Podcast Episode

Meet the most AI-pilled CEO on Earth

ACCESS
Jul 30, 20261 hr 12 minEN
Hosts
EHEllis HamburgerAHAlex Heath

Who’s talking

7 speakers
Dan Shipper38 min · 54%
Ellis Hamburger17 min · 24%
Alex Heath15 min · 21%
Speaker 629 sec · 1%
Speaker 528 sec · 1%
Speaker 24 sec · 0%
1 other1 sec · 0%

Transcript

154 turns · 1 hr 13 min
EH
Ellis Hamburger0:00

Support for this show is brought to you by Salonis. Does it ever feel like you're being told to wave a magic AI wand and everything will just get better? Sure, AI can chat and summarize, but what about big business problems like rerouting stock through a canal that won't unblock? Enterprise AI needs context for that, so Solonis provides it. The Solonis context model gives AI operational clarity, so agents know how your unique business runs and how to prove it. Meet the model at clonis.com slash

Speaker 20:31

context. Learn more at Accenture.com slash Spotify.

AH
Alex Heath1:06

Before we get into today's show, a quick note. This is the final episode of Access. You can keep following me at sources.news and Ellis at meaning.company. But don't go anywhere. The show's feed is going to live on and you're going to be hearing a lot more from me here very soon. We'll

EH
Ellis Hamburger1:23

talk more after the show about why we're winding everything down, what we're proud of, and what's next for the both of us.

DS
Dan Shipper1:29

There are some interesting nuances to evaluating models. A model that you don't like on day one can be a trash model. If you use it on low or medium thinking and you get rid of your skills, it's way better.

AH
Alex Heath1:40

Today, we have a guest who has been living at the edge of how AI is changing work, media, and the internet. Dan Schipper, CEO of

DS
Dan Shipper1:47

Evry. You have to be willing to waste tokens. I was testing 5.6, so I woke up the next day to a message from Ariel saying, did you spend 2 billion tokens overnight? I was like, fuck.

EH
Ellis Hamburger2:00

Dan breaks down why he thinks we'll all spend more time inside AI coding agents, how every is building agents to scale human taste, and why the real opportunity is not automating everything, it's figuring out what humans do next.

AH
Alex Heath2:13

Plus, his two phone system, billion token experiments, and why Siri may become a much bigger threat to ChatGPT than people realize. Welcome to Access.

EH
Ellis Hamburger2:26

I love your little every mic cube thing. Is there like an industry term for that that I'm not aware of?

DS
Dan Shipper2:32

I have no idea. This is like I actually I was thinking about this because this is the first time I've seen this. And I thought you guys would like this because it's sort of it's one of those things where you spend like six and a half years like pushing a boulder up a hill and you're like. um every single thing that happens you did and then and then eventually it starts rolling down the hill and you show up to your like recording studio and one of these is there and you're and you're like i had nothing to do with this but i love it you know and like that's a really nice feeling

AH
Alex Heath2:59

that's when you know you've made it

DS
Dan Shipper3:01

that's when you know you've made it is you got the little cube and you didn't do the cube

EH
Ellis Hamburger3:07

But was it an actual member of the team that did it? Or was it a gun for hire hired by one of your agents who just swooped into the office, added that to your mic? It was

DS
Dan Shipper3:17

obviously Codex.

EH
Ellis Hamburger3:20

I mean, Poke did this the other day. Did you guys see this? They had Poke Humans launch on July 4th, and you could have them go do Task Gravity things, but I was too afraid to use it.

AH
Alex Heath3:29

Poke bought by Cognition. Shout out Marvin, former guest. A sentence that only a certain cohort of people will understand. Cognition buying Poke. Those are two

Episode notes

Alex and Ellis sit down with Dan Shipper, CEO of Every, to talk about what it actually looks like to be the most AI-pilled company that isn't a frontier lab. They get into how Dan uses ChatGPT Voice to read Heidegger at 6am, why he accidentally spent 2 billion tokens overnight, and why he thinks OpenAI just pulled off something that almost never happens in tech. They also discuss what the new Siri could do to ChatGPT, and whether that's the Instagram Stories moment for the assistant market, why the developers building at the edge today are a 12-month preview of what every knowledge worker will be doing next, and what "after automation" means for people whose job is to have original ideas. Then Alex and Ellis reflect on one year of Access, why they're winding the show down, and what's next for both of them.

Follow ACCESS on Instagram:

⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://www.instagram.com/accesspodcast/⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠

Follow Alex:

Showing a preview — retrieve the full transcript via the Exa API.

Powered by the Exa API