Apple’s friendly new hardware arrives as Washington picks its favorite AI lab
Apple wants your upgrade, Washington wants your Artificial Intelligence lab.
Apple’s big news this week was new hardware. The government’s big news was picking which AI lab it wants in the room — and which one to ice out.
Neo and Air: the “student Mac” grows up
The MacBook Neo is Apple’s new entry‑level laptop, and it’s the first “student Mac” in a while that doesn’t feel like an apology. It’s colorful, thin and unmistakably a Mac the second you open it: aluminum body, side‑firing speakers and a deck that looks more playful than anything else Apple sells right now.
The keyboard does a lot of that work. The keys are color‑matched to the shell, the wells are a little more sunken, and from a distance it looks suspiciously like the butterfly era sneaking back in through a side door. Up close, it’s still a scissor design, but the visual throwback is hard to miss. It is absolutely the kind of laptop you’ll spot across a lecture hall, which is probably the point.
Under the hood, the Neo runs on Apple’s A18 Pro — an iPhone‑class chip that breezes through web browsing, note‑taking and basic spreadsheets, even if it’s not meant to chew through 4K video edits all night.
One step up the line, the new MacBook Air quietly becomes the default laptop for anyone who mostly lives in a browser and a notes app. Apple bumped the chip, base storage and memory, and leaned into the “this will survive your 47‑tab meltdown the night before a midterm” pitch. If the Neo is for getting into the ecosystem, the Air is for living there.
The base model will cost $599, or $499 using the student discount.
iPhone 17e: modern brain, 2022 face
Apple’s first phone of 2026, the iPhone 17e, is a weird mix of progress and pettiness. On the inside, it’s great: current‑gen chip, more storage than older base models, and support for the new on‑device AI tricks. On the outside, it looks like someone raided the iPhone 14 parts closet and told industrial design to make it “vaguely 2026.”
The most obvious sore spot is the display. Midrange Android phones have treated high‑refresh screens as table stakes for years, and Apple is still happy to ship a brand‑new iPhone with a 60‑hertz panel. Most people who buy the 17e won’t know or care what refresh rate their screen is running at. That doesn’t make it less absurd from a company that loves to lecture everyone about ProMotion higher up the line.
Then there’s the front of the phone: big notch, no Dynamic Island. It’s obvious what’s happening here — this is a parts‑bin iPhone built on leftover 14‑era shells, dressed up just enough to call it new. That’s fine if the goal is to squeeze more life out of old tooling. It’s less fine when the marketing story is about how modern and “intelligent” the lineup is, and the phone you’re actually going to see in carrier promos all year looks like it time‑traveled from 2022.
The 17e gets the invisible stuff right: speed, battery, cameras and AI. But the notch and the 60‑hertz screen send a familiar message: at a certain price, Apple is comfortable telling you that you don’t get the full experience.
Shiny consumer AI vs. classified AI
While Apple was polishing hardware we actually touch, a different kind of upgrade was happening in D.C. OpenAI has agreed to put its models to work inside the national security apparatus, after Anthropic walked away from a similar request over issues like fully autonomous weapons and mass domestic surveillance.
Anthropic wanted hard bans baked into the deal. The government wouldn’t go that far, so Anthropic said no. OpenAI, by contrast, agreed to a looser framework built around “red lines” and review processes — limits that sound firm but can be reinterpreted over time.
Civil liberties groups and AI policy researchers see a clear trade‑off: OpenAI stays in the room on a huge national‑security contract by accepting softer rules, while Anthropic tried to keep its existing guardrails and got frozen out. After Anthropic refused to loosen its policies, the administration moved to label the company a potential risk to the national‑security supply chain and quietly told agencies and contractors to start backing away. On paper, it’s framed as a procurement and reliability issue. In practice, it looks a lot like punishing the lab that wouldn’t give the government as much freedom, while deepening ties with the one that did.
Even that isn’t enough for some defense officials, who are already grumbling that OpenAI’s “red lines” might interfere with missions if they’re interpreted too strictly. So you end up with a strange loop: one lab is punished for being too cautious, another is nudged not to be cautious enough, and the rest of us are told to trust guardrails we’re not allowed to read.
The gap we’re living in
Put the two stories side by side and you get a sharp split‑screen. On one half, Apple is shipping color‑matched keyboards, friendly‑looking laptops and a “budget” iPhone that can run the same on‑device AI as the flagship, just with a notch and a 60‑hertz ceiling. On the other, OpenAI is wiring its models into national‑security work, Anthropic is getting slapped for saying no, and the rules for how far this tech can go in war and surveillance are being written in rooms none of us can walk into.
On campus, that gap isn’t abstract. When you buy a MacBook Neo or an iPhone 17e, you’re opting into an ecosystem where AI mostly shows up as nicer autocorrect, smarter photos and essay‑polishing suggestions. When you fire up a chatbot to brainstorm a paper or summarize a reading, you’re also plugging into the same companies deciding how comfortable they are selling those capabilities to the government.
You don’t have to boycott new phones or swear off AI tools to see the tension. Hardware is getting cheaper, prettier and friendlier at the exact moment the power behind it is being concentrated in a handful of firms cutting high‑stakes deals over targeting, surveillance and “acceptable risk.” The real question isn’t whether the MacBook Neo looks good (it does) or whether a 60‑hertz iPhone in 2026 is silly (it is). It’s how much say any of us actually have in the future we’re buying into when we tap “Agree” on the software terms and “Pay Monthly” on the upgrade page.
