AI’s awful week is a peek into the future

Scott, Tech in AsiaOct 26, 2026

The Prompt

Welcome to The Prompt, your guide to the biggest issues and developments shaping AI, with Asia front and center. Want full access to all our AI reporting? Subscribe here.

Hello [subscriber],

There’s an old Silicon Valley joke that goes: If buildings were constructed the same way developers wrote software, the first woodpecker that comes along would destroy civilization.

That pecker has arrived, and it is AI.

Except that’s not really true. Everyone assumes it’s AI, but it’s still people that cause problems – AI is just uncovering those issues.

Last Wednesday, researchers from two independent security firms – US-based Calif and Israeli-American company Wiz – announced that they had found vulnerabilities in a decades-old version of PHP, the software that handles security credentials for an estimated 70% of all websites. These flaws, which were discovered via AI tools, could be used by hackers to completely take over a web server.

This resulted in no small amount of panic globally, with some of the offered solutions arguably exacerbating the panic. At least one cybersecurity company has been advising firms to checks note> take everything offline.

Image credit: Made by Ulla/Tech in Asia with the help of AI

Sure, let’s do that. After all, governments and companies have been pushing for years to take everything we do online. So let’s just turn it all off – no problem.

I expected a much better response, although kudos to the IT teams around the world who worked to roll out a patch.

But this isn’t the first time that such a vulnerability has been detected, and it won’t be the last. It was also inevitable – humans have been writing buggy code for decades.

That AI is able to find these previously unknown bugs is a blessing, albeit an incredibly expensive and potentially dangerous one. But it is better to find these errors now instead of when something bigger – like quantum computing – comes along.

The panic, however, just adds to the growing number of reasons why people are freaking out over AI. Many people don’t understand that many of the problems “caused” by AI are nothing more than this new technology unearthing our own mistakes. And while the cybersecurity error was bad, it’s nothing compared to some of the issues waiting on the horizon.

Ride that bandwagon, baby!

Let’s start with the money issue. Humans – not AI – have been dumping investments into chips, data centers and models development for almost three years now, and no one is making a profit.

Case in point: Anthropic announced Wednesday that it had lost US$80.7 billion in the first half of this year. I’d add more examples, but this is a newsletter, not a book.

Stock market investors have finally woken up and taken notice, leading to another mass sell-off. From October 6 to 23, Nvidia has been down about 11%, although many other semiconductor firms on Nasdaq have seen their stocks drop over 20%. The stocks of top data center firms have slid by as much as 30%.

Google, which owns Gemini, is down about 7%. And OpenAI and Anthropic? Trick question. They’re still private companies, although Anthropic is still considering an IPO.

What’s worse (at least to someone like me who remembers the Great Shitstorm, aka the 2008 financial crisis)is that an unknown amount of the investment here – the billions spent on AI chips – are financed, meaning the debt will linger even after the bubble bursts.

Courts in the US are already scrambling to handle the bankruptcy of Character.AI, a chatbot company with an estimated US$50 million to US$100 million in liabilities. There’s also speculation about CoreWeave: The US data center operator has about $20 million in bond debt, and many are worried that it will default on this loan.

My concern is what happens here in Southeast Asia. If demand for GPU rentals tanks because of a downturn in the markets, who gets left holding the debt for the data center boom across the region?

Forbes has already started referring to this as “Too-Big-To-Fail 3.0.”

And just to clarify, this isn’t an AI-induced issue. Humans did this.

Where this is all going

Adding to my list of concerns is a question I heard asked at a conference earlier this month: “What do we need these data centers for, anyway?”

I hate this question because after a few weeks of thinking about it, I have to admit that the person might have a point.

AI was originally designed to make human lives easier. The data centers were built for that purpose – at least, that was the idea that companies sold us on. AI was supposed to replace mind-numbing work like processing data, automate basic research, and find patterns in systems that humans could then use to create better things.

To do that, though, we need to train people on how to use AI, establish businesses that would utilize AI agents, and create datasets that help us save time by reducing complexity. That should have been the goal of every company and government out there.

But judging from everyone I’ve talked to, that hasn’t happened at all.

Instead, we use AI to have conversations with machines, create microdramas and videos instead of paying for human creativity, and find new ways to rob each other.

Our new worldwide infrastructure project is in danger of going off the rails, so we need to change course quickly. Otherwise, all this panic over cybersecurity, market meltdowns, and environmental costs will have been for nothing.

AI isn’t a toy; it’s a tool. Most of the problems surrounding it are caused by humans, not the technology itself. Humans need to find a way to make AI work for us, not against us.

In the meantime, check out our recent AI coverage:

Scott Shuey, senior editor

Catch up on past issues of the newsletter here.

P.S. Don't miss out on the biggest tech news and analysis. Add [email protected] to your address book, contacts, or safe sender list. Or simply move us into your inbox.

Tech in Asia thanks our partners for supporting our newsletters. Sponsorship has no influence on editorial content.

You can also remove yourself from all Tech in Asia emails over here.