Hannah's web log

I was recently in a situation where I needed some Powerpoint slides. I was strapped for time. I had done a lot of work on a project, and I just needed to make a presentation about it. The slides weren't super important - it was just to support a casual 10 minute preso to an internal audience.

I was the perfect candidate to use Gamma – an AI Powerpoint builder that lets you upload your notes and it turns it into an attractive slide deck.

I concede that generative AI can give you a "good enough" output for a lot of text-based tasks based on a descriptive prompt. Alberto Romero's great essay Bullshit tools for bullshit jobs points out the subversive dignity of using AI to outsource one's meaningless labour. However I think you lose something when you outsource writing, even emails and work reports. Doing it long term will probably turn your brain into a perfectly smooth sphere.

But then again, unlike writing, dicking around in Powerpoint is not a task that builds character or knowledge. I've made many a Powerpoint, and the time sunk into resizing text boxes and fiddling with font sizes while a public speaking obligation looms over you is no fun. Unlike the writing process, I had no moral or educational opposition to shortcutting this shit.

So I made a Gamma. I watched it render a lovely presentation with a consistent font theme and colour scheme, nice diagrams, that captured the main points of my written content and distilled them into 8 slides. Usage was disclosed with a "made with Gamma" badge that appeared in the corners. I just had to do a few minor tweaks with the easy-to-use editor – and I was golden. This tool had just saved me hours of busywork, right when I needed those hours.

It's moments like that where I wonder if and how I should be doing more to use AI to outsource the mundane parts of my job. The promise of AI productivity shines most in the lazy high of having just generated a "good enough" output. Are manually made PowerPoints a thing of the past?

The illusion was shattered on presentation day. I delivered my slides, no dramas. Another group presented after me. They had clearly used Gamma as well. I could tell because their presentation, despite using a different "theme", looked exactly the same as mine. The same flow, the same balance of elements, the exact same graphical concept diagrams. What had initially appeared stylish in my first ever encounter with the tool had already lost its appeal. I had for some reason assumed my presentation was unique, customised to my content. Not so – it was wrangled into the same template as everyone else. I imagined a future where everybody made their slides in Gamma, and they all looked exactly the same, and what is the point. Most AI "productivity" tools suffer this problem. Everyone thinks they're getting away with something, but you end up with homogenised blobs of nothing.

Going forward, I don't know if there's a middle ground between saving time and maintaining my voice in presentations. My thinking was let the human focus on the content and let AI handle the polished form. But seeing AI's uniform aesthetic output in practice changed my perspective.

I had told myself the slides didn't matter – that they weren't "real" work. The thinking part was done and the rest was just packaging. But that turned out to be a false distinction. Translating your ideas into something others can engage with is part of the work. It's not just about making things look good, it's about deciding what deserves emphasis, what can be left out, how tone and structure shape understanding. It's not busywork, it's communication.

Even "working backwards" – letting AI produce something and then editing it, is a productivity trap. You get a dopamine hit from instantly getting something that looks "nearly finished" rather than a blank page – but if you're truly committed to making it something you are proud to present, it can require more time and tweaking than if you'd just done it yourself from the beginning. AI productivity tools promise to free us from drudgery, but they often just trade one problem for another. Instead of spending time on execution, the more rewarding process, you spend it on revision, the painful process, and one that is harder for your brain to access if you didn't do the actual execution. The time savings are offset by the creative cost - a cost that may only become apparent when you're sitting in a room full of people who made the same trade-off.

Comment