Pixel 11's Aug 12 date, India's recharge hike & open-source AI's biggest week
Today's tech: a reported Aug 12 Made by Google date for the Pixel 11, warnings of a 15-20% Jio/Airtel recharge hike, an Android trick to lend your phone safely, and the biggest week open-source AI has had yet. Plus, are free open models finally good enough to cancel your AI subscription?
- commentary
Made by Google looks locked for Aug 12
Made by Google looks set for Aug 12, and it's a big one. Per leaks, the lineup includes the Pixel 11 and 11 Pro, a Pixel 11 Pro Fold, and a Pixel Watch 5, all running a new 2nm Tensor G6 chip. The honest bit: Google is leaning hard on on-device AI, but price hikes look likely across the range. So the real question isn't whether it's good, it's whether it's worth the new price. Keep it labelled clearly, though: the date and specs are rumor until Google confirms. π±
- The problem
- Every major phone launch now leaks well in advance, but pricing and value questions usually can't be answered until the actual reveal β leaving buyers to guess whether to wait or buy now.
- What changed
- Made by Google looks set for Aug 12, with leaks pointing to the Pixel 11 and 11 Pro, a Pixel 11 Pro Fold, and a Pixel Watch 5, all running a new 2nm Tensor G6 chip, with Google leaning hard on on-device AI across the range.
- The catch
- Price hikes look likely across the range, and the date and specs are still rumor until Google confirms them β the on-device AI push doesn't guarantee the value equation improves for buyers.
- What it means for you
- The real question to hold onto until Aug 12 isn't whether the new Pixels are good β it's whether they're worth the higher price Google appears to be charging for them.
- news
India's mobile recharges may get pricier
Brace your wallet, India: reports say Jio and Airtel are prepping another tariff hike, roughly 15-20% on recharges this year. Nothing is official yet, so file it under report, not confirmed. But the logic checks out: the 5G buildout plus higher revenue targets means someone pays, and that someone is usually us. If a hike does land, locking in an annual plan before it kicks in can save you a decent chunk. πΈ
- The problem
- Mobile carriers periodically raise prices to fund network upgrades, and those costs almost always land on subscribers rather than the companies making the infrastructure investment.
- What changed
- Reports say Jio and Airtel are prepping another tariff hike, roughly 15-20% on recharges this year, driven by the ongoing 5G buildout and higher revenue targets.
- The catch
- Nothing is official yet β this is a report, not a confirmed price change β so exact timing and the final percentage could still shift before either carrier makes it official.
- What it means for you
- If a hike does land, locking in an annual recharge plan before it kicks in can save you a decent chunk β worth doing now rather than waiting for the official announcement.
- tip
The Android setting that lets you lend your phone safely
App Pinning (aka Screen Pinning) is one of Android's most underrated settings. When someone wants to 'quickly check something' on your phone, you can pin a single app and hand it over, and they can't swipe to your gallery, chats, or anything else until you unpin with your PIN. Turn it on under Settings, then Security (or Security & privacy), then App pinning, then open the app, tap Recents, and pin it. To exit, hold Back and Recents together. Zero downloads, and it's already on your phone. π
- The problem
- Handing your phone to someone who wants to 'quickly check something' means trusting them not to swipe into your gallery, chats, or anything else you'd rather keep private.
- What changed
- App Pinning (Screen Pinning) lets you lock your phone to a single app before handing it over β turn it on under Settings, Security, App pinning, then open the app, tap Recents, and pin it, and the other person can't leave that app until you unpin it with your PIN.
- The catch
- You have to remember to pin the app before handing your phone over each time β it's not a permanent setting, so it only protects you if you actually use it in the moment.
- What it means for you
- Turn on App Pinning now, before the next time someone asks to borrow your phone β it's a built-in, zero-download way to hand over exactly one app and nothing else.
- ai
Frontier AI is quietly going open-source
This might be the biggest week open-source AI has ever had. Three drops worth knowing: Kimi K3 from Moonshot brings 2.8T parameters, a 1M-token context window, native vision, and a spot around #4 globally, with open weights landing July 27. DeepSeek V4 pairs open weights with a genuinely cheap API. And Meta's Muse Spark 1.1 is an agentic model that can actually use your desktop and browser. The neutral, product-focused read: near-frontier capability is becoming something you no longer have to rent. For anyone building, or just paying for a chatbot, that's a big deal. π€
- The problem
- Near-frontier AI capability has historically meant paying a premium subscription to a closed lab β open alternatives existed, but usually trailed well behind the best paid models.
- What changed
- Three major open drops landed in one week: Kimi K3 from Moonshot brings 2.8T parameters, a 1M-token context window, native vision, and a spot around #4 globally, with open weights landing July 27; DeepSeek V4 pairs open weights with a genuinely cheap API; and Meta's Muse Spark 1.1 is an agentic model that can actually use your desktop and browser.
- The catch
- 'Near-frontier' still isn't 'frontier' β these models close the gap significantly but don't yet claim the top spot outright, and open weights landing July 27 for Kimi K3 means full independent verification is still pending.
- What it means for you
- If you're paying for a premium AI subscription purely for general capability, this week is a good prompt to test Kimi K3 or DeepSeek V4 against what you're used to β near-frontier capability is becoming something you no longer have to rent.
- community
Still paying for ChatGPT/Gemini, or switched to open models?
Open models like Kimi K3 and DeepSeek V4 are now near-frontier and basically free to use, yet a lot of us are still paying premium subscription prices every month for ChatGPT Plus or Gemini. So, honestly: are you still paying for a premium AI subscription, or have you gone full free and open-source? And if you're paying, what's the one feature keeping you locked in?