching a deadly airstrike, or deporting someone.

In some ways, Palantir can be seen as an amplifier of people’s intentions and biases. It helps them make evermore precise and intentional decisions, for better or for worse. But this may not always be obvious to Palantir’s users. They may only experience a sophisticated platform, sold to them using the vocabulary of warfare and hegemony. It may feel as if objective conclusions are flowing naturally from the data. When Gotham users connect disparate pieces of information about a person, it could seem like they are reading their whole life story, rather than just a slice of it.

I've been working and testing to switch my main PC (used for work like audio recording, music, and general multimedia) and have been playing with Ubuntu Studio on my laptop. Loving it so far but I keep seeing people talk about CachyOS, Bazzite, or the new Debian Trixie.

I'm having trouble finding what's really different about all these distros aside from how they look or slight changes in how they do things (I know Ubuntu Studio has a low latency kernel which seems important for what I need to do). Is there a big difference? Like, if I go with Ubuntu Studio am I gonna end up wiping everything and installing CachyOS or Bazzite or something in a month because it's better? Or are all these distros basically the same thing with a different look and feel and as long as I choose one that gets regular updates, it doesn't matter fundamentally?

I'm trying to grasp the Linux concept but being a Windows user my whole life I'm struggling to 'get it'. Instead of trying to understand in the contex of Windows or Mac, is a better comparison Apple/Android? Like iPhones would be similar to both Mac and Windows (you don't get to choose much) and Android would be Linux (I know it's built on it haha) and it's really just a bunch of different options to do the same thing?

Between fall 2019 and fall 2023, public school enrollment fell from 50.8 million to 49.5 million, a loss of more than 1.2 million students, or 2.5%. The pandemic accelerated that decline, but enrollment was already falling in some grades and communities before COVID, and that trend is expected to continue: The National Center for Education Statistics (NCES) projects overall enrollment will fall below 47 million by 2031.

Yet a FutureEd analysis based on data from NCES, the Center for Disease Control and Prevention and other sources, finds that enrollment trends vary widely by race, grade level, geography and schools. That variation offers important insight into how the education landscape is shifting, where the challenges are most acute and what it means for the future of the public school system.

When COVID struck and public schools switched to virtual learning, some families began educating their children at home; others opted for private schools, many of which resumed in-person learning sooner than public schools.

Though many families eventually returned, by fall 2022, an estimated 3.4% of American children were homeschooled, compared with 2.8% in fall 2019 — and advocates say the number continues to rise. Private school enrollment also grew, reaching roughly 7 million in 2021, an estimated 22% increase over pre-pandemic levels. Newly expanded school choice policies are likely to accelerate that shift: Sixteen states offer or intend to offer public funding for private school tuition to any student in the state. While these programs haven’t yet triggered a mass public school exodus, they are likely to reshape some family preferences and expand the supply of private alternatives.

Still, family choices aren’t the sole or even primary driver of the decline. The U.S. birthrate has fallen steadily for more than a decade and is now —at just under 1.6 — the lowest in history. Between 2010 and 2020, the number of children under 5 fell by 1.8 million nationwide, from 20.2 million to 18.4 million. That shrinking pipeline is now reaching classrooms.

Better social media is possible.

In his new book Move Slowly and Build Bridges, Robert W. Gehl tells the story of the activists, software developers, artists, and everyday people who have built the fediverse, a noncentralized alternative social media system. Unlike big tech corporations like Facebook, TikTok, or X, the fediverse is comprised of thousands of small, independent communities who use a Web protocol to communicate with one another.

These small communities govern themselves and moderate content at the human scale — compare that to Facebook and X, which try to moderate global conversations. And the fediverse isn’t built in order to gather user data and sell attention to marketers — it’s a more privacy-respecting social media alternative.

The most notable part of the fediverse is Mastodon. Founded in 2016, Mastodon was positioned as an alternative to Twitter. Like Twitter (or X), Mastodon members can post, like, share, and connect with one another across the world. Unlike Twitter/X, Mastodon can be completely under the control of its members, from how it’s run to its underlying software.

Making a noncentralized, ethically run social media system isn’t easy. The people building the fediverse have faced long hours, burnout, angry debates, and, worst of all, bigotry, death threats, and discrimination. They face constant, nagging doubts: Can we really do this? Can noncentralized social media survive in a world that is used to corporate social media? Can we—all of us—have our own social media?

As Move Slowly and Build Bridges shows us, the answer is yes, but it’s going to take a struggle.

Better social media is possible.

In his new book Move Slowly and Build Bridges, Robert W. Gehl tells the story of the activists, software developers, artists, and everyday people who have built the fediverse, a noncentralized alternative social media system. Unlike big tech corporations like Facebook, TikTok, or X, the fediverse is comprised of thousands of small, independent communities who use a Web protocol to communicate with one another.

These small communities govern themselves and moderate content at the human scale — compare that to Facebook and X, which try to moderate global conversations. And the fediverse isn’t built in order to gather user data and sell attention to marketers — it’s a more privacy-respecting social media alternative.

The most notable part of the fediverse is Mastodon. Founded in 2016, Mastodon was positioned as an alternative to Twitter. Like Twitter (or X), Mastodon members can post, like, share, and connect with one another across the world. Unlike Twitter/X, Mastodon can be completely under the control of its members, from how it’s run to its underlying software.

Making a noncentralized, ethically run social media system isn’t easy. The people building the fediverse have faced long hours, burnout, angry debates, and, worst of all, bigotry, death threats, and discrimination. They face constant, nagging doubts: Can we really do this? Can noncentralized social media survive in a world that is used to corporate social media? Can we—all of us—have our own social media?

As Move Slowly and Build Bridges shows us, the answer is yes, but it’s going to take a struggle.