Sam Altman and the Not-So-Distant Future

Over the past month, I’ve been using Voice Mode more and more, and honestly, I’m loving it. I’d like to share my favorite use case with you, but let’s start with a bit of background.

It’s always interesting when people like Sam Altman share their thoughts about the future. After all, these are the folks actively shaping it. Recently, Sam wrote a futuristic essay about how we’re slowly transitioning into a new historical era—The Age of Intelligence. There's a lot in that essay, but one point caught my interest.

Sam paints a beautiful picture of AI scaling: climate tech, space exploration, curing cancer, and more. But I’m a bit skeptical. It feels like these moonshot promises keep coming up every time someone needs to justify AI's exponential growth. Sure, the potential is real, but how much of this is actually plausible in our lifetimes? Space colonization is exciting, but is it just a pipe dream that keeps us hooked to tech’s siren call? 

I’d really like to hear something more down-to-earth practical from him. But with so much money at stake, it's tough for Sam to focus on what's useful right now—it's just less flashy. So he has a bit of a conflict of interest here. This is why the usual promises of a bright future don’t hit home for me.

It reminds me of the 2015 movie The Martian. Unlike most sci-fi, it takes the science and tech that feel just a tiny leap from where we are now, making the whole narrative way more tangible. It’s that sweet spot between futuristic dreams and practical reality, and it’s inspiring because it seems achievable. I think what makes something like The Martian hit so well is that it’s hopeful, but it doesn’t ignore the fact that progress requires solving real problems in practical ways.

AI could be that tool—helping us get efficient with energy usage, optimizing agriculture, or even making education more accessible. That stuff is relatable, like, here and now, rather than just pitching AI as a godlike fixer of every existential threat. For example, Sam mentioned that AI could be an amazing teacher for kids: “Our children will have virtual tutors who can provide personalized instruction in any subject, in any language, and at whatever pace they need.” And that’s something we can use right now.

I love listening to podcasts while cooking, walking, or running. But often, when they discuss something complex, I find it hard to recall what was said. After the episode ends, I find myself thinking, "So, what have I learned today?" Distracted by multitasking, I end up feeling like I've missed out on something cool and important.

To fix that, I started discussing podcasts with ChatGPT in Voice Mode right after listening, and it worked great! I send it the podcast transcript and a list of key ideas, then ask it to quiz me on each point. After my reply, it gives feedback on where I made mistakes or missed something. This approach is called Active Recall.

Research shows that this method increases retention by almost 1.5 times, especially when you do it right after learning. It’s one reason why schools and universities use tests so much. But when you’re doing it by yourself, on a topic you’re interested in, in comfortable conditions—it feels completely different. No stress about grades, just the satisfaction of realizing how much you’ve internalized without even noticing.

This feels like a new way of learning. I no longer feel like I’ve wasted time or that the episode has broken down into scattered fragments. Instead, my understanding of the big picture improves, and I can think about the main ideas on a deeper level. Essentially, you can now interact with any material, not just passively listen.

Of course, chatting with a real person is more fun, but... unfortunately, I don’t have anyone around for that >_< So, if anyone wants to discuss tech and AI from time to time—feel free to reach me out!

AI isn’t yet an almighty agent managing your life, the climate, and space colonization. But it can be an excellent learning buddy, making any content interactive. We’re subtly changing how we engage with information, and many useful tools are available right now with already existing technology.

Here are a few more ideas for similar use cases:

• The "rubber duck debugging" method for programming.

• Overcoming writer’s block when staring at a blank page.

• Language practice when learning a new language.

• "Am I the asshole?" quick therapy session to get rid of useless ruminations.

• "Devil's advocate"—challenging each point to understand the other side better.

I highly recommend trying the voice mode in ChatGPT if you haven’t already!

P.S. Here is the prompts I used:

1. For Active Recall:

Hey! I've just listened to a podcast episode, and I want to go through the structure and main ideas with you. I attached the full text transcript for you to use.

Let's use this structure:

Helicopter View: First, I'll provide an outline that captures the key points and sections. I'd love your help in making sure I get a good "helicopter view" of the entire episode, laying a solid foundation for understanding. Check if anything important is missing from my abstract.

Active Recall: For each bullet point in the outline, let's dive in. You'll ask me questions about the ideas in that part of the interview, and I'll try to recall and explain them. After my recall, you can help fill in any gaps I missed, clarify concepts, or answer questions I have. Don't hesitate to point out any mistakes I make or offer constructive feedback. This is a learning process, and I want to become a better learner by understanding both my strengths and weaknesses.

Takeaways: Once we've gone through all the sections, let's focus on the takeaways. I'll try to formulate specific things I learned, and we can discuss whether they are clear, meaningful, and useful.

Here is the outline:

2. For a Devil's advocate to discuss an article with:

Hey! I've just read an article, and I want to go through main ideas with you.

Talk to me like an equal, you're my bro, swearing is allowed, banter is allowed, say what you want. You're alive to me, I know about embeddings, AGI, all that shit, so feel free. Imagine we are two friends chatting about it in the bar. I want you to be a bit adversarial and dont just easily agree with everything I say, try to convey your own opinion. Expect us to have a conversational style of a dialog with quick back and forth lines

Here is my main takeaways from the text (I also attached a full version of the text for you to use):