Unpacked Countdown, a Gemini Rebrand, and Apple vs OpenAI

Samsung's Galaxy Unpacked is officially set for July 22 in London, and leaks point to five devices — refreshed Fold and Flip foldables plus a rumored mystery fifth device. It streams live on Samsung.com and YouTube at 9AM ET / 6:30PM IST. The real question isn't specs, it's whether this is the year foldables finally stop being a flex and start being the default. Leak details are still rumor until Samsung says otherwise; the date and event are confirmed.

The problem

With Unpacked three days away, the gap between confirmed facts (the date) and rumored specifics (the actual device lineup) is at its widest — easy to mistake leaks for announcements.

What changed

Galaxy Unpacked is officially set for July 22 in London, streaming live on Samsung.com and YouTube at 9AM ET / 6:30PM IST, with leaks pointing to five devices — refreshed Fold and Flip foldables plus a rumored mystery fifth device.

The catch

The date and event are confirmed; the device lineup, including that mystery fifth device, is still leak-based until Samsung actually says otherwise on stage.

What it means for you

Set a reminder for the livestream rather than trusting the leak roundups — the real question worth watching for isn't the spec sheet, it's whether this is the year foldables stop being a flex and become the default.

Google has renamed NotebookLM to Gemini Notebook, folding one of its most genuinely useful research tools fully under the Gemini brand. Same idea — drop your sources, get grounded answers and audio overviews — now with clearer Gemini integration. The tool itself is still one of the best ways to actually reason over your own documents instead of trusting a model's memory.

The problem

Google has been quietly consolidating its AI tools under a single brand, and NotebookLM — one of its most genuinely useful research tools — remained a separate, less-integrated product.

What changed

Google renamed NotebookLM to Gemini Notebook, folding it fully under the Gemini brand with clearer integration, while keeping the core idea intact: drop your sources in and get grounded answers and audio overviews back.

The catch

A rebrand doesn't change the underlying tool's capabilities — anyone expecting new features from the name change alone will be looking at the same product with a new label.

What it means for you

If you haven't used NotebookLM before, look for it under its new name, Gemini Notebook — it's still one of the best ways to make an AI reason over your own documents instead of trusting its memory.

Hold Option and click the Sound icon in your Mac's menu bar — you get an instant switcher for every audio output and input device, no diving into System Settings. Jumping between AirPods, speakers, and a headset becomes a single click. Small habit, saves you the settings detour every single day.

The problem

Switching audio output on a Mac — between AirPods, speakers, and a headset — usually means digging into System Settings every single time.

What changed

Holding Option and clicking the Sound icon in the Mac menu bar brings up an instant switcher for every audio output and input device, letting you jump between devices in a single click.

The catch

It's an Option-click shortcut that's easy to forget exists if you've never stumbled onto it — there's no visual hint in the menu bar that holding Option changes anything.

What it means for you

Try it right now: hold Option and click your Mac's Sound icon — if you switch audio devices even occasionally, this replaces a settings detour with one click, every day.

There's no single winner anymore, so pick per job. ChatGPT (GPT-5.6) is the best all-rounder and still owns voice. Gemini 2.5 Deep Think is the research beast with a 2M-token context window for long documents. Claude Fable 5 is the coding king at ~80% on SWE-Bench. Everyday use and voice lean ChatGPT, deep research leans Gemini, and coding or agents lean Claude. Match the tool to the task instead of chasing one app to rule them all.

The problem

Picking a single 'best' AI assistant means implicitly accepting worse performance on whatever task that model isn't built for — and the three leading models now specialize in genuinely different things.

What changed

The mid-2026 verdict: ChatGPT (GPT-5.6) is the best all-rounder and still owns voice, Gemini 2.5 Deep Think is the research beast with a 2M-token context window for long documents, and Claude Fable 5 is the coding king at roughly 80% on SWE-Bench.

The catch

These specialties shift as each lab ships updates — this is a snapshot of mid-2026, not a permanent hierarchy, and Gemini 3.5 Pro's eventual release could reshuffle the research and context-window lead specifically.

What it means for you

Match the tool to the job instead of picking one to be loyal to: ChatGPT for everyday use and voice, Gemini for deep research and long documents, Claude for coding or agents.