One of the founders of startup accelerator Y Combinator offered unsparing criticism this weekend of the controversial data analytics company Palantir, leading a company executive to offer an extensive defense of Palantir’s work.

The back-and-forth came after federal filings showed that U.S. Immigration and Customs Enforcement (ICE) — tasked with carrying out the Trump administration’s aggressive deportation strategy — is paying Palantir $30 million to create what it’s calling the Immigration Lifecycle Operating System, or ImmigrationOS, to help ICE decide who to target for deportation, as well as offering “near real-time visibility” into self-deportations.

Y Combinator founder Paul Graham shared headlines about Palantir’s contract on X, writing, “It’s a very exciting time in tech right now. If you’re a first-rate programmer, there are a huge number of other places you can go work rather than at the company building the infrastructure of the police state.”

In response, Palantir’s global head of commercial Ted Mabrey wrote that he’s “looking forward to the next set of hires that decided to apply to Palantir after reading your post.”

Mabrey did not discuss the specifics of Palantir’s current work with ICE, but he said the company started working with the Department of Homeland Security (under which ICE operates) “in the immediate response to the murder of Agent Jaime Zapata by the Zetas in an effort dubbed Operation Fallen Hero.”

“When people are alive because of what you built, and others are dead because what you built was not yet good enough, you develop a very different perspective on the meaning of your work,” Mabrey said.

He also compared Graham’s criticism to protests over Google’s Project Maven in 2018, which eventually prompted the company to stop its work analyzing drone images for the military. (Google has subsequently signaled that it’s become more open to defense work again.)

Mabrey urged anyone interested in working for Palantir to read CEO Alexander Karp’s new book “The Technological Republic,” which argues that the software industry needs to rebuild its relationship with the government. (The company has been recruiting on college campuses with signs declaring that “a moment of reckoning has arrived for the West.”)

“We hire believers,” Mabrey continued. “Not in the sense of homogeneity of belief but in the intrinsic capacity to believe in something bigger than yourself. Belief is required because 1) our work is very, very hard and 2) you should expect to weather attacks like this all the time; from all sides of the political aisle.”

Graham then pressed Mabrey to “commit publicly on behalf of Palantir not to build things that help the government violate the US constitution,” though he acknowledged in another post that such a commitment would have “no legal force.”

“But I’m hoping that if they [make the commitment], and some Palantir employee is one day asked to do something illegal, he’ll say ‘I didn’t sign up for this’ and refuse,” Graham wrote.

Mabrey in turn compared Graham’s question to “the ‘will you promise to stop beating your wife’ court room parlor trick,” but he added that the company has “made this promise so many ways from Sunday,” starting with a commitment to “the 3500 enormously thoughtful people who are grinding only because they believe they are making the world a better place every single day as they see first hand what we are actually doing.”

“I wonder how much money OpenAI has lost in electricity costs from people saying ‘please’ and ‘thank you’ to their models.”

It was a seemingly random question posed by a user on X (formerly Twitter), but OpenAI CEO Sam Altman jumped in to reply that typing those words has added up to “tens of millions of dollars well spent — you never know.”

Judging from Altman’s tongue-in-cheek tone, it’s probably safe to assume he didn’t do a precise calculation. But his response prompted Futurism to speculate about whether it’s actually a waste of time and electricity to be polite to ChatGPT and other generative AI chatbots.

Apparently, being polite to AI isn’t just an unnecessary habit, misplaced anthropomorphism, or fear of our future computer overlords. instead, Kurt Beavers, a director on the design team for Microsoft Copilot, said that “using polite language sets a tone for the response,” and that when an AI model “clocks politeness, it’s more likely to be polite back.”

That said, profanity has its uses, too.

3The leaders of the House Committee on Energy and Commerce said they are investigating how 23andMe’s bankruptcy might affect customers’ data.

Representatives Brett Guthrie, Gus Bilirakis, and Gary Palmer (all Republicans) sent a letter Thursday to the genetic testing company’s interim CEO Joe Selsavage asking a number of questions about how 23andMe will handle customer data if the company is sold.

The letter also says that some customers have reported problems deleting their data from the 23andMe website, and it notes that direct-to-consumer companies like 23andMe are generally not covered by the protections of the Health Insurance Portability and Accountability Act (HIPAA).

“Given the lack of HIPAA protections, the patchwork of state laws covering genetic privacy, and the uncertainty surrounding what happens to customer information should a sale of a company or customer data and information transpire, we are concerned that this trove of sensitive information is at risk of being comprised,” the representatives write.

23andMe, which settled a data breach lawsuit for $30 million last year, filed for chapter 11 bankruptcy in March, with co-founder and CEO Anne Wojciki saying she was resigning to become a private bidder for the company.