Dear Reader,

Just about 2 weeks ago, I received my PhD examination reports.

Some 30 pages of critique, just 6 days before the defense.

Students often have weeks or months to prepare — my uni gave me six days.

On top of that, I hadn’t touched my research in way over 4 months (turned in March, and my son was born in early May).

While one examiner wrote an extremely positive letter, another made numerous smaller suggestions and questions. And then there is the third examiner (quite a prominent one too) who started with the words: “In short, the thesis is framed around a range-shifting view not supported by the empirical evidence in New Zealand”. (My research is about forest ecology in New Zealand, btw).

He dismissed my entire thesis on page 1 and used the remaining 18 pages to pluck apart nearly everything!

Wow, was I angry…

(In all honesty, he wasn’t wrong. It’s just that, in my opinion, his arguments generalised insights from a handful of narrow ecological experiments onto all of New Zealand’s stunningly diverse forests.)

The fight was on…

So in 6 days I had to run an entire literature review on “Climate Change in New Zealand”, and I had to pull every trick to get it done in time. (I kid you not, I also got a fever for a few days over that weekend).

Two things saved me during that time:

1. My collection of notes, and especially the project documentation for the papers in the thesis. These notes contained dozens of (unpublished) decisions the reviewers questioned, making it easy to rebut their criticisms. (See this blogpost and that for some examples of my direction-result-progress system)
2. The AI + Obsidian integration I have been building for the past 3 months. This allowed me to find and aggregate 100s of nearly forgotten insights, scan and cross-compare studies the examiners mentioned (sometimes in 2-3 conversations in parallel), and identify weak points in my own argumentation. It is so beyond the "ChatGPT in a browser tab" experience.

Here is an example of how AI helped me find a screenshot of a plot I saved more than 2 years ago in 2024. No chance I could have remembered it in 2026!

You have probably felt that too…

I think these learnings are applicable to almost any academic. Since you are reading this, you probably at some point

• Felt like you have too little time for too much work
• Were overwhelmed by the sheer amount of literature to work through
• Could not find notes or remember facts you clearly wrote down somewhere

That’s exactly what a connected notes-and-AI system fixes!

Academia is genuinely challenging, but I think increasingly we do have systems and tools to help. In the future, using AI might even be expected of an academic, but for now it is an unfair advantage if you bother to learn & set up this technology.

Three kinds of academics

AI does to our minds what processed sugar does to our bodies. It feels too good to resist, and we’re wired to crave the easy calories (or easy ideas). That’s why using it well takes a conscious choice.

Right now, I see academics reacting in three ways:

1. Those who surrender. They let AI reason, generate ideas, and summarise the literature for them. In fact, most academic tools want you here, but study after study shows the cost: a slow erosion of critical thinking.
2. Those who refuse. They have many good arguments on their side (psychological, ethical, or environmental). But the genie is out of the bottle, and in a world (unfortunately) obsessed with productivity, refusing outright could mean falling behind in the long run.
3. Those who stay in control. They hand AI the searching, extracting, and aggregating - the reproducible “grunt work” - but use the results for better ideas and deeper insights, keeping the thinking their own. AI remains a key strategic partner to speed things up, but stays in the background.

Which one do you want to be?

I want you in the third group. So do you, I suspect — or you wouldn’t have read this far.

Over the next 3 days, I want to show you what this means (at least from my AI research practice), and how you can implement this system.

Until then, wishing you my very best,

Ilya, The Effortless Academic

Tomorrow, I’ll tell you how my PhD defence ended and reveal how you can start building the exact AI system that transformed my academic work! You will learn why an AI agent is nothing like the AI chat you’re used to, and why we’re truly at a technological inflection point. If you can't wait to get started, click here.

This email is part of a 4-day series on AI agents. If you don't want to know about it, unsubscribe from this series.​
Want to leave forever: Unsubscribe from everything (never hear from me again)