First Pour.
In a world shaped by AI, is college…cooked?
I get this question constantly in my DMs, especially from tech folks who find themselves overeducated and underemployed. And it’s not like they did something wrong to get there. We’ve been told our entire lives that a 4-year degree was the golden ticket to long-term economic security.
Now look where we are. Tech workers are getting laid off weekly. CS grads are working at Starbucks. And AI upskilling is expected to be an extracurricular activity.
Serious question: if AI can write your paper, summarize your reading, run the analysis, and even draft your job application, what exactly are we spending six figures of tuition on?
Here’s my slightly bitter Tea: higher education, as currently structured, is not built for an AI world. If universities want to stay relevant, they need to take a hard look in the mirror (and the market) and accept that knowledge has become commoditized. So they can’t just teach it anymore.
They’ll have to shift their focus. To areas where they can actually add value. And IMO, it’ll be about teaching judgement:
The ability to ask better questions than the machine.
The creativity to deploy AI productively.
The resourcefulness to lead through ambiguity…when the prompt has no template.
In an AI economy, memorization loses value, and interpretation, synthesis, taste, and nuance become our value-add. Experiential learning must become the new default in higher ed.
Something to sip on ☕️
Today’s Tech Menu:
OpenAI’s device leak signals a hardware land grab
Reports point to a family of AI-native devices, including glasses, a speaker, and a lamp. This move is bigger than a new gadget launch — it’s about owning the AI interface layer and closing the loop on the ecosystem. My take linked.Last week’s India AI Impact Summit signals a multipolar AI race
Global CEOs, capital commitments, and clear sovereign ambition. India is pushing for a seat at the AI table. See “The Steep” below for my take.Frontier LLMs enter national security workflows
As OpenAI, xAI, and Anthropic deepen U.S. government contracts, AI models are moving into defense and intelligence environments — raising new questions about oversight and use. Tensions emerge as Anthropic resists broad Pentagon usage terms tied to surveillance.Big Tech commits ~$650B to AI infrastructure in 2026.
Microsoft, Google, Amazon, and Meta are still funneling historic capital into data centers and AI buildouts as they gear up for the next phase of AI growth.Can AI move from cost center to profit source?
Microsoft is reporting roughly 15M paid Copilot seats and explicitly citing AI as a growth driver in earnings. After two years of heavy infrastructure spending, it’s clear what the markets are demanding: AI must begin to drive revenue, not just narratives around innovation.
The Steep.
Is India a Sleeper AI Superpower?
In case you were living under an AI-generated rock, the India AI Impact Summit was last week. Everyone from Sam Altman to Sundar Pichai showed up.
And I’ll admit it. At first, it felt random to me. India? Now? Why?
Indian Prime Minister Narendra Modi hand-in-hand with the top Global AI CEOs at India’s AI Impact Summit
But the more I followed the summit, the clearer it became: India is staking a serious claim in the AI race. It’s no longer content serving as the back office to global tech like it has for so many years. This time, it wants a real seat at the table. And I think India might just pull it off.
Let’s look at India’s structural advantages:
• A massive developer base
• A billion-plus consumer market
• English fluency
• Mobile-first adoption at scale
Geopolitics matter a lot here too.
As U.S. immigration tightens and more engineering talent opts to stay local, domestic ecosystems will likely compound. Talent begets capital. Capital accelerates innovation. And eventually, market leadership emerges.
A few months ago, I met with Ronnie Chatterji, OpenAI’s Chief Economist. His take might be a hot one:
At the infrastructure level, India may be late. The frontier model race is dominated by the U.S. and China; compute and chips are expensive and supply-constrained.
The real opportunity for India may not be in training the biggest models. It may instead be in applying them.
In a billion-person market, a beta test is larger than most startups’ entire TAM. (no wonder Sam showed up).
The brew I’m sipping on this week? Definitely a Chai. ☕️
Final Sip.
Well guys, you might just be looking at a newly-minted internet celebrity 😂. The Waymo video (below) went mega viral with 17M views on Instagram (and counting). It’s been a wild week overall: lots of inbound, lots of internet trolling, and lots of ideas too.
It’s like…ofc. The girl who’s been laser-focused on producing serious, thoughtful AI content goes viral for a chaotic robotaxi moment.
The video still cracks me up because it’s truly such an accurate snapshot of my life: always messing with some new tech thing, wearing the outfit, and honestly, not taking myself too seriously along the way.
So I’m going to lean in. I’m going to start doing more lifestyle content on socials and bring you along for the ride in my fast-paced, tech-forward life. Follow along on Instagram if you don’t already. And don’t worry, the Tea in Tech episodes aren’t going anywhere.
Until the next brew,
Meghana