Like many tech firms, Google ($GOOGL) is still riding the successes of decisions it made over a decade ago — things like buying YouTube, DoubleClick, and Android. These moves helped make the company a formidable data giant and, therefore, an advertising machine. They’ve also made up for innumerable missteps — Google+ didn’t become Facebook, its hardware still can’t compete with Apple, and its AI isn’t selling like OpenAI’s does. Really, its latest crowning achievements in product have been in the sheer volume of products Google kills — which might be why it’s embracing violence.

• Last week, Google deleted a pledge that said it wouldn’t use AI for weapons or surveillance, a commitment it made in 2018 as part of its “Responsible AI” principles.
• The decision follows in the footsteps of OpenAI, which made a similar change to its terms before partnering with weapons start-up Anduril on military-use AI.

It’s all about AI: Maybe Google just had a bad week, but it spent a lot of last week copying other companies. It announced plans to spend $75B on capital expenditures for AI this year (what tech firm isn’t doing that?). It then followed Meta in scrapping its already-useless DEI programs (Google’s workforce was already very unrepresentative of the US) and must have been so inspired by Palantir’s invigorating shareholder letter on violence and Western superiority that they decided, “Hey, that could be us, too!” But as Google banks on a mix of AI and government contracts to unlock a fresh future, success might be more a matter of whether it can commit to a product long enough — even more so than fixing its messy AI.