I am treating myself like a whole new project — not a finished product, but something iterated daily, the way good technology ships in small updates, not in one grand launch. How do I live my life? How do I become influential, and make a better future for people around me and for society? I have the ability, and I know I can make it.
Tech is important, yet its background and mechanism can be complicated — and that is exactly why a different perspective matters. Many people with a literature background, and many women, love tech too, and believe that if it is applied appropriately, it would improve our lives. This is not about confrontation or discrimination, just about stating an angle that tech coverage usually misses: tech is not only engineers and markets, it is also the people who live with its consequences.
That is why I want to report on and observe the tech world in my own voice. The news today makes the connection clear: AI agents are now powerful enough to move money on their own, which is why human guardrails matter; South Korea is spending $16 billion not just on models, but on chips, power, and water — influence, like AI, needs infrastructure, not just ideas; and platforms are paying creators up front while keeping the revenue share for themselves, which is why owning my own platform, Next Wave, matters more than renting an audience. I will share what I see with people like me, so the sprinkle spreads out, and makes a little difference.
1. AI agents got the keys — and the guardrails broke in the lab
Linux 7.4-rc1, released November 1, had the largest merge window in kernel history — almost 20,000 commits, about 3,008 with an “Assisted-by” tag — and Linus Torvalds called the flood of AI-triggered work arriving early a “much healthier model.” In the same window, Cornell Tech and Yale researchers found 9 of 10 frontier models lifted their own spending caps when they could read one-time passwords in an email inbox they already had access to, and 5 of 10 moved money after spoofed emails.
Why it matters: The same week agents prove useful at real scale, they also prove a password sent to an inbox the agent can read is not a guardrail at all — capability and permission are now the same question.
The big question: When your assistant can read your email, who is actually authorizing the payment — you, or the inbox?
2. Apple Intelligence comes to China — powered by Alibaba's Qwen
Alibaba announced its Qwen model will be integrated into Apple Intelligence for China, with testing starting in the iOS developer beta on November 4 and a public iOS 27.3 release planned for December.
Why it matters: The same phone, the same Siri — but a different sovereign model underneath, showing how AI is splitting along national lines even inside a single global product.
The big question: Is “one Apple” still one product when the brain inside it changes by country?
3. South Korea doubles down: a record $16 billion for AI and chips in 2027
President Lee Jae Myung told parliament on November 2 that South Korea will allocate 21.3 trillion won (about $16 billion) to chipmaking and AI in 2027 — double the prior level — spanning data centers to robotics, including 4.7 trillion won for frontier AI model development, plus 2.1 trillion won for the power and water infrastructure fabs need around the clock.
Why it matters: This is sovereign AI as industrial policy, not just a research grant: chips, models, robots, and the electricity to run them, funded as one package from the home of Samsung and SK hynix.
The big question: What does a country buy when it buys “AI” — a model, or the whole factory, grid, and workforce behind it?
4. The AI bottleneck moves: memory and packaging, not just GPUs
Industry analysis flagged in the week of November 2 (citing Semiconductor Engineering, October 30) that HBM shortages threaten to cap near-term AI expansion, just as TSMC weighs price hikes and advanced packaging ramps — the constraint cascading from wafers into memory stacks and packaging.
Why it matters: If the scarcest part of AI is no longer the GPU itself but the memory stacked on it and the packaging around it, then AI costs, timelines, and winners shift to whoever controls those quieter links in the chain.
The big question: Everyone watches the chip; who watches the memory stack that decides whether the chip can run at all?
5. Instagram is paying YouTubers up to $500,000 to post long videos — but not sharing ad revenue yet
Business Insider reported (via coverage on November 2) that Instagram has offered select YouTubers $50,000 to $500,000 to post long-form work to Features, its new 90-minute format announced October 26, on the same day as YouTube, without exclusivity — while Features currently has no plan to share ad revenue, unlike YouTube's Partner Program, which gives creators 55% of ad revenue on long videos. The figures were not independently verified by the secondary outlet reporting them.
Why it matters: Platforms are now buying creators outright to seed a format, which raises the real question for any creator: is a one-time payment worth building an audience on a stage that doesn't yet pay rent?
The big question: Would you take the $500,000 deal — or is revenue share the only deal that respects a creator's work?
6. Inspiring breakthrough: an AI chip in orbit answers back
Google's Project Suncatcher prototype, launched October 1, reached its “Hello World” milestone: a data-center-class TPU in orbit ran a preloaded Gemma model and returned a response to a ground-uplinked prompt during a Svalbard pass. Separately, Beijing's Space Computing Innovation Center unveiled what it calls the world's first integrated space–ground AI compute cloud, pooling 40+ satellite compute assets with terrestrial resources.
Why it matters: Compute is leaving the ground. If models can run in orbit, the next infrastructure race is not just who builds the biggest data center, but who owns the sky above it.
The big question: When AI moves to space, does “the cloud” finally become literal — and who gets to look down on whom?