Tech Stories

Issue #65

Hello Reader,

I was in San Francisco earlier this week for HubSpot's Inbound 2025 conference, and my head is still spinning from how AI is changing content and marketing. I thought I was decent at using AI, but it turns out I’ve only been scratching the surface.

For today's newsletter, let's talk about how Singapore is working to redefine the rules of data centres, contextual engineering, and optimising content for AI search engines. And finally, I explain why it pays to listen to those with no vested interests.

Jurong Island: Singapore's next data centre hub?

In Singapore this week, probably the biggest news that most publications missed was a report on Zaobao about JTC's plan to study the risks and feasibility of operating data centres on Jurong Island.

You might wonder what's the big deal about a feasibility study. This is Singapore, you see, and such announcements are never made lightly. Such news in a mainstream newspaper means data centres on Jurong Island is almost a certainty.

Now that it's public knowledge, I can talk freely about it. Let's just say it jibes with what I've previously heard from data centre leaders.

Anyway, why Jurong Island, given the inherent risks of critical infrastructure sitting on land right next to potentially hazardous petrochemical plants?

My take, which I elaborated in my post, is this: We've always built data centres based on the needs of the day, be it speed-to-market, availability of resources, or land. And if you ignore your knee-jerk reaction, Jurong Island actually ticks all these boxes when it comes to the 300MW of data centres announced back in May 2024.

Are you still at level 1 AI competency?

Sorry about the quality - had to deskew this photo

I've previously shared how I use AI to write. In fact, I currently have paid versions of ChatGPT, Claude, Gemini and Perplexity. So I'm no newbie. But some of the concepts I saw this week simply blew my socks off.

On reflection, practically all of us are operating at what's called "Horizontal AI" or Level 1 competency. This basic use often produces mediocre results, also known as "AI slop." Why? Because large language models are the sum of their training. When you average out the exceptional and the poor, you're left with mediocre outputs. Many people give up here when AI fails to produce what they need.

Good prompts, examples, and iterations help, but this takes time and often isn't replicable across chat sessions. In his talk "40-min AI sprint: branded AI training and take-home playbook," marketing agency founder Kevin Barber was candid: he wouldn't want to pay an employee to spend hours a day "prompting AI."

Moving to level 2 AI competency requires going from prompt engineering to contextual engineering. This means incorporating context to an extreme degree, with five distinct layers that Kevin shared. If I were to summarise: effective AI use combines systems thinking, structures that inject empathy, and extreme contextual detail.

Level 3 competency involves AI agents executing step-by-step workflows. According to Kevin, this entails the use of AI agents executing step-by-step workflows. He uses this for full-stack marketing, so your mileage might vary for other tasks. But similar workflows can be adapted for your specific needs.

Even level 3 has limits. As Kevin observed: "[AI] may never sound exactly like you." So I'll keep writing these newsletters by hand.

Ranking with AI search engines

Another topic that attracted loads of attention is AI engine optimisation (AEO) - ensuring that brands are mentioned when users query AI search engines.

What's the big deal about this? Why, it might be because 60% of searches now end without a click, courtesy of users searching with AI and features like Google's AI Overviews.

It doesn't help that the "Awareness" portion of digital marketing has scattered. Today, buyers are on scores of disparate platforms: TikTok, YouTube, Reddit, or listening to their favourite podcasts. With web traffic tanking and the link from search to company blog broken, marketers are panicking. How will they bring in leads, much less convert them?

This is where AEO comes in. There's no singular strategy, but rather a constantly evolving list of best practices and techniques to get picked up by AI search engines. Confusingly, some of what works contradicts traditional SEO.

For example, content freshness matters more for AEO (one year or less) while SEO favours established content. To succeed with AEO, at least one speaker talked about making regular updates to content - using AI - to keep it "fresh."

You can read more AEO tips here.

In closing: Don’t ask the barber if you need a haircut

By the way, my post pointing out that the typical enterprise with their CPU-heavy workloads is unlikely to require liquid cooling met with strong denial from at least one responder. After a couple of fruitless exchanges, I realised one thing: Don't ask the barber if you need a haircut.

To be clear, I have a soft spot for liquid cooling, and in fact have made it a point to learn about the various liquid cooling solutions out there. Unfortunately, many of the voices clamouring for 130kW racks today and how every data centre must be fully liquid cooled tomorrow have deeply vested interests.

It's one reason why I keep doing what I do: We need to hear both sides of the story, and not allow biased statements get repeated so often they become accepted as the truth. Of course, I don't know everything, and it's possible that I could be wrong. So please keep talking to me - I'm going to need all your help to stay up to date.

As usual, you can reach me by hitting reply.

Regards,
Paul Mah.

You can now comment on stories published on techstories.co. If you received this newsletter from me, your account was created automatically.

All you need to do:

Scroll to Member discussion and sign in with the email used for this newsletter.
An email message with a magic link is sent immediately. Click to log in.
Comment away.

Spotlight