Palantir published a 22 point manifesto defending military AI, dismissing pluralism as hollow, and calling some cultures 'regressive and harmful.' It lands as Congress presses the company on ICE contracts. Bellingcat calls it a sales pitch.
In Newslit Daily.
newslit.substack.com/p/vercel-hac...
The Pentagon told agencies to drop Anthropic in February. The NSA is already running its newest model, Mythos Preview, for cybersecurity work. About forty organizations have access. Dario Amodei just met the White House chief of staff.
In today's Newslit Daily.
newslit.substack.com/p/vercel-hac...
The full rundown, plus what Vercel customers should rotate today, is in this morning's Newslit Daily.
newslit.substack.com/p/vercel-hac...
Vercel confirms attackers pivoted in from a compromised Google Workspace account at a third party vendor, then pulled source code, API tokens, and environment variables marked 'non sensitive.' ShinyHunters is selling 580 employee records.
The EU’s new age verification app was pitched as privacy-friendly and open source, but researchers quickly found serious flaws: biometric bypasses, exposed data on the device, and a design that can let others reuse an adult verification.
In today's Newslit Daily
newslit.substack.com/p/google-pai...
OpenAI just made Codex on Mac far more hands-on. It can see your screen, click through apps, type, run multiple agents, and now includes an in-app browser, image generation, memory across sessions, and dozens of plugins.
Full breakdown in today's Newslit Daily
newslit.substack.com/p/google-pai...
Full breakdown (plus today's other stories) in today's Newslit Daily.
newslit.substack.com/p/google-pai...
Google is changing how Chrome search feels on desktop.
Links from AI Mode now open beside the chat, so you can compare pages and keep asking questions without losing context.
Open tabs, notes, and images can also feed directly into search.
newslit.substack.com/p/google-pai...
Google rolled out a Gemini text-to-speech model, and the idea is control.
Developers can now steer tone, pacing, accent, and multi-speaker dialogue with simple prompts, which makes AI voice tools feel a lot more flexible.
Read more in today’s full newsletter:
newslit.substack.com/p/claude-id-...
A lot of people moved to Claude because it felt like the more privacy-minded AI option.
Now Anthropic is asking some users for a government ID and even a live selfie to unlock certain features or pass integrity checks.
Claude is reworking coding app around the idea: developers are working with more than one agent.
This setup is for juggling sessions, reviewing diffs, running tests, and editing files without bouncing between tools.
Check out this story in the full newsletter.
newslit.substack.com/p/openai-cyb...
Microsoft just rolled out a cheaper, faster image model, which sounds like a routine product update until you look at what it says about the bigger strategy.
Check out this story, and more, in the full newsletter.
newslit.substack.com/p/openai-cyb...
OpenAI is rolling out a new cyber model, but only to a small circle of trusted security users for now. On the surface, this is a product launch.
Underneath, it looks like the next phase of the AI race, where companies are building models for specific high-stakes jobs instead of just general chat.
Roblox is changing how younger users navigate the platform, with new account types for kids and teens in games and chats now.
It is a product update, but also a story about platforms trying to rebuild trust around child safety.
Read more in the full newsletter.
newslit.substack.com/p/amazon-app...
Google is cracking down on sites that hijack the back button and send users to pages they never meant to open.
It now classifies that tactic as spam, opening search penalties.
That policy shift could ripple across publishers ad tech and web browsing experience.
newslit.substack.com/p/amazon-app...
Amazon is making a much bigger play in satellite connectivity than this headline alone suggests. The company plans to buy Globalstar, build out direct-to-device service through Amazon Leo, and keep powering satellite features for iPhone and Apple Watch.
A key web preservation tool faces a problem: publishers are making it harder for the Wayback Machine to archive pages.
It sounds niche, but it matters for anyone trying to verify how the web changes over time.
See this story and more in the full newsletter.
newslit.substack.com/p/rockstar-h...
The suspect in the Molotov attack on Sam Altman’s home was reportedly tied to a public Discord server critical of AI development.
Police say the case also involved threats aimed at OpenAI’s HQ.
Check out this story and more in the full newsletter.
newslit.substack.com/p/rockstar-h...
Rockstar just got pulled into another messy security story, this time tied to a third-party breach and a ransom threat from ShinyHunters. The company says the impact is limited and doesn’t affect players, but the bigger question is what kind of internal data may now be floating around.
Google is pushing encrypted Gmail further onto mobile.
Enterprise users can now read and write end-to-end encrypted emails directly inside the Gmail app on Android and iPhone, without extra tools getting in the way.
More in the full newsletter:
newslit.substack.com/p/anthropic-...
A private warning from top US officials to big banks is not something you see every day.
This one was about advanced AI models and the possibility that the next wave of tools could create new kinds of cyber risk for the financial system.
A newly uncovered Adobe Reader flaw was exploited for months through booby-trapped PDFs, with researchers warning it could steal data and enable deeper attacks.
It starts with routine file open and gets more interesting fast
This story in the newsletter:
newslit.substack.com/p/metas-muse...
Tubi just made a smart move by putting its recommendation experience directly inside ChatGPT.
Instead of keeping discovery locked inside a streaming app, it is meeting people where they are already searching for what to watch.
More in the full newsletter.
newslit.substack.com/p/metas-muse...
Meta just introduced Muse Spark, the first public model from its new Superintelligence Labs, and it feels like more than just another model launch.
The bigger story is how Meta wants its AI to pull directly from the social universe it already owns, from Facebook to Instagram to Threads.
