Published on LinkedIn, Substack and AmitabhApte.com on August 24, 2025
In spotlight this week: AI assistants go mainstream – Apple eyes Gemini, Google embeds it.
What if your phone didn’t just respond to you, but anticipated your needs before you spoke? This week, AI assistants took a giant leap from cloud-based helpers to embedded, proactive companions. Apple is reportedly in talks with Google to integrate Gemini into Siri, potentially transforming its underwhelming assistant into a multimodal powerhouse. While no deal is confirmed, the move signals Apple’s openness to external AI partnerships, including ongoing discussions with OpenAI and Anthropic. The goal: to bring richer, more conversational intelligence to iPhones and across Apple’s ecosystem.
Meanwhile, Google has taken a decisive leap forward with the launch of its Pixel 10 smartphone lineup, embedding Gemini AI directly into the device via its new Tensor G5 chip. The Pixel 10 series introduces features like Magic Cue, which proactively surfaces relevant info across apps, and Gemini Live, which offers real-time visual assistance based on what the phone sees. Other AI-powered upgrades include Voice Translate for multilingual calls, NotebookLM integration for smarter notetaking, and Pixel Journal for wellbeing tracking. The Pixel 10 Pro models even come bundled with a year of Google AI Pro subscription, unlocking creative tools like Imagen 4 and Veo 3.
My key takeaway: The battleground is no longer just software. It’s the device, the chip, and the ecosystem. Apple is pivoting strategically, Google is executing decisively. Both point to the same future: assistants that are native, multimodal, and deeply personal.
Noteworthy this week: what caught my eye in the AI and tech world
Meta + Midjourney – Meta has struck a deal to license Midjourney’s image-generation tech for future products. It’s a boost in visual creativity and a possible hedge against the lukewarm response to its own Llama 4. Partnerships like this are signals. Meta knows it needs an edge beyond its own labs.
Intel agrees 10% U.S. stake –Intel is selling a 10% stake to the U.S. government, one of the largest federal equity moves since the 2008 auto bailout. It strengthens chip sovereignty but also raises a hard question: what happens when governments become shareholders in the engines of tomorrow?
New turn in Nvidia’s chip for China – Nvidia has stopped producing its H20 chip for China after Beijing told local firms not to buy it, despite U.S. approval. It’s the latest flashpoint in the U.S.–China tech standoff. Critical AI infrastructure is seen as differentiator in increasing tense geo-political scenarios.
Coinbase firing engineers for not onboarding AI – Engineers who failed to adopt tools like GitHub Copilot were let go. Coinbase calls it “AI fluency or out.” Yes, fluency in AI is now non-negotiable. But enforced adoption without empathy risks losing talent and trust.
TCS opens AI-led operations centre in LATAM – A new AI-led operations centre marks its eighth in the region. Jobs, skills, and digital transformation are the pitch. Indian IT giants are exporting AI at scale, and LATAM is the next growth frontier.
TikTok to replace UK staff with AI –Over 85% of moderation is now automated, with thousands of roles at risk. Efficiency is up but user trust may not be. Platforms can’t trade human oversight for pure automation without ethical safeguards.
In summary: my key takeaway this weekend
This week marks a turning point. AI assistants are no longer cloud novelties, they’re becoming embedded essentials. Apple is courting Gemini. Google is hard-wiring it into Pixel 10. Meta, Intel, Nvidia, Coinbase, TikTok, each move adds to the same message: AI isn’t just a feature. It’s the new operating system of everything. The question now is not if you’ll use an assistant, but whose ecosystem you’ll live in.
Amitabh Apte 