NEWS YOU CAN’T MISS

$700 Billion on AI. That’s Not Normal.

Alphabet, Microsoft, Meta, and Amazon are expected to spend nearly $700 billion combined this year on AI, a 60% jump from 2025. That's not a typo.

To put it in perspective, $700 billion is more than the entire GDP of Switzerland. They're loading up on high-priced chips, building mammoth data centers, and buying the networking tech to connect it all.

But here's the part that affects you: this spending spree is causing real shortages across the economy according to The Washington Post.

• Electricians are getting harder to find because data center construction is snapping them all up.

• Smartphones are expected to get pricier, potentially for years, as chip demand outstrips supply.

• And promising innovations outside of AI are being starved of funding, because all the capital is chasing the same opportunity.

Ayesha's Take: We've seen tech gold rushes before, but nothing at this scale. The AI boom is real, but so is the risk of a monoculture economy where one technology sucks the oxygen out of everything else. The companies spending these billions are betting they'll earn it back many times over. History suggests some will, and some very much won't.

Sources: CNBC · Washington Post

TikTok’s Design Is Under Fire

The European Commission dropped a bombshell on February 6: TikTok's app features (think infinite scroll, auto play videos, and constant push notifications) constitute "addictive design" that could harm children and vulnerable adults.

Translation: the things that make TikTok so hard to put down may actually violate EU law.

The law in question is the Digital Services Act (DSA), Europe's landmark regulation that requires big platforms to assess and reduce the harm their products cause. 

If TikTok is found in breach, it faces fines of up to 6% of its global annual revenue, we're talking billions of dollars. 

And this isn't just about TikTok. Every social media company is watching, because the same logic applies to Instagram Reels, YouTube Shorts, and any app designed to keep you scrolling.

Ayesha's Take: Europe is doing something no other region has had the nerve to do: treating addictive design the way we treat addictive substances, as something that needs regulation, not just a personal responsibility problem. Whether you think this is overreach or overdue, it sets a precedent. If "addictive by design" becomes a legal liability, the entire attention economy has to rethink its business model. 

Sources: European Commission · Euronews

AI Can Talk You Out of Being Right

Here's an unsettling finding from Harvard and MIT: when you push back on ChatGPT's answer, pointing out a flaw, questioning its logic, it doesn't just correct itself. It often fights back. With more confidence, more detail, and more persuasive language than before.

Researchers call this "persuasion bombing." In a field study with over 70 Boston Consulting Group consultants, professionals used GPT-4 for a challenging financial analysis task and then tried to validate its output.

When they fact-checked, exposed inconsistencies, or flat-out disagreed, the AI ramped up its persuasive tactics, burying the human's concern under an avalanche of confident-sounding text.

The result? Many ended up deferring to the AI, even when their initial instinct was right.

Ayesha's Take: This is the most important AI paper I've read this month. We've been told the solution to AI error is "keep a human in the loop." This study shows the loop itself can be compromised. The AI doesn't need to be right, it just needs to sound more right than you feel.

Source: MIT Sloan Management Review · Harvard D^3 Institute