The AI race was never just about models. This week proved it's about who controls the infrastructure of power. In 24 hours, OpenAI turned Anthropic's ethical stand into a business opportunity. The question isn't who has better AI.

It's who has the better deal.

WHAT HAPPENDED:

The Trump administration ordered all federal agencies to immediately stop using Anthropic's technology. Defense Secretary Pete Hegseth designated the company a "supply chain risk to national security" - a label previously reserved for foreign adversaries.

Hours later, OpenAI announced a deal to deploy its models inside the Pentagon's classified networks, while raising $110 billion at a $730 billion valuation, backed by SoftBank ($30B), Nvidia ($30B), and Amazon ($50B).

WHY IT MATTERS:

Anthropic drew two hard lines: no autonomous weapons, no mass surveillance of Americans. The Pentagon wanted "all lawful purposes" with no explicit carve-outs. Talks collapsed. What followed was unprecedented: the first time in US history a domestic company was designated a national security supply chain risk - not for ties to foreign adversaries, but for refusing to sign a contract on the government's terms.

OpenAI moved fast. Altman claimed the same red lines as Anthropic, yet still landed the deal. Legal experts are skeptical. Altman himself admitted it was "definitely rushed" and "the optics don't look good." Critics point to Executive Order 12333, which some analysts argue still permits bulk data collection under the agreement's language.

Anthropic chose principles. OpenAI chose access. Both decisions carry consequences, but only one comes with a Pentagon contract and a $730B valuation. The deeper question isn't who's right on ethics. It's whether any AI company can hold meaningful safety commitments once government and capital are on the other side of the table. As Fortune noted: every general counsel at every Fortune 500 with Pentagon exposure is now asking whether using Claude is worth the risk. That reputational damage will take years to reverse, regardless of how the courts rule.

If you’re a decision maker at your company, you need to be on the bleeding edge of, well, everything. But before you go signing up for seminars, conferences, lunch ‘n learns, and all that jazz, just know there’s a far better (and simpler) way: Subscribing to The Deep View.

This daily newsletter condenses everything you need to know about the latest and greatest AI developments into a 5-minute read. Squeeze it into your morning coffee break and before you know it, you’ll be an expert too.

Subscribe right here. It’s totally free, wildly informative, and trusted by 600,000+ readers at Google, Meta, Microsoft, and beyond.

Boris Cherny, creator of Claude Code at Anthropic, told Lenny's Podcast that the job title "software engineer" will be replaced by "builder" by end of year and it will be painful. At OpenAI, 95% of engineers already use Codex. Those who embrace it open 70% more pull requests than peers who don't.

→ When the person building AI coding tools says this, it's not a prediction. It's a warning from the inside.

Researchers from MIT, WashU, and UCLA published a paper on the economics of AGI. Core argument: when AI can execute most tasks, the scarce human resource becomes verification checking AI output, auditing intent, owning accountability.

They warn of a "Hollow Economy" where AI satisfies every measurable metric while accumulating hidden debt between output and actual human intent.

→ The most valuable people won't be those who produce the most. They'll be the ones who catch mistakes fastest.

A piece circulating widely in product and design circles: as AI agents become the primary users of digital platforms, the entire UX playbook needs rewriting. Agents don't need beautiful onboarding flows. They don't read tooltips. They need clean APIs, consistent schemas, and machine-readable error messages.

→ Does your product roadmap account for non-human users? This is a quiet shift with fundamental consequences.

You want to build an app? prototype? or agent? But you don't have data? And you don't want to pay to start?

Ok. I have something for you:

The solution Someone published a list of over 320,000 free public APIs on GitHub, community-curated and verified as working. Covers weather, finance, news, sports, crypto, AI and ML APIs, government open data, maps, geolocation, and entertainment. Fully categorized, searchable, and open source.

To use it: Go to github.com/public-apis/public-apis, search by category or keyword, copy the endpoint, and test immediately with curl — or paste the API docs directly into Claude or ChatGPT to build fast. No setup required. No credit card.

Who it's for: Developers, vibe coders, product builders, data analysts,.. I mean anyone at the stage of "I need data to test this idea."

The problem You spend two days on a polished Figma mockup. Then user feedback from a real prototype tells you something completely different.

The solution A field study published on UX Design found that working prototypes consistently outperform mockups for gathering actionable feedback. The reason is simple: users can't give accurate feedback on something they haven't actually used. Polished mockups create politeness bias - users say nice things instead of surfacing real problems.

How to use it Use AI tools like Lovable, Bolt, v0, or Claude Code to build a working prototype in 1 to 2 hours instead of spending two days on mockups. Put it in front of 3 to 5 real users immediately. The feedback will be harder to hear and that's exactly why you need it.

Anthropic held its principles. OpenAI took the contract. Block laid off 4,000 people. The creator of Claude Code said the job title "software engineer" is going away. All of this happened in the same week.

AI is no longer a conversation about the future.

👉 Forward it to someone trying to keep up with AI.

Read online: techzip.beehiiv.com

Tech.zip Newsletter
| Zipping what truly matters in the AI era.