DAN SHIPPER (Every CEO) - Introduction & Proof/Anecdote Demo

=== DAN SHIPPER (Every CEO) - Introduction & Proof/Anecdote Demo ===

(00:00:03): Hello, everybody.

(00:00:04): I was just muted, but I'm here now.

(00:00:10): Please let me know if you can indeed hear me.

(00:00:14): So excited to have you here.

(00:00:15): Welcome to Vibe Code Camp.

(00:00:18): This is an all-day marathon where we watch the best Vibe Coders in the world do their thing.

(00:00:23): I'm Dan Shipper.

(00:00:24): I'm the co-founder and CEO of Everee.

(00:00:25): I will be demoing first.

(00:00:30): If you're here, if you're watching, we're going to be here all day.

(00:00:32): It's eight hours of streaming.

(00:00:35): There's going to be a lot of really amazing stuff.

(00:00:37): So please stay tuned.

(00:00:39): We have an agenda.

(00:00:40): If you want to see what's coming up, go to every.to slash agenda.

(00:00:45): and follow along.

(00:00:47): Maybe also while you're here,

(00:00:49): if you're here,

(00:00:50): fire up some Cloud Code instances and some Codex instances and start making some

(00:00:54): stuff.

(00:00:55): Please share in the chat.

(00:00:57): Hopefully this is one of those things where as you're going through your

(00:00:59): day-to-day,

(00:01:00): you can just tune in and out and just be working on stuff and also hanging out with

(00:01:04): us

(00:01:05): watching the best live coders and vibe coders in the world do their thing.

(00:01:07): So first I want to just introduce you to every,

(00:01:10): then we'll go through the agenda and then I'm going to demo some cool stuff.

(00:01:13): I've been building some really cool stuff.

(00:01:14): So I'm going to start with my workflow.

(00:01:19): What is Every?

(00:01:20): If you're here, hopefully you know what Every is.

(00:01:22): If you don't know, now you know.

(00:01:24): Every is the only subscription you need to stay at the edge of AI.

(00:01:28): It's every.to.

(00:01:29): You should go there and subscribe.

(00:01:30): For paid subscribers, we do three things.

(00:01:32): We have three parts of the business, ideas, apps, and training.

(00:01:36): On the ideas side, we have a daily newsletter.

(00:01:39): We cover every new model and every new product when it comes out.

(00:01:42): So on the day it's released,

(00:01:44): you have a hands-on vibe check from us on what's good and what's not.

(00:01:47): We also do long-form essays on how we use AI in our work and in our lives.

(00:01:51): So if you want to keep up,

(00:01:54): if you want to stay at the edge,

(00:01:56): the ideas part of every will keep you there.

(00:01:58): We also have a suite of apps that we build for ourselves to help us work better with AI.

(00:02:02): So we have an app called Cora, which is an AI email assistant.

(00:02:06): We have an app called Monologue,

(00:02:07): which is a smart dictation app,

(00:02:09): sort of like Super Whisper and Whisper Flow.

(00:02:10): we have spiral which is a ai ghostwriter with taste and we also have sparkle which

(00:02:15): is an ai file organizer if you are someone who wants to stay at the edge we build

(00:02:20): these tools for ourselves and uh it all they all come with one under one

(00:02:25): subscription under the every subscription

(00:02:27): The last thing that we do is we do training.

(00:02:30): So we have camps and courses camps like this one.

(00:02:32): Usually these are paid this today because it's such a big marathon.

(00:02:35): We're doing this for free for everybody.

(00:02:37): But tomorrow,

(00:02:38): for example,

(00:02:38): we have a camp with all the internal every team talking about how we build agent

(00:02:43): native software inside of every.

(00:02:45): So if you're interested,

(00:02:47): if any of this peaks your interest,

(00:02:49): check out every and get ideas,

(00:02:52): apps and training all for one subscription to stay at the edge of AI.

(00:02:55): All right.

(00:02:56): Let's talk through the agenda.

(00:02:58): So I'm going to pull that up.

(00:03:00): If you want to look at the agenda yourself at any time during the day to see where

(00:03:02): we are,

(00:03:03): go to every.to slash agenda.

(00:03:06): And let me just pull this up for you guys.

(00:03:10): Hopefully you can see this.

(00:03:13): OK, this is a stacked lineup.

(00:03:16): This is going to be a crazy, crazy day.

(00:03:18): So to start, I'm going to demo some stuff for you.

(00:03:24): I'm Dan Shipper, co-founder and CEO of Every.

(00:03:26): I'm going to talk about a few apps I've been building,

(00:03:27): a few agent-native apps I've been building that I'm really excited about.

(00:03:30): Then we're going to have my friend Ben Tossel from Ben's Bytes.

(00:03:34): Ben is an incredible person, incredibly smart, and definitely at the edge.

(00:03:40): He's working at Droid right now,

(00:03:42): in addition to being part of Ben's Bytes,

(00:03:43): and he's going to show us how he's shipping lots and lots of software,

(00:03:46): even as a non-technical person.

(00:03:48): Then we're going to have my friend Ash from Harf.

(00:03:51): She's going to come on.

(00:03:52): She's amazing.

(00:03:53): She's going to show us how she is building Chrome extensions to work on X.

(00:03:58): Then we've got Ryan Carson,

(00:03:59): who was until recently at AMP,

(00:04:02): and now he is building his own company.

(00:04:04): So he's going to demo Ralph Wiggum for us.

(00:04:07): He's awesome.

(00:04:09): So that's going to be the morning block.

(00:04:11): A lot of really, really cool people.

(00:04:13): Then around lunchtime, we're going to have a bunch of internal Every demos.

(00:04:16): So we're going to have Natalia and Nitesh from Every who work on our consulting side.

(00:04:22): We work with a lot of big companies helping them to implement AI in their work and

(00:04:25): in their lives.

(00:04:27): Austin, are you seeing this?

(00:04:29): I just want to make sure you guys are all seeing the slideshow.

(00:04:36): austin you are you're good okay cool so sorry continue just uh just wanted to make

(00:04:42): sure we're all on the same page so uh we've got internal every demos from natalia

(00:04:46): and natesh they work on our consulting business um and they're going to show they

(00:04:50): have really seriously totally changed how they use how they how they run our

(00:04:55): consulting business which is a seven-figure business working with some of the top

(00:04:58): hedge funds and pe firms and fortune 500 companies in the world and it's like

(00:05:01): completely just

(00:05:03): It's mostly just cloud code at this point doing stuff,

(00:05:05): and so they're going to demo how they did that and how they're operationalizing

(00:05:09): cloud code inside the consulting business.

(00:05:11): Then we've got Katie Parrott,

(00:05:12): the one and only Katie Parrott,

(00:05:13): who is one of our lead writers,

(00:05:16): and she's going to show us how she's using the cloud code desktop app.

(00:05:22): and using Cloud Code in her editorial writing workflow.

(00:05:26): Then at 1230, we're going to have my good friend Nat Eliason, who hopefully you all know.

(00:05:31): If you don't, you should know him.

(00:05:32): He's awesome.

(00:05:34): He's a writer.

(00:05:35): He's an author.

(00:05:36): He's also a course guy.

(00:05:37): He's done a lot of really, really cool courses over the years, some on coding with AI.

(00:05:41): Back in the day, he did a Rome research course.

(00:05:45): uh he's he's a best-selling author and he also used to run a marketing agency he's

(00:05:49): amazing he's been doing a ton of like really really cool interesting stuff uh with

(00:05:53): claude uh multi-agent so he's going to come and talk about that

(00:05:57): Then we've got Tina He from Pace Capital.

(00:06:00): Tina is awesome.

(00:06:02): She's going to talk about Vibe coding as creative practice.

(00:06:04): You should check her out.

(00:06:06): Then we've got Paula from Portola.

(00:06:08): Paula is amazing.

(00:06:10): She,

(00:06:11): if you haven't read it recently,

(00:06:12): she just published a case study that she wrote with OpenAI where they've,

(00:06:16): she works at Portola.

(00:06:18): They do, they make a little alien character that has like 200,000 monthly active users.

(00:06:23): That's all AI.

(00:06:24): So she's going to talk about

(00:06:26): how she does iOS dev with Claude.

(00:06:29): She just published a big case study on it, went super viral.

(00:06:31): She's awesome.

(00:06:34): After that, we're going to have CJ Hess from 10X.

(00:06:36): If you have heard of 10X, they're one of the top AI consulting companies run by Alex Lieberman.

(00:06:43): So he's going to talk about Figma and basically a Figma-like flowchart and UI tool he's built.

(00:06:50): So that's really awesome.

(00:06:52): Then as we get further into the afternoon,

(00:06:54): we've got some really,

(00:06:55): really amazing people coming up.

(00:06:56): We've got Logan and Amar from Google.

(00:07:00): So they're going to talk about all of the stuff they're doing inside of Google AI

(00:07:03): Studios and show us some demos of how they're building things.

(00:07:06): Then we've got Jeffrey Litt from Notion.

(00:07:09): He's going to demo some amazing stuff,

(00:07:12): workflows that he's been building using Notion as a Kanban board for Cloud Code.

(00:07:16): Then we've got Kevin Rose, the one and only Kevin Rose.

(00:07:21): from True Ventures,

(00:07:22): and he is going to be interviewed by the one and only Kieran Klassen,

(00:07:26): who is the GM of Cora at Every,

(00:07:28): and they're gonna talk about compound engineering.

(00:07:30): That's gonna be a big, big discussion.

(00:07:32): It's gonna be so fun.

(00:07:33): You should definitely be there.

(00:07:35): Then we've got Tariq from Anthropic, and he's going to talk about inside Cloud Code.

(00:07:41): He literally works on Cloud Code,

(00:07:43): and he's going to talk about how Cloud Code works,

(00:07:44): how they're building it,

(00:07:46): and then he's going to do some demos of things that he's building and his workflow.

(00:07:50): Then we've got Naveen, who is the GM of Monologue.

(00:07:53): Naveen is awesome.

(00:07:55): He's building Monologue, which is

(00:07:58): An internal every product that is run by just him and it's competing with companies

(00:08:01): that have like 30 or 50 employees.

(00:08:03): It's really amazing.

(00:08:04): And it's all because of a he's so smart, but also his workflow is really incredible.

(00:08:09): So he's going to talk about how he's building an iOS app and all the stuff he's

(00:08:13): learned about building an iOS for Monologue.

(00:08:17): Then we've got Yash, also amazing.

(00:08:18): He's a GM of Sparkle.

(00:08:20): And Sparkle is our AI file organizer.

(00:08:22): And so he's going to talk about building a file organizing app.

(00:08:27): And then we've got,

(00:08:29): lastly,

(00:08:30): Brooker Belcourt from Every,

(00:08:31): who is one of the other people on our consulting side.

(00:08:34): And he's basically built like a whole...

(00:08:37): uh coded hedge fund skill that does trading for him um and it's like actually

(00:08:43): really really good so he's gonna i think he's gonna talk about that and talk about

(00:08:46): the intersection of vibe coding and consulting so we've got a an incredible

(00:08:51): incredible agenda um

(00:08:53): I am about, I'm going to do a demo.

(00:08:55): I'm going to start demoing you my workflow before I do that.

(00:08:58): I'm just going to check in with my producers and make sure all of this is all set up.

(00:09:01): This is the first time that we've done this.

(00:09:03): We've got a lot of people on the stream and I'm super excited for this,

(00:09:06): but I'm going to be back in about 30 seconds.

(00:09:09): So just give me, give me.

(00:09:40): All right, I am back and I'm going to start some demos.

(00:09:45): So the thing I want to demo, here's my agenda for today.

(00:09:50): Hopefully you can see the agenda.

(00:09:52): You cannot see it.

(00:09:54): So here's my agenda for today.

(00:09:57): And I wrote it in a markdown file.

(00:10:01): Here's the overall run of show.

(00:10:05): The cool thing is that I'm showing you my agenda in an app that I built.

(00:10:10): This is an app called Proof.

(00:10:13): It's an agent native markdown editor.

(00:10:16): Basically, it's a markdown editor that is built to be used by humans and by agents together.

(00:10:21): I built this for myself.

(00:10:23): I really needed it for all of the stuff that I'm doing,

(00:10:25): like when you build a plan in cloud code or in codex.

(00:10:28): I just don't want to watch that Xcode thing bounce.

(00:10:32): And so I really needed like a really simple markdown editor that had a couple of

(00:10:35): features that made it useful for both humans and agents to collaborate together on

(00:10:39): plans.

(00:10:40): But then I realized there's a lot of other places that we can use it around every.

(00:10:43): So I'm going to I'm going to talk about that,

(00:10:45): but I want to give you a little bit of a tour of proof.

(00:10:48): And then I'm going to and then I have a bunch of stuff that I want to make proof

(00:10:54): better about a bunch of ways I want to make proof better.

(00:10:57): And generally I want to make proof ready for internal testing by our editorial

(00:11:01): team,

(00:11:01): not by our programmers.

(00:11:02): A lot of our lot of the engineering side of every is already using it,

(00:11:05): but by the editorial team and basically I'm going to be on this stream all day,

(00:11:09): so I'm going to be.

(00:11:11): Hopefully the overall goal is to build enough improvements that I can actually

(00:11:15): launch it to the internal team today and you'll be able to watch me do that through

(00:11:18): the day today.

(00:11:20): But let me give you a demo of proof so you kind of get what it is.

(00:11:22): So obviously it's just a basic markdown editor.

(00:11:25): What's really cool is it has this what's written by human and what's written by AI

(00:11:30): here and it's really easy for me,

(00:11:31): for example,

(00:11:32): to go into a part of the document and

(00:11:36): press a command shortcut and you'll see it turns purple.

(00:11:39): And then we have written by AI.

(00:11:41): And what's what's also really interesting is if you open a plan that was written

(00:11:44): all by an AI,

(00:11:45): it will just show all purple and then you will be able to track,

(00:11:48): OK,

(00:11:48): what parts of it have I changed and what parts of it are from me versus from the

(00:11:51): AI,

(00:11:52): which is the first thing that I felt like I needed because I just want to have some

(00:11:55): sense of like who wrote what and what have I looked at?

(00:11:58): What have I what I what have I not looked at?

(00:12:01): but it gets more interesting from there.

(00:12:03): So another thing I can do is I can do the basic stuff like comments,

(00:12:07): like this is cool,

(00:12:09): or I can do track changes.

(00:12:11): So I can say, oops, that didn't work.

(00:12:15): How do we, let me put on,

(00:12:18): Track changes so yeah, so I can do track changes.

(00:12:21): I can suggest changes.

(00:12:24): It's all all the kind of like basic stuff that you might.

(00:12:28): You might expect in a basic good Markdown editor and I just want to stop and say

(00:12:34): that's fucking crazy that I could build this in between meetings and it's I started

(00:12:39): this like.

(00:12:40): Less than two weeks ago and I've only demoed you like half of what's here.

(00:12:45): This is a project that should have taken like engineers like a group of 300 years,

(00:12:50): like six months to do.

(00:12:51): And this is something that I I just vibe coded in my spare time and I have

(00:12:56): absolutely no idea how any of the code works.

(00:12:58): We're living in.

(00:12:59): We're living in crazy times.

(00:13:01): Let me show you some of the more crazy things.

(00:13:04): One thing I can do is I have this review,

(00:13:08): this review menu and I can say run review every style guide.

(00:13:11): So we have a style guide at every that that says here's how we here's how we write.

(00:13:15): Here's what good writing is and it takes a long time for our editors to like go

(00:13:20): through everything that we publish and say like,

(00:13:22): you know,

(00:13:23): correct all of the all of the spelling errors or all the different things that

(00:13:28): don't match our style guide.

(00:13:29): So I can press every style guide.

(00:13:31): And what it does is it launches a proof agent.

(00:13:33): So proof has an is an agent native.

(00:13:34): So it has an agent inside of it that is going to go through the document and make

(00:13:38): changes as we're as we're talking.

(00:13:40): Hopefully this is a live demo, so it may it may mess up.

(00:13:44): But basically that's that's the basic idea is like I have an agent inside of here

(00:13:48): that I can say like at proof.

(00:13:50): What do you think of this?

(00:13:53): And now we have two agents running.

(00:13:55): And so and so look, look, look, look.

(00:13:58): So it's starting to make changes according to the style guide.

(00:14:01): So it said new anecdote test flight test flight in our style guide.

(00:14:04): We would we would capitalize T and F. So I press apply and it applies it.

(00:14:10): And so this is kind of crazy.

(00:14:12): We've got like agents running in our in our markdown editor and it gets sort of

(00:14:19): crazier from here.

(00:14:21): what's what's kind of cool about this so far is it has an agent inside of it.

(00:14:26): But what I really also want is a way to have other agents interact with the

(00:14:31): document I have open.

(00:14:33): So I'm going to open warp here, which has cloud code over here and then codex over here.

(00:14:39): And I'm going to open up a more complex document.

(00:14:42): So this is a this is like a traditional plan file.

(00:14:45): So if you're familiar with if you're doing a lot of odd coding,

(00:14:48): you know these are the kind of plan files that cloud or codex are going to produce.

(00:14:51): You can see it's all purple, which means I have not written any of it, but I can.

(00:14:56): This is by a human.

(00:14:57): I can put my human stuff in there and it's really easy for me to see like.

(00:15:01): what I've done versus what the agent has done.

(00:15:03): And it's also really easy for Claude or any agent to interact with the document I've opened.

(00:15:10): And I have a little demo of this.

(00:15:13): So I put it in a little slash command, but basically,

(00:15:19): I can just have Claude.

(00:15:21): I can just say Claude,

(00:15:21): hey,

(00:15:22): like go look at the document I've opened and make comments or flag anything you

(00:15:25): think I should look at.

(00:15:26): This is all kind of wrapped in a presence demo so you can see.

(00:15:29): But basically I'll say agent presence demo.

(00:15:32): Claude's going to connect to proof and you'll see that there's a.

(00:15:36): Hopefully it'll pop up right here.

(00:15:40): And let's see.

(00:15:44): Let's see if it works.

(00:15:49): It's it's be booping.

(00:15:51): OK,

(00:15:52): so it's starting to make changes and what's really cool is it'll follow the agent

(00:15:57): around to different parts of the document as it's making changes.

(00:16:01): And like that's just I don't know.

(00:16:02): I just feel like this is kind of fucking wild and it's wild that this works and I

(00:16:10): have no idea how this works.

(00:16:11): And it's also just very useful for me.

(00:16:13): So what I'm going to be doing over the next couple of hours is I'm going to be

(00:16:18): making this even better.

(00:16:22): And hopefully, basically, if I feel like it's good,

(00:16:26): Um,

(00:16:26): I'll release it internally at every,

(00:16:28): and then we will release it to every paid subscribers as well.

(00:16:30): Um, but this is what's possible.

(00:16:32): This is what's possible.

(00:16:33): I have, I have not looked at any of this code.

(00:16:37): Um, and, uh, I just, I, I think that too few people understand that this is a thing.

(00:16:43): So, uh, let me, I'm going to start with, uh, my plan.

(00:16:50): Let's look at, let's look at my plan.

(00:16:51): I'm gonna resolve these.

(00:16:54): And I'm just going to say.

(00:16:57): Reject all changes because those are all just sort of demo changes.

(00:17:01): OK, so.

(00:17:05): A couple of things I want to do.

(00:17:07): There's just some general I've been working on the sort of like agent presence part

(00:17:11): of this,

(00:17:12): and so there are some general agent presence things that I want to get done.

(00:17:15): One is I think it'd be really fun when an agent's in your document if it shows the

(00:17:19): logo of the agent so you know if it's a cloud agent or if it's a

(00:17:22): codex agent or whatever.

(00:17:26): Another thing is I really don't like this.

(00:17:30): I don't think it should say active agents.

(00:17:32): I just wanted to like pop up with the agent toast thing if an agent is there,

(00:17:38): but otherwise it should just say nothing just to keep the UI a little bit cleaner.

(00:17:47): Another thing that I really wanna do is I wanted to have like a little bit of a follow glow

(00:17:54): So if you notice like when Claude in Chrome is using Chrome,

(00:17:59): it like has a little orange glow around the screen.

(00:18:02): And what I really want to do is have the one when you're following an agent around

(00:18:07): your document,

(00:18:08): I really want the it to glow a color so you can kind of like see where the agent is

(00:18:14): and it just feels like there's a different thing happening.

(00:18:17): So that's kind of cool.

(00:18:18): Also the scrolling was not really working,

(00:18:19): so I just want to like debug some of the scrolling issues.

(00:18:23): um and then there's a couple things there's a couple other things that i want to do

(00:18:27): these are things i'll try to try to get done through the rest of the day but i want

(00:18:31): to just start here and like see if i can figure out how to show you how i would how

(00:18:37): i would do this normally so i'm going to go into claude let's say um i already have

(00:18:43): a work tree going so let's say

(00:18:48): Uh, I'm going to say, okay, so I'm using monologue here and monologue is this thing over here.

(00:18:53): I just normally just like talk into it.

(00:18:55): This is, uh, built by every as well.

(00:18:58): Um,

(00:18:58): built by,

(00:18:58): by Naveen,

(00:18:59): he'll be presenting later today,

(00:19:00): but basically,

(00:19:01): uh,

(00:19:03): with monologue,

(00:19:03): with monologue,

(00:19:04): I'm just going to talk into my plan and let me just cancel this.

(00:19:06): Cause there's a lot of, uh, oops.

(00:19:08): Okay.

(00:19:09): So, okay.

(00:19:13): Alright, I want to do a bunch of agent presence improvements.

(00:19:16): We already have a work tree going,

(00:19:18): so I want you to do all these in the work tree,

(00:19:20): but first I want to make a plan.

(00:19:22): So here are a couple of the things that I want to make better.

(00:19:26): One is when an external agent joins the

(00:19:32): Joins the session.

(00:19:33): I want to show the if the logo of the agent if we can and do it in a way that like

(00:19:39): makes sense,

(00:19:40): like within the within the little agent status thing.

(00:19:44): So if it's a cloud agent,

(00:19:45): I want to show a little like tiny cloud logo with the chat GPT agent or codex

(00:19:49): agent.

(00:19:49): I want to show that if we don't know what it is, we should just have like a default.

(00:19:52): So that's like that's one thing.

(00:19:56): Another thing is I want.

(00:20:00): I don't think that we need active agents and no agents there on the sidebar unless

(00:20:05): there are active agents.

(00:20:07): If there are activations, we should just show their their UI.

(00:20:11): But otherwise just yeah,

(00:20:13): like just to keep things clean,

(00:20:15): like just have no active agents in there.

(00:20:19): I also want to do a follow glow.

(00:20:24): And the way that I want you to do that is.

(00:20:28): When we're following an agent around the around the document,

(00:20:32): I want the little border of the like document window to have a little bit of like a

(00:20:37): glow to it.

(00:20:38): Maybe in the purple color that we're using for agents so that you know I'm I'm

(00:20:44): attached to an agent and ideally it's sort of like a.

(00:20:48): It's the same sort of glow that Claude uses around Chrome.

(00:20:52): I don't know exactly.

(00:20:52): Maybe it's like a sort of internal drop shadow or something.

(00:20:55): But yeah, something like that.

(00:20:57): There are also a couple of bugs that I want to improve.

(00:21:00): So one is.

(00:21:05): I think following is broken for external agents because I'm observing that when I

(00:21:10): do the agent presence demo,

(00:21:13): it pops the agent finish toast even when it shouldn't.

(00:21:17): So it's like popping the finished thing in between steps as opposed to at the end.

(00:21:22): And I think when it pops the finished thing, it ends the following.

(00:21:25): So we got to figure out what that's about.

(00:21:26): And you can look at the agent presence skill to figure that out.

(00:21:30): Also following doesn't work for the internal agents.

(00:21:32): We want to make sure that that works.

(00:21:36): And what I'm going to do is I'm just going to press this and it's going to

(00:21:42): transcribe it hopefully and I'm going to paste in here and we've got our whole.

(00:21:49): We've got our whole.

(00:21:53): all of me monologuing into this thing.

(00:21:54): And I'm going to use,

(00:21:56): so we have a compound engineering plugin built by Kieran,

(00:22:00): who is GMO of Cora at Every.

(00:22:03): And I'm going to use his planning feature here because I think it's really, really good.

(00:22:08): And this is going to do like a really in-depth plan.

(00:22:12): I'm going to say, make sure you do this in the current presence work tree.

(00:22:19): OK, cool.

(00:22:20): So now that's off.

(00:22:21): We're sending that off.

(00:22:25): And let me see if there's anything else to do while we are.

(00:22:31): Oh, here's a good here's a good one.

(00:22:33): The review skill is like not quite good enough and I want to figure out like.

(00:22:40): While this is going, I want to do some digging into like how the actual.

(00:22:47): review skill works and uh make it better so the review skill for people who haven't

(00:22:52): seen it i can say run review and run the every style guide thing but it's like not

(00:22:56): that good and i need to spend a little bit of time figuring out like how does that

(00:23:01): actually uh how does it work and why is it not working that well so i'm just gonna

(00:23:06): say um

(00:23:11): Hey, can you, I wanna understand the review skill.

(00:23:16): I feel like the review skill inside of the app doesn't really work that well.

(00:23:21): Can you just help me first understand how it works and then interview me to help me understand

(00:23:29): To help to help you understand like what should what would make it better.

(00:23:32): It's also I'm also finding that it's failing a lot and I don't understand it like

(00:23:36): why it would fail.

(00:23:37): So just take a look at it, review it in depth and then.

(00:23:44): And then let's talk about it.

(00:23:46): So I'm going to let that go.

(00:23:47): I'm going to check in on this.

(00:23:49): It looks like.

(00:23:51): doing some presence improvements i'm going to say we've done a bunch of presence

(00:23:55): improvements already so make sure you review those plans too you probably will but

(00:23:59): just take a look um and so that's that's all running one thing that i think i could

(00:24:07): demo for you guys while we're uh while we're waiting is um i also have another app

(00:24:14): that i built for myself called anecdote

(00:24:17): And I have a couple of PRs that I need to merge into anecdote.

(00:24:21): So,

(00:24:21): uh,

(00:24:22): while we're waiting for all this to happen,

(00:24:24): I'm just gonna like open up anecdote and demo it for you.

(00:24:26): Cause I think it's a really good example of an agent native app.

(00:24:30): Um, so I'm gonna share my screen.

(00:24:34): Hopefully this will work.

(00:24:46): Is it working?

(00:24:48): Not yet.

(00:24:58): Live demos, folks, live demos.

(00:25:09): Alright,

(00:25:10): for some reason I cannot get my iPhone to screen share properly,

(00:25:16): so I'm going to stop mirroring,

(00:25:18): but maybe I can open it up in a simulator and show you guys in the simulator.

(00:25:25): So let's just say.

(00:25:29): Alright.

(00:25:32): Can you run anecdote in the simulator?

(00:25:35): Alright, let's see what happens, but I'll I'll describe it so that you guys get a sense for it.

(00:25:40): So.

(00:25:42): We've been talking a lot at every about agent native apps,

(00:25:45): and if you're curious about how to build agent native apps,

(00:25:48): you should go here.

(00:25:48): Every dot to slash guide slash agent native and you can think about agent native

(00:25:53): apps as the core of the software you're building.

(00:25:57): is just an agent.

(00:25:58): It's like cloud code at the at the center of of all your software or working.

(00:26:04): And and that's very different from and basically every time you press any kind of

(00:26:09): button or anything in the app,

(00:26:11): it's actually sending a prompt to an agent,

(00:26:13): and that's very different from traditional software where the software is pre built

(00:26:18): with rules that a programmer writes in an agent native app.

(00:26:24): Each feature is essentially a skill that the agent can do that has prompts and

(00:26:29): tools and all that kind of stuff,

(00:26:30): and it runs in a loop until it's done.

(00:26:32): And there are new principles for how to build agent native apps.

(00:26:35): So there's principles like parity.

(00:26:37): So like whatever the agent can do through the UI that the whatever the user can do

(00:26:41): through the UI,

(00:26:42): the agent should be able to do.

(00:26:43): And you can see that in the way that proof works.

(00:26:45): Let me just see.

(00:26:49): We're installing or building and installing, so that's good.

(00:26:53): So parity, granularity.

(00:26:55): So granularity means you want the tools that the agent has access to to be smaller

(00:27:03): than the features.

(00:27:05): So the agent can combine tools in new ways that you didn't necessarily predict to

(00:27:10): do things that you might want.

(00:27:13): So like an example of granularity, which is really interesting,

(00:27:17): is one thing that this app does not have yet is a clear all comments function.

(00:27:25): And that's something that like ordinarily,

(00:27:27): in order to get any of that to work,

(00:27:31): I would need to like code that.

(00:27:33): I would need to go ask code or codex or cloud code like,

(00:27:35): okay,

(00:27:36): code a function that like allows you to clear all,

(00:27:39): reject all comments,

(00:27:40): just like we have a reject all changes function.

(00:27:42): But because we have parity and granularity, what I,

(00:27:48): I can accidentally minimize my document.

(00:27:52): We're living in the future, folks.

(00:27:54): What I can do is I can say, at proof, can you clear all of the comments here?

(00:28:00): And because the agent can do anything the user can do,

(00:28:04): even if I haven't built this feature explicitly,

(00:28:05): it just knows how to do it.

(00:28:07): And you'll watch the comments disappear in a second.

(00:28:12): And look, there are no comments anymore.

(00:28:14): And so that's like the really interesting,

(00:28:17): that's one of the really interesting properties of building truly agent native apps

(00:28:23): like this is if the agent has the ability to use any part of the tool,

(00:28:29): the users can use it for things that you didn't explicitly build,

(00:28:31): but are really powerful and valuable.

(00:28:33): Um, okay.

(00:28:34): We have one minute before our next guest.

(00:28:37): So I just want to show you, um, I want to show you anecdote real quick.

(00:28:41): This is another agent native app.

(00:28:43): It's a health app.

(00:28:45): Um,

(00:28:46): Basically, the core of the app is a chat.

(00:28:49): It's like an agent, hey, can you look at how I slept last night and compare it to my glucose?

(00:28:58): It's an agent that connects to your Apple Health data.

(00:29:00): It's very much like what Anthropocon, hopefully slightly better or better.

(00:29:07): It came out earlier and basically I used this as a way it helped me figure out that

(00:29:12): I was not sleeping well and I was not sleeping well because my glucose was

(00:29:16): crashing.

(00:29:18): And anyway,

(00:29:19): this isn't a simulator,

(00:29:20): but basically has a chat has access to all your Apple health data.

(00:29:23): It also has a feed so you can write journal entries.

(00:29:25): Hey.

(00:29:27): I slept like shit last night.

(00:29:30): You can write journal entries that that anecdote can read an anecdote itself.

(00:29:39): can write journal entries too.

(00:29:41): So you can have this collaborative thing going on where you are collaborating with

(00:29:48): an agent,

(00:29:49): you have access to the same data and the same feed of journal entries and

(00:29:52): documents.

(00:29:55): you know, genetic documents and stuff like that.

(00:29:57): And it can help you kind of figure out your health.

(00:29:59): It also does a morning brief where it will like learn about your health goals and

(00:30:05): then help you figure out,

(00:30:08): hey,

(00:30:08): like what do you want to track and what do you want to know about yourself in the

(00:30:12): morning?

(00:30:13): And again, this is all agent native.

(00:30:15): So like all this stuff is just prompts to the agent.

(00:30:18): So it's quite flexible.

(00:30:20): And I see my good friend Ben from Ben's Bites is here.

(00:30:23): Ben, welcome.

(00:30:25): Hello.

(00:30:25): I was going to say, you should try having three kids and worry about your sleep.

(00:30:30): I, uh, yes.

(00:30:32): I hero, absolutely hero.

(00:30:35): Do you have three?

(00:30:36): I think you only had two.

(00:30:38): Three under three, yeah, two and a half year twins.

(00:30:40): And then the last one just came along six weeks ago.

(00:30:44): Congratulations, how you doing?

(00:30:45): Thanks, I'm good.

(00:30:46): Yeah, she sleeps a lot, so.

(00:30:48): She sleeps a lot, what's her name?

(00:30:49): Poppy.

(00:30:50): Poppy, amazing.

(00:30:51): So you're gonna show us,

(00:30:53): hopefully,

(00:30:53): how you vibe code without knowing how to code and with three kids under three.

(00:30:58): Yeah, I've got a bunch of stuff that I've built recently, so I'm just going to dip into those.

(00:31:07): Before we do that, let's just introduce yourself.

(00:31:10): We've been friends for a long time.

(00:31:11): You run Ben's Bytes, an incredible newsletter, one of the AI newsletters I read all the time.

(00:31:16): You also work at Droid.

(00:31:19): Just give us a little bit of background on yourself, and then we'll jump into your demo.