YouTube kills the Shorts dislike button, Apple lets you rent your iPhone
Today's tech: YouTube's big Shorts overhaul reshapes how creators get reach, Apple's lease-to-own program goes live, a hidden macOS batch-rename trick, an AI 'noise-filter' prompt, and a poll on humanoid robots.
- commentary
YouTube just killed the Shorts dislike button
YouTube has given Shorts a TikTok-style makeover, and creators should pay attention. The dislike button is gone, replaced by a heart. You can now tap and hold to play at 2x speed, and a new Clear Screen mode hides every icon with one tap. The bigger shift is under the hood: with dislikes frozen, the algorithm now leans harder on watch time, rewatches, comments and shares, so your hook and retention matter more than ever.
- The problem
- For years, Shorts creators had noisy, low-signal feedback: a dislike count that could be brigaded and a like button that said little about whether a video actually held attention. Reach felt unpredictable and hard to optimize for.
- What changed
- In its June-July 2026 rollout, YouTube removed the Shorts dislike button entirely and replaced the thumbs-up with a TikTok-style heart. It added 2x-speed playback (tap and hold), the 'most requested' Shorts feature, plus a Clear Screen mode that hides all on-player UI. Dislike counts stopped updating by the end of June; creators can still see historical dislikes in YouTube Studio.
- The catch
- Losing the public dislike signal cuts both ways. Low-quality or misleading Shorts lose an easy warning flag, and creators lose one honest data point about what is landing. The algorithm's heavier reliance on watch time can also reward bait that holds attention over content that genuinely helps.
- What it means for you
- Stop optimizing for likes and start optimizing for the first two seconds. Track watch time, rewatches and shares in Studio, and treat a strong hook plus tight retention as your real growth levers going forward.
- news
Apple will now let you rent your iPhone
A big shift in how Apple sells hardware goes live today (July 28, US). 'Apple Upgrade,' backed by Klarna, lets you lease an iPhone or Apple Watch for up to 24 months and an iPad or Mac for up to 36. You can pay it off early, upgrade mid-way, or hand the device back at the end. It makes expensive devices feel affordable, but lease-to-own means you may keep renting instead of owning, so it rewards serial upgraders and can trap everyone else. Read the fees before you sign.
- The problem
- Apple hardware keeps getting pricier, worsened by a memory-chip shortage, and big upfront costs push buyers toward older models or away from Apple entirely. Existing carrier and card financing is fragmented and often unclear.
- What changed
- Apple and Klarna launch 'Apple Upgrade' in the US on July 28, 2026, one of the largest-ever changes to how Apple sells products. It offers leases up to 24 months on iPhone and Apple Watch and up to 36 months on iPad and Mac, with options to pay off early, upgrade to a newer model mid-term, or return the device at the end.
- The catch
- 'Lease-to-own' can quietly become perpetual renting: upgrade every cycle and you never actually own anything, and some upgrade transactions carry extra fees. It is US-only at launch, and total cost over a full term can exceed buying outright.
- What it means for you
- If you upgrade every year, this can be convenient and cash-flow friendly. If you tend to keep a phone for 3-4 years, buying outright is almost certainly cheaper. Either way, read the fee schedule and end-of-term terms first.
- tip
Rename 50 files at once on your Mac
A macOS hidden gem for anyone with a messy Downloads folder: you don't need an app to batch-rename files. Select all the files in Finder, right-click and choose 'Rename [N] Items,' then pick Replace Text, Add a prefix or suffix, or Format to auto-number them (Photo-1, Photo-2 and so on). It takes about 30 seconds and needs zero downloads, and it's perfect for exported screenshots or invoices.
- The problem
- Cleaning up dozens of messily named files (screenshots, invoices, exports) one by one is tedious, and most people install a third-party app for something macOS already does natively.
- What changed
- Finder has a built-in multi-file rename: select several files, right-click and choose 'Rename [N] Items,' then use Replace Text, Add prefix/suffix, or Format for sequential numbering. No app, no Terminal.
- The catch
- It is easy to miss because it only appears when multiple files are selected, and there is no one-click undo for the batch beyond standard Cmd-Z right after, so double-check your pattern before applying it to a big folder.
- What it means for you
- Next time you export a pile of screenshots or receipts, select them all, right-click, Rename, and use Format to number them cleanly in one shot.
- ai
The AI prompt that filters hype from any article
Save this prompt: it turns any long article, PDF or report into a 20-second read. Paste your text, then add: 'Give me a 1-line TL;DR, 3 key facts, 1 claim that's overhyped, and what I should actually do next. Be blunt, no filler.' It works because most AI summaries just shorten; this one forces the model to separate signal from hype and hand you an action. Works in ChatGPT, Gemini or Claude.
- The problem
- Default AI summaries just compress text. They don't tell you which claims are marketing fluff or what you should actually do with the information, so you still have to re-read to make a decision.
- What changed
- A simple prompt structure fixes that: ask for a one-line TL;DR, three key facts, one overhyped claim, and a concrete next action, and tell the model to be blunt with no filler. The 'flag the hype' and 'give me an action' steps are what make it useful.
- The catch
- The model can still be wrong or over-confident about what's 'hype,' so treat its verdict as a fast first pass, not gospel, and sanity-check anything high-stakes against the source.
- What it means for you
- Copy the prompt, paste any long read after it, and get signal plus a next step in seconds. It works across ChatGPT, Gemini and Claude.
- community
Humanoid robots doing your chores by 2030 - inevitable or overhyped?
Humanoid robots are suddenly everywhere. Unitree just launched commercially in Europe, and BYD and XPeng are building full-stack humanoids in-house, with demos that look genuinely wild. Genuine question for you: will a humanoid robot be doing your household chores by 2030, or is this overhyped for now? Where do you land, and why?