BEN TOSSELL (Ben's Bites / Factory) - Non-technical builder, session sharing

=== BEN TOSSELL (Ben's Bites / Factory) - Non-technical builder, session sharing ===

(00:31:23): Yeah, well, quick background is I can't code, still can't.

(00:31:28): I've tried and just couldn't never get it to work for me.

(00:31:31): So I built a company called MakerPad where we taught people no code back in the day.

(00:31:35): So this is 2019, drag and drop type stuff.

(00:31:38): I sold that to Zapier.

(00:31:40): It was there for a few years.

(00:31:43): Started Ben's Bytes before Chatwt came around and then just sort of rode the AI

(00:31:47): wave and then I've been investing and building ever since.

(00:31:50): And I'm an investor in a factory and have joined them recently to help on the DevRel side.

(00:31:56): And yeah, that's kind of what I'm doing.

(00:31:58): So just build stuff all day and all night as much as possible.

(00:32:02): Fuck yeah.

(00:32:03): Okay, what do you want to show us?

(00:32:06): Lots of stuff.

(00:32:08): So I will share my screen, I suppose.

(00:32:12): here let me know if you can see it not yet uh no but it's coming i see it now cool so this is

(00:32:38): A couple of things.

(00:32:38): This is one session reviewer that I have built for myself.

(00:32:44): So you can basically see my whole entire session with my agent.

(00:32:47): So basically export the session from the terminal where I'm talking to Droid all

(00:32:51): day and it presents it in this UI.

(00:32:54): So you can sort of see I don't want to see any of the tools.

(00:32:57): I just want to see the user message.

(00:32:59): Then you can basically scroll down and just see the actual like what I actually say

(00:33:04): to the agent and what did it actually do.

(00:33:07): this one i did 10 minutes ago to say quick i've got to present this to dan's

(00:33:11): audience in 10 minutes one shot put the pop-up on exit to effectively show i gave

(00:33:18): it this code my snippet from my newsletter so that's it's just embedding there

(00:33:22): because it renders it on this on this site um so then if i go to the site and go to

(00:33:29): leave it it should come up with a pop-up

(00:33:34): So built that,

(00:33:35): yeah,

(00:33:35): 10 minutes ago,

(00:33:37): just as like a,

(00:33:38): let me do a quick thing to see if I can just quickly one shot something.

(00:33:43): So that's one thing, but like even building this session builder

(00:33:47): It's kind of a format that I follow a lot, and I've got a few here.

(00:33:51): This is where I stole it from, basically.

(00:33:53): So there's a developer called Mario who has built an agent called Pi,

(00:33:59): and he's got a share command on his agent that basically renders a HTML page like

(00:34:04): this,

(00:34:05): exactly like this,

(00:34:06): actually,

(00:34:06): but colors are slightly different.

(00:34:08): So I just said, find the share command code from his coding agent,

(00:34:16): So then the agent went and did a bunch of that stuff.

(00:34:19): Let me try and get the raw file, blah, blah, blah, blah, blah.

(00:34:22): All the different tools it's doing.

(00:34:23): Then it found the share command and sort of what it does.

(00:34:27): So then I say, make a plan to copy the implementation and replace what we currently have.

(00:34:34): Because I think I just had, basically a GitHub gist is created.

(00:34:38): So this is showing you the spec that Droid comes up with.

(00:34:42): And then I basically just said,

(00:34:47): I was trying to figure out how you submit it and all that kind of stuff.

(00:34:52): Then this was me building this thing,

(00:34:54): just saying yes,

(00:34:55): approving the plan,

(00:34:56): going through the plan a few times and then wanted a web view.

(00:35:00): So I just hook it up to my website and all that kind of stuff.

(00:35:03): And then it did it.

(00:35:05): So every session, like I just shared that one, I just do a share session command.

(00:35:12): And then it renders, it basically exports the whole session to a GitHub gist that is a secret.

(00:35:18): And then it populates my website with this view so I can show you and everyone else

(00:35:25): watching how you can do these sessions.

(00:35:26): And I think for me,

(00:35:28): how I'm learning to build stuff is often just being in a session and talking to an

(00:35:34): agent.

(00:35:35): I'm just doing stuff I'm not supposed to be allowed to do.

(00:35:38): because you can.

(00:35:39): And then you just start picking up those things as you go past.

(00:35:43): You start seeing like Glob, what the fuck is Glob?

(00:35:45): Like, oh, then you realize it's like, it's a search tool.

(00:35:48): It's searching for something globally across your project for certain things.

(00:35:54): You just start recognizing the patterns and feeling more confident with what you're doing.

(00:35:58): And I actually just published

(00:36:00): this which was like all the terminology and everything you kind of need to know and

(00:36:06): what i feel like i learned as i've been doing this of how to get comfortable with

(00:36:10): working with agents because something i have built is this sort of cookbook site

(00:36:16): for myself and i want to the idea is to like share all of my sessions that i'm

(00:36:21): doing and then do a walkthrough of my session

(00:36:24): And obviously in building this, I can share the sessions of building this as well.

(00:36:28): So if we go to this, this is where I was.

(00:36:32): I saw this tweet, which many people might have.

(00:36:35): It's got this fun button that has these different states and it kind of changes and

(00:36:40): has an animation.

(00:36:41): And I had had these things called nutshells, which is an open source project.

(00:36:48): And it basically is an expandable that then explain in line.

(00:36:54): So my idea was I can go through a session of how I'm reverse engineering something

(00:36:58): or building something.

(00:36:59): And anytime there's a technical term, it will come up with a nutshell.

(00:37:05): You can be like, okay, what the hell is that?

(00:37:07): And then these,

(00:37:08): I try to make sure I'm manually writing them or curating them at least to make them

(00:37:13): actually easy to understand.

(00:37:15): So that's what I've been doing.

(00:37:18): And then, yeah, this is what I wrote to my agent.

(00:37:21): Use a bird tool to get this tweet.

(00:37:23): download the video,

(00:37:24): analyze it to reverse engineer it and create a similar component that we could use.

(00:37:27): And then again, you can go to the full session and review the full session.

(00:37:34): And then you just talk about what actually happens then in this session here or in

(00:37:39): this assistant part of the session and then go through and open it.

(00:37:47): It's just all so simple stuff that I think people get scared of

(00:37:51): but they don't need to be.

(00:37:51): And I was like, how you replace our nutshell with this?

(00:37:54): And then do it like it comes up with a plan.

(00:37:56): I think then it comes up with two options or three options because models love

(00:38:01): coming up with a hybrid option for some reason.

(00:38:05): So then I just said,

(00:38:05): yeah,

(00:38:06): do a start the server,

(00:38:08): start the buttons with more 3D,

(00:38:10): better,

(00:38:10): it's too big,

(00:38:11): make it a bit smaller.

(00:38:12): And that's actually what you see now on the site is what I built here.

(00:38:16): So I think that's pretty cool.

(00:38:18): I love just being able to share my session.

(00:38:20): I think that's how I've learned is poking around other people's code,

(00:38:24): reverse engineering it,

(00:38:25): going through a session,

(00:38:27): just asking questions and doing dumb stuff.

(00:38:30): So that's a couple of things I've built.

(00:38:33): Wait, I want to pause you there.

(00:38:34): This is super cool.

(00:38:36): And so basically like if I'm interested in what you're doing,

(00:38:39): I can go look at your almost basically your raw five coding sessions and then your

(00:38:46): output of,

(00:38:46): you know,

(00:38:48): almost the story version of this is how I did something.

(00:38:51): I think that's so helpful for technical people.

(00:38:55): Can you talk for a second about like how far are you able to take this?

(00:39:00): So

(00:39:01): You're not technical.

(00:39:02): Obviously you're building this stuff,

(00:39:03): but I think someone looking at it could be like,

(00:39:05): yeah,

(00:39:06): like,

(00:39:06): but it's all static sites or whatever.

(00:39:08): Like, what are you, what's, what are the limits of what you're able to do?

(00:39:11): Like, are you, are you pushing you work at, at, at factory?

(00:39:14): Are you, are you pushing code for factory or like possible?

(00:39:18): Yeah.

(00:39:18): Yeah, I mean, I've pushed code to the actual code base, which was like a big moment for me.

(00:39:24): And this is one I wanted to show,

(00:39:27): which is we had my brother in Bubble,

(00:39:30): which is a no code tool from back in the day,

(00:39:32): which I found very,

(00:39:34): very difficult to ever use.

(00:39:36): So he bought this sort of newsletter booking system that we used for our newsletter.

(00:39:41): So this is like our ad platform.

(00:39:43): Instead of like Passion Fruit or anything like that,

(00:39:45): I didn't want to talk to any sponsors at the time.

(00:39:47): So we just had this.

(00:39:48): People would just book main ad, a tool placement, whatever.

(00:39:51): So this...

(00:39:54): which works as you would expect it to, fill in a form, pick some dates, all that kind of stuff.

(00:40:00): But it stopped working recently and we can't figure it out.

(00:40:03): So I thought, well, I can just rebuild this.

(00:40:07): And this is a very, very long session.

(00:40:09): I think there's 766 messages here.

(00:40:13): But what I said was I need to clone a piece of software we built on Bubble several years ago.

(00:40:17): Bubble's an old platform.

(00:40:19): Rebuilt in the simplest possible way.

(00:40:21): Least lines of code, dependencies, et cetera, but look good, work the same way.

(00:40:25): If anyone from Bubble is watching, we apologize.

(00:40:27): I know, I apologize.

(00:40:29): Yeah, I didn't think about this when I was exporting more sessions.

(00:40:31): Yeah.

(00:40:33): Use the dev browser skill to go through the website,

(00:40:35): click every link and button,

(00:40:36): explore every interaction,

(00:40:37): come up with a master plan on how you rebuild it.

(00:40:39): I forgot to give it to the link.

(00:40:40): So I said, good point, give it a link.

(00:40:43): And then I also have this in my agent's MD,

(00:40:46): I've got this line from Cool Runnings,

(00:40:48): just so I know that the agent has read my agent's MD before it starts.

(00:40:55): So yeah, good point.

(00:40:57): And then it comes up with a plan somewhere here.

(00:41:02): A big plan.

(00:41:02): And I basically have been building it now.

(00:41:05): So it's currently here.

(00:41:07): This is the staging site.

(00:41:09): So we've got a calendar.

(00:41:10): It won't work now, probably.

(00:41:13): So it's got all the sessions here and stuff.

(00:41:16): You can like hold it, remove it.

(00:41:18): You can like add available hold.

(00:41:26): So add things there, there's orders.

(00:41:28): So you can like create a new order.

(00:41:30): Okay, someone wants these dates, so you can sort of mark them for them.

(00:41:35): You can send them an IO, we can sign it.

(00:41:37): Like we've built a lot, or I've built a lot with all these connective pieces.

(00:41:43): Obviously there's super bases connected behind the hood.

(00:41:46): So you can have things changing in real time from the database.

(00:41:51): All of this is just, yeah, it's basically all vibe coded from that session.

(00:41:57): Even got an audit log to just see like everything that's changed and happened at any point.

(00:42:01): So we know that anything, you can see my agent also going through and testing stuff.

(00:42:07): So, I mean, it goes to a fully fledged, it just goes all the way.

(00:42:12): It's the same principles.

(00:42:14): It's just, you've kind of got to have a bit of a pain threshold

(00:42:20): And I think in that post that I mentioned,

(00:42:21): I posted on Twitter,

(00:42:23): it talks about like,

(00:42:24): I mean,

(00:42:25): everyone gets there where you're like,

(00:42:26): fix it,

(00:42:27): fix it,

(00:42:27): fix it.

(00:42:27): And you're sort of in that loop and you're like, why the fuck is it?

(00:42:29): This agent is so smart that I can't fix this thing that's a bug that seems like an

(00:42:33): obvious thing.

(00:42:34): And you've given it documentation, you've given it whatever else.

(00:42:36): And it's just, I actually just quite like figuring out

(00:42:41): It's always a me problem, and I know that because I can't code.

(00:42:44): So it's always my fault,

(00:42:46): but just knowing that and saying,

(00:42:47): well,

(00:42:47): yes,

(00:42:48): obviously you don't know how to fix this because you can't code.

(00:42:50): So your job is to figure out how you can guide the agent in the best way to then

(00:42:54): get it to figure out what it needs to do next,

(00:42:56): no matter how complex it is.

(00:42:59): So I just play around with that, and I switch up the models.

(00:43:02): I use a lot of Opus 4.5, but more recently, especially on this platform,

(00:43:07): project i've been using gpt 5.2 codex and obviously with factory you can just

(00:43:11): switch like in the middle of a session um so it's been really helpful to like

(00:43:15): review the code that opus has written or make a plan with codex and then switch

(00:43:19): over to opus to implement it and all that kind of stuff but yeah i mean i've hit a

(00:43:23): fair amount of issues with building this platform for example but um i mean in all

(00:43:30): of these sessions there's going to be a bunch of stuff

(00:43:32): me just like swearing at it or just like getting really frustrated but that's just

(00:43:36): part of part of how this whole thing goes i think um let me just stop you there

(00:43:42): because i think this is really interesting and very helpful for people who are

(00:43:46): maybe non-technical like when you get stuck the first thing that you said that

(00:43:53): really

(00:43:54): struck me is you're like this is my fault it's not the agent's fault it's my fault

(00:43:59): i can't code yeah and i think that's really really interesting so once you've taken

(00:44:03): that responsibility and you're stuck um what's there what's your first step that

(00:44:09): you take to help yourself get unstuck well i do a few different things so often to

(00:44:15): validate if something is working like an implementation that the agent's working on

(00:44:19): is going to work i don't know anything about tests but i say make sure you write

(00:44:23): tests

(00:44:24): And also I used, I used to use dev browser.

(00:44:28): I now use the agent browser from Vercel.

(00:44:31): So I say, use the agent browser, go through everything and like log exactly what's happening.

(00:44:36): And then you come up with,

(00:44:38): like,

(00:44:38): I also can't be bothered to keep going onto the site where it says it's done

(00:44:41): something.

(00:44:42): Okay.

(00:44:42): It doesn't work.

(00:44:43): I'm not going to copy paste console logs and Vercel logs and all the rest of it.

(00:44:47): Like the agent can figure that out and do it itself.

(00:44:49): Yeah.

(00:44:51): So I kind of get it to do that kind of stuff, to see that.

(00:44:54): And what I did,

(00:44:55): I was stuck on a database issue,

(00:44:58): especially writing to the wrong database for a start.

(00:45:00): So that's not a good sign.

(00:45:04): So what I did more recently was kind of taking some

(00:45:10): take some concepts from that route,

(00:45:11): the Ralph sort of phenomenon that's happening,

(00:45:14): which is just everything's running in a loop and you've got to have a file that has

(00:45:17): like,

(00:45:17): this is your plan.

(00:45:18): And then you've got a file that is like,

(00:45:19): this is your progress and you're taking stuff off as you go.

(00:45:22): So every agent knows where you are at every stage.

(00:45:26): So I just said,

(00:45:26): right,

(00:45:27): let's go through and list all the possible issues of what could be happening with

(00:45:31): this like bug.

(00:45:33): The problem with, you basically couldn't create orders, inquiries or customers.

(00:45:37): Nothing was getting written to Superbase.

(00:45:40): So I knew that there was the issue,

(00:45:42): like nothing's going from this site to Superbase,

(00:45:44): but it was all working locally.

(00:45:47): So say, come up with all the ideas that you have of all of the bugs.

(00:45:52): Now let's go through them one by one.

(00:45:53): Or you can say to the agent, which do you think is the most likely culprit?

(00:45:58): I also, after several iterations of trying to fix the bug,

(00:46:02): I realized I was in that loop and thinking,

(00:46:04): OK,

(00:46:04): I know now it's not going to be any of these sort of things that seemingly are

(00:46:09): going to fix it.

(00:46:10): I think it's a super base issue somewhere.

(00:46:12): There's no connection with it somewhere.

(00:46:14): And I was like, well, we're not even using super base functions.

(00:46:18): So.

(00:46:19): maybe use that and i just sort of guided it from doing that like planning of think

(00:46:24): about all the bugs this could be run through some of them validate with tests and

(00:46:29): the agent browser and then if it's not working just keep going like keep going

(00:46:33): through that loop um

(00:46:35): And I just pointed it to Superbase Functions.

(00:46:38): Sure enough, that was the issue.

(00:46:39): We didn't have it set up properly.

(00:46:41): It went through, validated it.

(00:46:43): And then it's basically like the agent is then using your site as a first class

(00:46:47): user,

(00:46:48): doing everything you would be doing to test that it goes wrong.

(00:46:51): And then it just like is in a loop with itself.

(00:46:54): So then it can just continually like fix stuff,

(00:46:58): update its own plan and then continue from there.

(00:47:01): So I just try and think about how I can

(00:47:05): what is a what my markdown file what instruction can i give it to guide it to do

(00:47:09): something over and over and over again to like bang its own head against the wall

(00:47:12): to figure it out um and like are there things i've picked up from doing many

(00:47:18): projects or just this one project that feel like i feel like it's not that we've

(00:47:22): gone down that road before or like to me it feels like it must be to do with it's

(00:47:28): not that

(00:47:29): Like it's not that part of the code,

(00:47:31): it's gotta be this part because there's nothing in Superbase,

(00:47:33): for example.

(00:47:33): It's like obvious to me.

(00:47:38): So that's kind of how,

(00:47:40): and I've only gotten that intuition,

(00:47:42): I suppose,

(00:47:42): from just building a load of stuff.

(00:47:46): And the point is,

(00:47:47): I think if you're really non-technical and the point of that article is like,

(00:47:52): you're gonna hit issues all the time.

(00:47:54): Like professional developers hit issues all the time.

(00:47:57): Like you can't expect not to.

(00:47:59): it's just learning how to then navigate those issues and like if they come up again

(00:48:03): ask the agent why did that come up again do we need to add something to our agents

(00:48:07): md to like improve that so we don't come up with that again or is it something that

(00:48:11): i've given the instructions that are conflicting or um things like that so i think

(00:48:15): there's i i know you guys have your like compound engineering plug-in and people

(00:48:20): have really really extensive setups on slash commands and hooks and skills and mcps

(00:48:26): and all the rest of it and i think

(00:48:28): when you see people like Peter Steinberg and even Boris from Claude Code talking

(00:48:32): about like the simplest setups that they've got,

(00:48:35): and yes,

(00:48:35): they are very good programmers,

(00:48:37): but it's mostly up to the model.

(00:48:40): Like the model does most of the work.

(00:48:42): You're there to guide it.

(00:48:43): And if you really think about that, as you're going through your session,

(00:48:49): you start thinking about like,

(00:48:51): I'm being stupid here by telling it just to fix it,

(00:48:53): fix it,

(00:48:53): fix it.

(00:48:54): It's getting through its context.

(00:48:56): Like I'm not really helping it.

(00:48:58): I'm just thinking it should be able to have fixed this by now.

(00:49:01): I've given it no guidance or anything like, okay, maybe that's a bit unfair.

(00:49:04): Like maybe,

(00:49:05): okay,

(00:49:05): if I point it over here and then it gets it straight away and you're like,

(00:49:08): okay,

(00:49:09): that's a skill issue for me.

(00:49:10): Like that's fine.

(00:49:12): I know that, but now I'll know in the future,

(00:49:16): when that happened to me and that was painful for 24 hours,

(00:49:18): then hopefully we won't repeat that mistake again,

(00:49:20): but we all fall into the trap.

(00:49:22): Yeah.

(00:49:23): There's a, you said a lot of things I think are really important that I want to emphasize.

(00:49:27): One thing that you said that is totally true is I feel like people discount vibe

(00:49:34): coding because they're like,

(00:49:35): well,

(00:49:35): if something goes wrong,

(00:49:36): how are you going to know how to fix it?

(00:49:37): And it's like,

(00:49:40): professional developers all the time run into problems where they're like,

(00:49:43): I actually don't know what's wrong and I don't know how to fix this.

(00:49:46): And the mindset is the same.

(00:49:48): It's just like, I'm going to figure out why it's not working and I'm going to fix it.

(00:49:53): And if you're a professional developer, you do that by understanding all the code.

(00:49:57): If you're vibe coding,

(00:49:58): it's a process of you understanding what's the difference between what you think

(00:50:02): should happen and what is happening and what the agent thinks should happen and how

(00:50:06): do you close that gap?

(00:50:08): So it sounds like a lot of what you're doing

(00:50:10): when you get into one of those like frustration loops,

(00:50:12): which I find too,

(00:50:13): like I just demo this app proof.

(00:50:15): And two days ago,

(00:50:17): it was like,

(00:50:17): I was just like,

(00:50:18): make it work,

(00:50:18): make it work,

(00:50:19): make it work,

(00:50:19): make it work for like hours.

(00:50:21): And then eventually I was like, okay, I just need to spend the time to sit down

(00:50:25): and explain in detail all the things that are not working,

(00:50:29): and then go back and forth with Claude and Codex until we'd created a shared

(00:50:35): understanding of like what good looks like.

(00:50:37): And then I just said, Codex, go do it.

(00:50:39): And it did it, and it just one-shotted it in like an hour.

(00:50:41): Yeah.

(00:50:42): I think that is the, like, vibe coding gets a bad rap, much like no coders did.

(00:50:49): I think it's like, it's like it's a negative connotation of like the real version.

(00:50:53): You're caught, you're not there.

(00:50:54): So you're like the fake version of this and you're like pretending and you're just

(00:50:57): sort of like flying by the seat of your pants or whatever.

(00:51:02): But I think,

(00:51:04): I mean,

(00:51:04): I am vibe coding because I can't write code,

(00:51:07): but I wouldn't class myself as like,

(00:51:09): I'm one shotting Pomodoro timers on lovable.

(00:51:11): And then just like, that's kind of where I'm at.

(00:51:14): I'm looking at what the agents do.

(00:51:16): I'm thinking about the,

(00:51:17): How do I guide the context?

(00:51:19): I think really it is just engineering,

(00:51:22): not to insult any other engineers,

(00:51:23): but I think that's just like engineering on the very one end of the spectrum where

(00:51:30): what you've got to understand is how you give instructions to an agent,

(00:51:33): because yes,

(00:51:34): the agent will be able to figure it out most likely on what you're building and

(00:51:37): probably not building something that it can't figure out.

(00:51:40): But it's just guiding it.

(00:51:42): Like you said,

(00:51:43): it's like you've got to put yourself in the session with it and try and feel like

(00:51:48): you can be there as a resource to the agent rather than the agent's just going to

(00:51:53): go off and do all of this stuff.

(00:51:54): And maybe it will one day and hopefully soon and all the rest of it.

(00:51:57): But I actually find it really interesting to learn about...

(00:52:02): okay,

(00:52:02): there's all these errors happening and like bug fixes and as they're frustrating at

(00:52:05): the time,

(00:52:06): as soon as you,

(00:52:06): when you like figure it out,

(00:52:08): you just feel like,

(00:52:09): oh my God,

(00:52:09): I actually did that.

(00:52:10): Like, how did I do that?

(00:52:12): And like all these things that I'm, I reverse engineer everything.

(00:52:15): I just, there was one,

(00:52:17): I can't remember where it is now.

(00:52:18): There's like a video that Danny Postma put up of the Ralph Loop stuff.

(00:52:23): He built himself a UI for it and just did a video.

(00:52:28): So I looked at that and I think Matt Pocock had an actual system in his repo of

(00:52:33): like how he uses the Ralph Loop in his code.

(00:52:35): So I was like, pull these two things together, reverse engineer it and like build this system.

(00:52:41): like Ralph loop thing that can be triggered by creating a GitHub issue.

(00:52:46): Then it runs through the loop and then it like creates a PR and that's how you can

(00:52:51): get work done like remotely.

(00:52:53): And I did a video on that on Twitter and there's a repo that people can clone and

(00:52:58): stuff for that.

(00:52:58): So like I just do so much of exploring because I just find it interesting.

(00:53:03): And I feel like there's nothing I can't have a go at

(00:53:07): And it just will be a question of how long I'm willing to sit here and figure out

(00:53:13): how I need to guide the agent.

(00:53:14): Like, do I really want to guide the agent this much?

(00:53:16): Or like, do I want to keep going through this for days on end?

(00:53:19): But like things like the ad platform is,

(00:53:21): that's a core piece of software that we like to use for ads.

(00:53:24): And the alternative is sponsors emailing us all the time.

(00:53:28): So I'm going to get it done.

(00:53:30): So that's kind of,

(00:53:32): there's definitely incentives to it.

(00:53:33): And a lot of it is just like play, like tinkering.

(00:53:35): And I've got so many projects that I just don't do anything or they're just local

(00:53:39): or they're just in GitHub that I never finished.

(00:53:42): But that's part of it.

(00:53:44): That's not the point is not to just do one project ever.

(00:53:49): one of the things that i would really love your perspective on is i feel like

(00:53:53): you've been vibe coding for way longer than anybody else in the world because you

(00:53:56): were doing no code stuff yeah 10 years ago and so you've just been in this world

(00:54:03): for so for for a long time and watch it change can you just i'm just curious about

(00:54:08): you were reflecting on

(00:54:10): then and now and like how things have changed and how your perspective on all this

(00:54:14): has changed you know when you had the original sort of no code dream and got that

(00:54:19): bug to now yeah i think like the no code dream was oh my god i can actually build

(00:54:24): something like this like i can use webflow and air table and have zapier in the

(00:54:28): middle to like build a marketplace or build a job board um that was the extent of

(00:54:33): my dreams then um

(00:54:35): But I was like, oh, actually, anybody in the world can do this.

(00:54:40): This is actually so much more approachable and attainable for anybody than I've got

(00:54:44): to sit down and learn script tags and hello world and go through the 100 days of

(00:54:49): programming or whatever courses and stuff you should take to be a coder.

(00:54:55): and my argument was always that yes it's not going to be software that's like going

(00:55:00): to change the world or going to run a hundred thousand users concurrently but i

(00:55:05): spent so much less time on it i spent like an hour on it or a few hours or a

(00:55:08): weekend like if it doesn't work and no one ever uses it i can throw it away and i'm

(00:55:13): not emotionally invested in the fact that i spent 12 months coding to try and build

(00:55:17): this crap version of something that no one wants

(00:55:20): I think that's true today still.

(00:55:21): That's still the thing.

(00:55:22): You can just build stuff, throw it away.

(00:55:25): It doesn't matter.

(00:55:25): It took an hour of your time, two hours of your time.

(00:55:28): It's not like you've got to be in this to build.

(00:55:30): And I think more and more people are doing that now.

(00:55:32): You see projects flying around Twitter all day, every day.

(00:55:35): I still think that I think I mistakenly thought anyone could build and

(00:55:41): democratization of software is here with no code.

(00:55:44): But it was actually only the people who still had a certain threshold of pain to go through

(00:55:50): to build something like we still have to know to want to build something or have an

(00:55:53): incentive to build something and that is still true today because you still if you

(00:55:57): one shot something on lovable that's probably not going to go into production

(00:56:02): you're going to end up with five prompts 15 prompts 300 prompts before you can get

(00:56:06): it there are you willing to get like to get through that session and do that i

(00:56:12): don't know that's still out there so i don't know about the those kind of tools

(00:56:16): like i think they're great but i think there's

(00:56:20): for what purpose may they exist?

(00:56:23): I don't know.

(00:56:23): I think people would prefer...

(00:56:24): Lots of people would prefer this kind of...

(00:56:27): I felt like I built it.

(00:56:29): And even... Even no-code...

(00:56:34): what I was doing was programming.

(00:56:36): It was just a different abstraction.

(00:56:37): Like Webflow was the front end,

(00:56:38): Zapier was the roots,

(00:56:41): the API layer in the middle,

(00:56:43): and Airtable was the back end.

(00:56:45): So I kind of started picking up the fact that these three pieces and other pieces

(00:56:50): sort of need to talk together,

(00:56:52): need to like be connected.

(00:56:53): When something breaks, it all breaks and like data needs to flow in a certain way.

(00:56:57): So I think that probably helps with my understanding of

(00:57:03): generally how like projects are how you see a project and like what it is and like

(00:57:07): where what's happening behind the behind the scenes or underneath the hood um so

(00:57:12): that kind of helps but i think it's very easy to just be poking around and

(00:57:16): tinkering these days so

(00:57:18): There's lots of similarities, but I think this is what I actually meant back then.

(00:57:22): So when I was thinking of anyone could build anything,

(00:57:25): I meant AI actually,

(00:57:26): but I just didn't know anything about AI or was really using it.

(00:57:28): So there's a sort of right time,

(00:57:31): no,

(00:57:31): right message,

(00:57:33): wrong timing,

(00:57:35): which is the same as being wrong.

(00:57:36): But I think- Well, you stuck around long enough for it to be right.

(00:57:38): I know, luckily enough.

(00:57:40): Yeah, it wasn't too long away that I can kind of look right now.

(00:57:44): So that's fine.

(00:57:45): Awesome.

(00:57:47): Ben, thank you so much for joining.

(00:57:49): If people are looking for you or for Factory, where can they go find you to find out more?

(00:57:57): So at Droid on Twitter for Factory.

(00:57:59): And I recommend that agent over any others that you'll hear about today.

(00:58:03): Of course, I think it's just the best one.

(00:58:05): I don't have to worry about maintaining it.

(00:58:07): It just handles everything for me.

(00:58:09): Actually, give us 30 seconds on that.

(00:58:11): Well, it's multimodal.

(00:58:13): So like I said, you can change multimodal, sorry.

(00:58:15): So you can change between Opus and Gemini and GBT models all in the same session.

(00:58:21): The compaction that we have,

(00:58:23): so when your agent runs out of context,

(00:58:24): which it will,

(00:58:25): our compaction is phenomenal and it basically won't lose anything.

(00:58:29): You don't have to keep starting new sessions, hence my 766 message session.

(00:58:35): And yeah, I mean, it's the one that I feel like

(00:58:40): does more for me than others where others need more handholding.

(00:58:43): And I've got enough handholding to do myself, not knowing how to code.

(00:58:46): So actually I think it's obviously I work there,

(00:58:50): but yeah,

(00:58:51): it's been the best one that I can't stick to any others,

(00:58:54): but I'm on Twitter at Ben Tossel or bensbytes.com is our newsletter.

(00:59:02): Amazing.

(00:59:03): Thank you, Ben.

(00:59:04): Thank you.

(00:59:06): all right we've got our next guest coming up in a in a short second i just want to

(00:59:11): take a second to remind you all

(00:59:14): who we are and what we're doing here.

(00:59:16): Um, this is vibe code camp.

(00:59:19): Uh,

(00:59:19): we're,

(00:59:20): we're going to be here live all day with the best vibe coders in the world showing

(00:59:24): you about their workflow.

(00:59:26): We just had a bent also from Ben's bites coming on.

(00:59:28): We have a stacked guest list.

(00:59:30): Um,

(00:59:31): if you're interested in figuring out what's,

(00:59:33): uh,

(00:59:33): what the agenda is,

(00:59:34): you can go to every.to slash agenda.

(00:59:36): And I'm going to just throw that in a chat here.

(00:59:38): Um,

(00:59:40): We've got people from Google, from Anthropic, from AMP, from lots and lots of cool places.

(00:59:46): All the top body coders in the world are coming up.

(00:59:50): And now I'm excited to welcome our next guest.

(00:59:55): Ash, welcome.