Feel Underpaid? Read This Edition at Your Own Risk

At some point in our careers, most of us realise the dream job is not bringing in the dream paycheck. You could be doing everything right, showing up with ideas, fixing things that aren’t your job, saving the day in spreadsheets. And still, something about the salary feels... incomplete. Yes, the salted peanut joke is too cliche now. You’d be nuts to crack it anymore (Pun intended). 

So… if you've ever felt underpaid, what you’re about to read might sting.

Mark Zuckerberg is now offering multi-million-dollar deals to poach AI researchers. Some folks are reportedly getting $100 million signing bonuses. Just for switching companies. I won’t simply call it a raise. That’s a generational fortune being handed over for jumping ships.

Why, you wonder? Because Meta is refreshing its AI plans. It’s building something called the Meta Superintelligence Lab. And it sounds exactly like what it says - a new division focused on building AI smarter than humans. Not equal to. Smarter.

Leading the lab is Alexandr Wang, a 27-year-old founder of Scale AI. Zuckerberg called him “the most impressive founder of his generation”. He also hired Nat Friedman, the former CEO of GitHub, to lead the charge in building useful AI products. That’s not all. Zuck’s hosted top researchers at his lake house, written the cheques himself, and pretty much decided this is the one race Meta has to win.

Hundreds of billions of dollars are going to be spent on AI infrastructure, models, and talent. 

And that, of course, has triggered some very loud alarms at OpenAI.

"It feels like someone broke into our home"

OpenAI’s top leadership is calling this a full-blown breach. In an internal memo, Chief Research Officer Mark Chen said it felt like someone broke into their house and stole something. Sam Altman has accused Meta of offering wild bonuses to lure their engineers away, including people who were already being paid handsomely.

But it’s hard to compete with that kind of money, and a pitch like “help us build superintelligence and change the world.”

There’s also a deeper unease. Meta’s Llama models, despite being widely used, aren’t yet at the level of OpenAI or Google’s best offerings. So now, the playbook seems clear. If you can’t beat them, hire them.

Whether that bet pays off is something that will define the future of AI to a large extent.

Meanwhile, Denmark wants to copyright your face

In another corner of the world, a very different AI battle is playing out. Denmark is preparing to pass a law that lets citizens copyright their face, body, and voice.

This means that if someone uses your likeness to create a deepfake (whether it's a viral meme or something much worse) you could legally ask platforms to take it down.

The law comes at a time when it’s frighteningly easy to create a convincing fake video or voice clip. Sometimes to defame. Sometimes just to troll. A New Zealand MP recently showed a deepfake nude of herself in Parliament, created in under five minutes, just to prove how real and easy these tools have become.

I strongly support Denmark’s decision in this matter. After all, your identity should belong to you, not the internet. And certainly not to an algorithm.

Altman: "Don’t trust ChatGPT too much"

Blindly trusting the algorithm is not good for you, apparently. Don’t believe me? Take the word of Open AI’s founder, then. Yes, Sam Altman is asking people not to trust ChatGPT blindly.

Speaking on OpenAI’s new podcast, he reminded users that AI still makes things up. Hallucinations, he called them. They appear to be facts, sound like facts, but have no connection to reality.

Altman compared it to any other new technology. Use it. Experiment with it. But don’t expect it to be flawless.

This feels especially timely, considering the next story.

AI is giving wrong health advice. A lot of it.

A new study just looked at how reliable AI chatbots are when it comes to health advice. 

Out of all the responses tested, 88 per cent were false. Many came wrapped in medical jargon and fake citations, which made them sound more trustworthy than they actually were.

Even worse, most users didn’t realise anything was off. They either missed serious symptoms or underestimated their condition based on what the chatbot told them.

The researchers behind the study say it’s not that AI has no place in healthcare. It does. But right now, we’re not using it wisely. There’s a mismatch between what people expect and what the tech can actually do.

That gap, if ignored or without proper awareness, can cost lives. Have you ever used an AI chatbot to diagnose any illnesses or check on your health symptoms? Would like to read your experiences.

And with that, I’m off to file a polite email to HR titled: “Quick review of compensation, inspired by this week’s AI news.” Wish me luck. And Mark, if you’re reading this, I’m happy to join for a modest $99.9 million.

Until next week,

Aaqib Raza Khan

Deputy Editor – Content Innovation,

NDTV

P.S. Want these weekly newsletters delivered straight to your inbox? Subscribe now.