24 Comments
User's avatar
Herbie Bradley's avatar

> Because often enough my guests can’t just answer pretty complicated fractal questions in a satisfying way on the spot

Given this, perhaps you should try a dialogue with a guest over multiple emails, condensed into an essay. This could work in cases where you feel there is benefit from going deep and the guest has the time to dedicate to writing arguments and responding to yours?

Neil's avatar

Any more book reviews coming our way? I quite liked the one on Robert Caro's LBJ series, and Notes on The Prize was a good read as well.

Dwarkesh Patel's avatar

I'll consider doing more of these, thanks!

David M. Todd's avatar

Dwarkesh, the work you are doing is awesome. As was pointed out in another comment, please don't underestimate what others are learning from the podcasts. You have many very intelligent listeners, I'm sure, but few have the deep understanding of any particular topic that you and your guests have. Your hours of preparation allow for conversations that are deep, detailed and highly-informative for the rest of us. So while you may not be getting that extra kernel of knowledge you are looking for, the rest of us are getting a popcorn bag of full of insights and understanding.

And extra congrats for the string of unbelievable guests and conversations you've had of late. A+ stuff.

Sean Sullivan's avatar

This is great stuff, Dwarkesh. You're a shining example of the pay off one can get by just working hard at something other people are half assing.

Aditya Nath Jha's avatar

>> The main angst I’ve kept receding back to over and over is, “Okay what did I actually learn from this interview? And if I didn’t get that much concrete insight and understanding out of it, despite a week+ of research and hours of conversation, what hope is there for the audience? And if no one learned anything, what the fuck are we doing here?”<< Your audience doesn't have the benefit of your prior research; so your questions themselves are often a learning, even if the answers are not fully satisfying. In addition, if you do a thousand conversations and only ten survive as meaningful for future generations fifty years from now, that in itself would be worth the endeavour.

Elliot Olds's avatar

RE: helping people actually learn valuable things from podcast interviews:

The podcast that has probably given me the most valuable insights is Econtalk, and if I had to guess what Russ does that causes that I'd say that he's willing to go really slow and linger on one seemingly simple point for a long time and explore it from every angle.

I think in the moment it probably feels more entertaining to listeners to move on from topics faster, but doesn't leave much of a lasting impression because absorbing new ideas is hard. You already linger on points more than most other podcasters but I'd be curious what'd happen if you leaned into that even more.

Another thing I wonder is: how would your podcasts be different if you pretended the convo wasn't being recorded and your only goal was your own selfish learning during the time you talked to the guest? I think you're already way higher on the dimension of 'having the convo you want to have' than other podcasters but it seems like it'd be interesting to try to max that out.

Kevin Sukmin Son's avatar

The spirit of enlightenment and thanksgiving truly resonates.

In 2025, I can't think of anyone else who has had a bigger impact on AI discourse, particularly in a forward-looking and productive way.

A multi-channel approach makes sense going forward since it's ever more difficult to maintain foresight and to communicate. Some amount of personal sharing helps as well. It's good to share the ingredients of "taste."

Ignore the haters and keep up the good work!

Alex Volkov's avatar

Love these self reflections Dwarkesh! Thanks for posting them 🔥

Siddharth Manuja's avatar

The thought I had after listening to the podcast for a while had been that this has the potential to be the defining voice of the coming era (had shared the same in my email). Glad to see your vision for the podcast.

In 2026 - I would love to see the pod focus more on the real world impact (both guests / content, and the audience). Someone (really, a number of people, who hopefully would be aligned) is going to have to serve as the bridge between labs / big tech and the real world. It is definitely not going to be the politicians. It is not going to be tech leaders. It is not going to be the regular media. It is not going to even be most of the tech / new media (Conflicts, trust gaps, knowledge/understanding gaps). Fine needle to thread - but you are well positioned

Conor Bronsdon's avatar

The most impressive part of this whole post is that you can regularly get enough friends together to play Blood on the Clocktower!

Excited for the new essay era - and interested to see you experiment with more long-form narrative conversations on the podcast as well.

Rousol Altimimi's avatar

You've made me fall in love with AI. It feels like concrete philosophy that is impacting us TODAY. We're basically Plato.

Kian Kyars's avatar

Can't wait for you to find the next Sarah Paine

Rob Neuhaus's avatar

Zeynep Tufekci have an amazing and incredibly bleak talk about AI and the future at neurips today. I think she'd make a great guest. I'd love to hear you two disagree about basically everything.

Steeven's avatar

The Twitter drama was hilarious and I had a great time

Oz's avatar

Well done. Been an inspirational journey to watch from afar

Mihir Dave's avatar

This is cool as an opportunity to experiment with format. Perhaps a mix of pre-written questions provided in advance and spontaneous questions, interstitials where you break to elaborate on something like pablo torre's pod, exploring some concepts with the guest before you go live to give them a framework, and essays that frame a pod and orient the guest are some possibilities.

Mihir Dave's avatar

And perhaps having multiple guests paired selectively could work too