Apple rents you your iPhone, OpenAI ships a keypad, and a prompt to end overthinking

Apple is reportedly launching 'Apple Upgrade', a lease-to-own program backed by Klarna, on July 28 in the US — letting you spread the cost of an iPhone, iPad, Mac or Watch across years and upgrade whenever you like. 📱 For heavy upgraders it's genuinely convenient, but 'lease-to-own' quietly normalises renting your hardware forever. Worth reading the fine print before the EMI dopamine kicks in.

The problem

Buying a new iPhone, Mac, or Watch outright means a large one-time cost, which pushes a lot of people toward carrier financing plans that lock them into a specific network for years.

What changed

Apple is reportedly launching 'Apple Upgrade,' a lease-to-own program backed by Klarna, on July 28 in the US — letting you spread the cost of an iPhone, iPad, Mac, or Watch across years and upgrade whenever you like, independent of any carrier.

The catch

'Lease-to-own' is a gentler way of saying you're renting your hardware indefinitely — it's genuinely convenient for people who upgrade often, but it normalizes never fully owning your device outright, and the fine print on fees and upgrade timing matters more than the pitch.

What it means for you

If you're a heavy upgrader, this could actually save you money over carrier financing — but read the terms before signing up, since the convenience is the sales pitch and the ownership trade-off is the part that's easy to skim past.

OpenAI has shipped its first-ever hardware product, and it isn't the long-rumoured Jony Ive AI device — it's Codex Micro, a $230 mechanical control deck built with Work Louder for driving its Codex coding agent. 🤖 It packs 13 switches, a touch sensor, an analog stick and a rotary dial to crank the AI's 'reasoning level' up or down. It's a limited-run tool for hardcore developers rather than a mass-market product, but it's a real hint at where AI-native gadgets are heading.

The problem

AI coding agents like Codex are entirely software, controlled through a keyboard and mouse — there's been no physical, purpose-built way to interact with them, even as they've become central to how developers work.

What changed

OpenAI shipped its first-ever hardware product: Codex Micro, a $230 mechanical control deck built with Work Louder specifically for driving its Codex coding agent, packing 13 switches, a touch sensor, an analog stick, and a rotary dial to adjust the AI's 'reasoning level' on the fly.

The catch

This isn't the long-rumored Jony Ive AI device, and it's not trying to be — it's a limited-run tool for hardcore developers, not a mass-market gadget, so most people will never touch one.

What it means for you

If you're a developer living inside Codex daily, this is worth watching for when it's available to buy; for everyone else, treat it as an early signal that AI-native hardware is starting to show up in niche, purpose-built forms before it goes mainstream.

Most people never find it: long-press anywhere on Google Maps to drop a pin, tap the pin's card at the bottom, and choose 'Measure distance'. 📏 Move the map and tap to add points, and it shows the straight-line distance between any spots — handy for planning road trips, checking how far a plot really is, or mapping a run. No extra apps, no signup.

The problem

Measuring the distance between two points on a map — for a road trip, a plot of land, or a running route — usually means reaching for a separate app or doing rough math in your head.

What changed

Google Maps has a built-in ruler: long-press anywhere to drop a pin, tap the pin's card at the bottom, and choose 'Measure distance' — then move the map and tap to add points, and it shows the straight-line distance between any spots.

The catch

It measures straight-line distance, not actual walking or driving distance, so it's useful for rough estimates rather than exact trip planning.

What it means for you

Next time you're eyeballing how far something is, use this instead of guessing — no extra app, no signup, and it's already sitting inside an app you open every day.

Next time you're stuck on a decision, paste this into ChatGPT, Gemini or Claude: 'Act as my skeptical advisor. I'm deciding [your decision]. Give me the 3 strongest reasons for, the 3 strongest reasons against, what I'm probably not considering, and one question I should answer before deciding. Be blunt.' 🧠 Most people use AI to agree with them — this forces it to challenge you, so you get a real second opinion instead of a cheerleader.

The problem

Most people use AI chatbots to confirm a decision they've basically already made, which means the AI ends up agreeing with them instead of actually stress-testing the choice.

What changed

A specific prompt flips that: 'Act as my skeptical advisor. I'm deciding [your decision]. Give me the 3 strongest reasons for, the 3 strongest reasons against, what I'm probably not considering, and one question I should answer before deciding. Be blunt.' — forcing the model to argue against you instead of validating you.

The catch

The output is still only as good as the context you give it — a vague or one-sided description of your decision will still produce a shallow 'against' case, since the model can only push back on what you tell it.

What it means for you

Save this prompt somewhere you'll actually use it — next time you're stuck on a real decision, paste it in with real detail about your situation, and treat the 'against' list as the part worth reading twice.