Autumn Nash: Yeah. And I remember that there was this dad, and he was so offended that they thought his teenager was pregnant, and she really was pregnant… And it was like how they were going off of people’s purchases. So can you imagine, with so much detail of like human curation and like interests, how does this change algorithms, and then how is that used to profit off of? In a way, it’s kind of cool, because maybe it helps us to find more of our tribe out there, or people. But then also, how will that be monetized? So it’s both interesting.

Erik St. Martin: Yeah, I mean, step one, have cool Ashley McNamara logo. Step two, question mark. Step three, profit.

Feross Aboukhadijeh: Yeah, I don’t think they wanted you to type scripts in… I’m pretty sure they tried to stop scripts… But the CSS and all the other customization - I agree. I remember those days, it was really cool. You could try to make your profile stand out, and people learned HTML and CSS in order to make their profile cool. It was a real reason to learn to code. I remember that.

[07:53] Then what happened though - so the story continues, actually… So an hour after he hits a million friends, he actually realizes that – a friend contacts him and says “Hey, I can’t actually see my profile anymore, or anyone else’s profile, or anything. The site seems to be broken.” Messages start appearing on other websites, where people are talking about how MySpace is now down for maintenance. Everyone at MySpace is working on getting the site back up… Now he’s like “Oh, snap! Should I drive to their office and apologize to them? I think I broke the website.” [laughter]

Two or three hours after that the site seems to be back up, but all the code is gone… But the string “Samy is my hero” is still on everyone’s profiles.

Feross Aboukhadijeh: I think obviously getting money from companies has to be the strategy that we adopt going forward. So one route is contacting companies directly, and that’s what I’ve sort of tried. It sort of works if you are persistent and you’re willing to email a lot of people and explain to them the benefits of getting their logo on your open source projects page. You can sort of end up getting a few thousand dollars a month if you really work hard at it.

Obviously, the downside to that is now you’re spending a bunch of time emailing people and having meetings about sponsorship issues, instead of coding. And I think a lot of maintainers just don’t wanna do that. So I think that’s where the promise of something like Tidelift comes in.

The idea there is instead of maintainers having to interface with all these companies and trying to explain to them why they should be caring about their dependencies and the shape that those dependencies are in, Tidelift can just go out and do that, and they have a sales team of people who are just basically going out and talking to companies and trying to convince them to pay for open source.

[43:58] And then they turn around and they give half of the money that they collect from those companies straight to the maintainers. I think they’ve promised that they’re always gonna give at least half of their profits indefinitely.

That’s a cool model, because now suddenly I don’t have to worry about talking to people, I don’t have to worry about emailing people and doing all this sort of salesy stuff, which – I mean, I don’t mind doing it, because I like to push myself to learn new skills and to go outside my comfort zone, but I know a lot of maintainers don’t wanna be spending their time basically being a salesperson… So this is actually a promising model.

The way Tidelift does it is they look at what their customers/subscribers are using, they look at what packages are their customers relying on, and then they have some kind of an algorithm that tries to calculate how much of the money that they’ve collected should go to the different dependencies. As a maintainer, you just sign up and collect that every month, and there’s really very little you have to do in return at the moment. You just sort of collect it.

I’m right now making about $500/month from Tidelift. So it’s great, but all these things are all just – the thing that’s not ideal about all these different approaches is just that if I were to go just get a job at Google, I can make way more money.

James Snell: Okay, so right now what’s inside V8 - you have the compiler, and you optimize with the compiler: it takes the JavaScript and compiles it down to bytecode, optimizes it and does a number of tricks to make it faster. The TurboFan+Ignition is just a new compiler and a new optimizer inside V8, and for most people, whether you’re using the old Crankshaft stuff or the news stuff, there’s just not gonna be any impact, you’re not gonna see any difference. But the new compiler and optimizer is really geared towards optimizing the new language features, so things like class, or spread arguments, or any of the new language features. Crankshaft - which was the old optimizer - just wasn’t set up to handle these new language features very well at all. The new toolchain will make it so that those things can actually perform decently going forward. For most people, they won’t see a difference.

[11:45] The code that is going to see a significant difference is code that has been highly, highly optimized to take advantage of quirks in the old Crankshaft optimizer. We have some examples of this in Node core itself, where we actually will unroll for loops in order to make it faster. What that means is instead of just going through the entire loop every time, we’ll see “Does this thing only have one item or two items or three items?” and we will special-case handling those, and then loop only if none of the other more optimized cases apply. Under Crankshaft it ends up being significantly faster.

Code that is highly optimized for Crankshaft could actually end up running slower under TurboFan, because a lot of those optimization tricks are not longer there or are being done a completely different way.

We shouldn’t see code break, we shouldn’t see any existing stuff just stop working, but we could see a very different performance profile under the new tool chain once we enable that. That’s where the biggest difference people will see is - just the performance of their applications will change.

Una Kravets: Tool tips. Even like notifications, menus, a little profile icon… And the other cool thing is you could declaratively have the layout of that change, if you run out of screen space, for example - so you could use the viewport to actually reposition it without any JavaScript. You don’t need to have libraries. You could do this all in CSS. Even in a line of code [unintelligible 00:46:35.28] just to autoflip. And it’s so cool, and part of the new proposal is a tether pseudo-element, so you can create an arrow which would then auto flip to match the flips. You don’t have to do any of that in JavaScript; you don’t have to rewrite all these rules of resize observation, and all of that, scroll observation for your popovers, your little tooltips… It’s going to be something you can do in CSS and HTML; connecting it in HTML with the anchor, and then applying the styles in CSS. Or you could connect them in CSS, too. [laughs]

Jesús Espino: Yeah, with a profiler, right? With a profile guided optimization. So yeah, that can be very cool to see that is going to have a huge impact, because probably it’s going to have a certain impact there.

Jaana Dogan: If we’ll run out of time, it’s actually okay… This question really hit me hard, because I have exactly two weeks at the end of this year to do whatever I wanna do. And the question is – probably I will use Go if I wanna write code… And the question is “What am I going to do?” This is more of like an existential crisis type of question for me probably… [laughter] There was one crazy idea I had for a long time. I was wondering if it’s possible to highlight some of the concurrency related stuff in a text editor.

Imagine if a library – if you’re making a call, it starts the goroutine, whatever… You just don’t know from the API surface, but the goroutine just runs some stuff, and other goroutines, and whatever… I wonder if you can highlight in the editor that “Oh, some of the pieces here in this block is just going to run a different goroutine”, or may run in a different goroutine, or whatever.

I was thinking about this dynamic tool - you just run your program, it maybe collects some profile or whatever, and then you apply that profile to your text editor and it gives you all these different colors. “This has been run in these different goroutines” and so on. So it helps you to – it’s not a perfect solution, but it could be a good experimentation point and may influence maybe some other people to work on this type of problem.

We had this discussion last week - there’s no good way to say “Hey, I’m just going to run some stuff in a different goroutine.” Some libraries are doing a good job documenting this, but some others don’t. So that is an interesting area to work on.

Miriah Peterson: So I think this is a very interesting topic to bring back to Go… So yeah, I was doing research for a course I’ve been workshopping, trying to do more – anyway, the topic is Go and Data Engineering. And going into it kind of with this idea - I’ve been doing a lot of stuff with pprof, which is Go’s profiling tool, which we should circle back to… I’ve been doing a lot of stuff with pprof, trying to understand goroutines… Basically, at the orchestration layer, is it better to use a goroutine and have it in one program, or to scale out your programs? Should you do horizontal autoscaling in Kubernetes, or should you try doing workers internally? Those kinds of things. And part of the problem with that comes to essentially basic API calls, is how I’ll put it. So it doesn’t matter what program you’re building. Go uses the IO reader and IO writer for most of its – like, that is the interface behind all its orchestration, whether you’re connecting to a database, whether you’re connecting to a streaming service, whether you’re connecting to an API… It all goes back to that layer. So it doesn’t matter – like, if you’re writing a service that’s going to be handling hundreds of thousands of API calls, it’s the same system as if you’re handling hundreds of thousands of database writes. Or if you’re processing hundreds of thousands of data points, the latency in Go doesn’t tend to – unless you have built your program incorrectly, the latency and the slowness doesn’t tend to come from the actual data munging, or the manipulation of the data point. It comes from those connections to and from file systems, or to or from API points.

[18:12] So when you’re designing your service and you’re trying to optimize things in Go, that is the point where you tend to have memory issues. The memory issues often come from those connection points or those API points, the point where you’re sending data from one function to another, or one system to another; it doesn’t come from the actual horizontal scaling of the service. It is from those design choices at the IO read/write level; like, the “Oh crap, I forgot to close my writer. Oh crap, I have 15 connections open when I really only need one.” Those are the different things… And I feel like that’s the same no matter what program you’re building, because all of us are manipulating bytes at the end of the day… And those are the kinds of things that I find people tend to turn to Kubernetes logs, or other kinds of things, and not use some of the stuff that Go has built in to help us track that kind of stuff.

Kemal Akkoyun: We haven’t jumped into that, but right now, I’m working on a continuous profiling tool. We haven’t even basically defined the profiling, but let’s say we define - profiling is kind of an observability signal. It’s an action that you collect the resource usages of your processes that’s running on your machines, right? And what makes it fun is this continuous profiling part. So you are profiling all these processes, and then you release a software; maybe you scale up, you scale down, but you always have these snapshots of memory usage or CPU usage in a timespan, right? And then you can actually check between versions if you actually changed something. So when you have a continuous signal like that, you can actually optimize or do something about any of these actions, right?

[17:56] Same goes for metrics. You have a latency metric, and you can define some SLO-based alert on that. And that alerting part actually makes the difference, right? So all of a sudden, you release a version, and your latency has spiked, and it actually triggered an alert. Now that you know – since you put that alert in there, you know that you are actually exceeding your allowed boundaries, your error budgets maybe. And then you can actually get in there and try to optimize, try to do something about that. But it’s the fun part of that, right? We mentioned the profiling side of that, and we mentioned the metrics side of that, and we mentioned alerts, right? They come all together, actually. You get the alert, you check what’s going on with the process, down to the method and function, and then you decide to do something about it.

Brian Ketelsen: Yeah, who knows…? So Prism is a pretty nifty open source profiling tool for Go code, and the thing that I really like about Prism is that it shows historical diffs over time. So you can run a profile on your app, and it has pretty decent-looking output - better than pprof, obviously - and as you make changes to your application, you can view the difference in those profiling changes from the most current revision to the last one, or you can view the entire profile difference. So you can say “This change made me 5% faster, but since I started profiling, I’m 1000% faster.”

[43:57] It’s really neat to show historical changes in your profiling. And it’s got a nice interface, it looks really easy to use… I have not yet tried to use it, but it looks pretty nice. So Prism… I don’t even remember what company released it, but it looked pretty cool.

Olivier Lacan: [27:54] Right. So what I encourage listeners to do is basically what I’ve been starting to do in the last few months - I’ve put this page up and I’ve used this page as argument bullets to go inside of a project that I use. Basically, when I do an upgrade, there’s a shock and I don’t know why something doesn’t work, and I say “Okay, let’s see the changelog.” If there’s no changelog, I say “Okay, do you want me to make a changelog for you? This is what I offer, this is the format. Are you cool with that?” and at least two, or three, or four, or more people actually have been like “Yeah, sure. You can do it. Just send a PR.”

I think one of the contributors for Discourse, a really great British guy… I can’t remember his name. But yeah, if you go to Discourse, they use something to – rack-mini-profiler. It’s this really cool profiling thing that you can use in Ruby apps. Sam Saffron. And his changelog was strange, because it was backwards; so the oldest changes was at the top, which is kind of counter-intuitive… And the date format was a little strange, because he’s British; we can’t judge, but yeah…

Gerhard Lazu: …and I’m taking the whole Changelog application, which has migrated to Dagger, part also of episode 33, when I mentioned the Terrajet provider… So like all those three things we are using. Parca is the other tool for CPU profiling. So all those things we are already using in production. And what I’m thinking is what is the next step? What can we learn from Kuby? Where do we meet? Because I’m sure that we meet at some point, so where do we meet today? And where do we want to meet six months from now? What would make sense? Because I’m so keen on solving this problem once and for all. I mean, it’s maybe one of the reasons why I’m back in this CI/CD space full-time, because it feels like I have unfinished business there. So as we prepare to wrap up, what is your take-away for the users, for the listeners?

Jaana Dogan: Yeah, so all the pprof tools – pprof can be tweakable. You can turn it on dynamically, and you can turn it off. So what we do is basically turn it on for several minutes, collect the data, just parse it, store it, and then we aggregate all that data, and we have this daily, weekly etc. report… And you can take a look at “Oh, this service, particularly this handler, is often used, and all these particular functions are accounting for the most CPU time, or memory, or whatever”, and you can just go and dig and optimize those particular places.

I wish that Go had some tools around maybe supporting this type of more continuous profiling features. It’s possible to write a tool that aggregates multiple pprofiles. It could be possible to write a library that automatically just turns once-in-a-while reports to some central service, and then turns it off, and so on. I think we can do much better in this field; it’s just kind of up to the user right now to plan and design and do this type of thing… But that’s basically what we do. I wrote on this topic for a while.

Some companies are aware of these methodologies and some companies are not. It would be so nice if the community was producing more best practices, as well as more tooling around this.

Jaana Dogan: [01:00:01.11] Yeah, so all the pprof tools – pprof can be tweakable. You can turn it on dynamically, and you can turn it off. So what we do is basically turn it on for several minutes, collect the data, just parse it, store it, and then we aggregate all that data, and we have this daily, weekly etc. report… And you can take a look at “Oh, this service, particularly this handler, is often used, and all these particular functions are accounting for the most CPU time, or memory, or whatever”, and you can just go and dig and optimize those particular places.

I wish that Go had some tools around maybe supporting this type of more continuous profiling features. It’s possible to write a tool that aggregates multiple pprofiles. It could be possible to write a library that automatically just turns once-in-a-while reports to some central service, and then turns it off, and so on. I think we can do much better in this field; it’s just kind of up to the user right now to plan and design and do this type of thing… But that’s basically what we do. I wrote on this topic for a while.

Some companies are aware of these methodologies and some companies are not. It would be so nice if the community was producing more best practices, as well as more tooling around this.

Bill Kennedy: Cool! I’m kind of excited about two pieces of this update. The new profiling that you can get out of a tracing tool, and I’m really interested to see how the plugin stuff ends up working and being adopted, because some of that is a little – we can get in a lot of trouble with a plugin, but it’s also really interesting.

Erik St. Martin: Yeah, that’s what it is. Some other cool things - the HTTP library can support taking a context now, so you can cancel HTTP requests as part of a context. One of the other cool things I was excited about was the mutex contention profiling. Up until now you couldn’t actually see which goroutines and things were contending for a particular mutex, and then there’s like a whole series of other stuff… I’m trying to remember what all was in there, but those were some of the fun ones I saw.

Nick Sweeting: Yeah. So no matter what, I’m going to have to be both. There has to be a nonprofit component, there has to be a for-profit component. It’s going to be a sort of peer corporate structure relationship, similar to any company that does massive content re-hosting, like Archive.org, like OpenAI, like Mozilla, like MAPS… Basically, you have a nonprofit and you have LLC’s underneath it that do anything relating to money. The content is only ever hosted by the nonprofit, which is not earning revenue for it, but you can sell software that people use, that contributes to that pool of content.

And so the financial motivation to – basically, the financial motivations are kept separate. You’re not incentivized to profit off of the copyrighted material, which I think is important… Because as this eventually grows beyond just me, I don’t want to have sort of corporate structuring that is pushing it in the direction of destroying copyright.

Kris Brandow: [38:16] Yeah, one of the thoughts I had during what you were saying is - I express this in private to some people, and I always get kind of the “You’ve just said heresy” look or comments… But I have been wondering, is open source sustainable as the method of how we do things in this industry? Is this focus on sharing so much of code actually going to be something we can continue doing in the future, since there’s so much ambiguity around all of this? And quite frankly, I think it also atrophies the whole industry, because we’re not rewriting things. We’re not reimagining things. I think that’s one of the core problems with copyright in general - there’s this whole thing that Disney has done, where it’s just like… Copyright used to be like 20 years. Now it’s almost forever; or as Disney would like to have it, actually forever. So it’s like, “Oh, well, now these things are just kind of sticking around, and it’s so much harder to move things forward.”

If I remember the kind of genesis of copyright, or the vague genesis, it’s like “We want to protect people, making so they can profit off of their creative work for some time”, but then it goes back into the general pool of things so we can kind of continue making progress forward.

Chris Benson: I think there’s certainly a chance at it, and I would argue – it’s the same argument I’ve made in previous shows where we talked on similar topics, is that we’re seeing… As the AI industry is has been maturing these years at an incredibly rapid pace, but we’re still seeing many of the things occurring that we saw when the software world was really maturing over several decades. And the place where open source has really, really worked are in common touchpoints where all organizations or many organizations need a common thing. And they might build something differentiated on top of that, for their revenue, to drive profitability… But there’s so much that is underneath that point of differentiation that they and many other organizations can get the benefit out of a lot of effort, a lot of work; a lot of times they’ll have paid employees do it…

So there’s a point where working together and doing open stuff makes sense for business, and it drives profitability. It may not be your single point of differentiation, but if it’s anything under that, why not? Why not share the costs, and pool expertise for the best possible foundations? And so what I’m hoping is that we continue to see that play out in the AI space. We’re seeing – if you look at Hugging Face, we’ve already talked about the fact that a couple of months ago they announced that they were hosting a million models; those are all open source. Really, really impressive. And so I think that there is a good chance that a vibrant, open community around AI can and will continue, and it will have a lot of corporate players involved in it. So I’m very optimistic in that way.

Break: [00:13:19.26]

Jeff Smith: I would say the initial reaction is the eyebrows go up, and they lean in. Music is ubiquitous. It’s all around us. It’s amazing. And it touches our lives. And so you have this emotional currency that we all get. So they lean in and say “That’s amazing.” And then they’ll kind of sit back and say “Ooh, what does my music say about me? If I’m listening to Nine Inch Nails, does that mean I’m depressed? Or what’s going on here in my heart?”

So they go through that, and then they lean forward again and say “Oh, this is amazing. How do I incorporate this in my life, in my profession?” And I would say the surprise or the unique things that came out of those conversations is really these tributaries that were created.

So we built the tech, you patent it, you apply it to a sector that you understand, have influenced in, that has a large enough addressable market, good liquidity, there’s a budget item for engagement measurement… And you apply it in there. I didn’t expect then where this would take us, right? We’ve got mental health professionals saying “Hey, we want it as a screening tool to get ahead of certain things in our clinics”, as well as creating a profit center for them. We have athletes and sports psychologists looking at it for performance. We have the US military needing to address suicide rates and say “Hey, if we could just know more about how they’re doing…” It’s all about that early detection, early indicator on how they’re doing. And again, I want to stress, it’s a screening tool. It’s not a diagnostic or an assessment tool. So inside of therapy clinics, it’ll just get them to that BDI or GAD, the formal assessments quicker, which is kind of cool.

[12:18] So I think that was the biggest surprise for me, is I’ve launched a lot of companies, and it’s around the innovation or the relationships or the opportunity… This one was just – I use the word ‘ubiquitous’. It just – it’s emotional, it’s primal, it’s historic… It’s just in people’s lives.

We come to learn that the average person listens to 22 hours of music a week. It’s all around us. And so that’s just really cool. So I would say that was a big surprise for me.

Ismael Valenzuela: With replicants. There is no replicants as of today. There’s deepfakes. That’s a different thing, that could look like humans. That’s the closest thing. But there’s no autonomous agents that can do all of these things. Or we don’t see people that – I don’t know, like you are training dolphins for your entire life, and then all of a sudden, now because of AI, you can hack into companies and make a lot of profit out of that. Probably not. So what we see is a type of system that uses a tool essentially for the initial phases of the attack. And that means that they’re getting a lot better at writing phishing emails.

We have seen an increase in phishing emails with language that’s non-English, for example. Before, we would see some of these Eastern European organizations or Russian criminals sending emails in English that was like broken English. And you could quickly spot them and say “Oh yeah, this is phishing or spam.” These days, everybody speaks not only perfect English, also perfect Japanese. We have seen an increase in the number of phishing against Japanese companies, or other languages that hardly would be used by these cyber criminals. And that’s a clear use of LLMs.

Now, in terms of coding, there’s a lot of debate. It’s very controversial. Can you learn coding from scratch, or can’t you just use this to create code from scratch, and it’ll do these things? Probably not today. These models are getting better, but I still find out that every time I ask any of these agents to create some code for me, I still have to understand the code, understand what I’m trying to do, and being able to refine it and to tune it. Also, bear in mind that these models are crafting things based on the training that it has received, based on previous data that is already known. Therefore, when Greg maybe talks about predictive solutions and AI, that makes us also even more successful in the use of our AI, because we have trained these models with everything that has been seen in the past as well.

So at the end of the day, I think that AI is not going to be that much of an advantage to attackers. There’s always a little advantage, but just because they’re attackers, because they take the first step, and you’re on the defense side, and you don’t know if they’re coming tonight, or if they’re coming tomorrow morning, or if they’re coming next month. You may anticipate that, and that’s where my team does threat intelligence, which is looking at the geopolitics, looking at the weather forecast, what are the clouds signaling. And then based on that, you adapt your threat model. But you’re always one step behind, by nature. That’s what defense is about.

But even though defenders may have that temporal advantage, I think when used properly by defenders, the field could be leveled, and AI could be effectively used to do more things at scale, especially when you have a solid strategy.

John Nunemaker: Yeah, so price-wise, the multiple on profit was less than that. It was probably more like two and a quarter, or less than that. And again, it’s because – this is a massive credit to Dan. I want to give him props, because he could have just waited, and just collected checks… He could have tried to get a higher multiple for somebody that’s going to harvest the customers… He could have tried like a lot of different things, and he was like “I want the customers to be taken care of. A lot of these people are my friends. A lot of these people I’ve served for a long time. I don’t want it to just go away.” And so that, I think, was a big benefit to Garrett and I, and the others, is that we could come in and say “Look, this is what we’re going to do. We’re actually going to try to grow it, we’re going to make it better, we’re going to take your baby and we’re going to hopefully help it continue to grow, instead of just to sit and slowly kind of decline.”

So multiple-wise, he gave us a much better multiple than perhaps like – if any private equity company or somebody else came in, they would definitely have given probably a better multiple. He had a better offer that he turned down before us. So I think it was truly like – he’s like “These are the right people to run it.”

So I don’t have a great example of “Here’s how you figure that out.” What I’ve always done is – it’s basically napkins. It’s like, “Look, you have this much in revenue and you have this much in profit.” And you’ve got to just look at the business and say “Is it fair to just look at the profit, or is should I look at the revenue? Because I can cut the expenses and make the profit larger…” It’s kind of a holistic thing. If it’s at X hundred thousand and the expenses are like 20 percent, the fixed expenses of no one working on it, just the servers and software to make sure it’s up and running… 10, 20 percent, whatever it is, then you can say “I’ve got the rest of that 80 percent to pay debt service”, and also to like pay people to work on it, or pay myself to work on it, or any of those kinds of options. And if I did that, I’ve got this massive spreadsheet, with all these adjustable variables… And I learned these really cool functions thanks to ChatGPT, where I was like “How do I amortize alone over 15 years in an Excel function?” And it’s like “Oh, well, you need to use PMT.” And then like “How much principle is left if I do this?” So I used all those things, and I basically got it to where I could say “Okay, if I grow 0 percent, negative 5 percent, 5 percent, 10, 15…” And I’m conservative, so I’m I’m not going to say that I’m going to grow 50 percent. I don’t feel like that’s going to happen. Realistically, we’re going to put in a good 5, 10, 15, 20 percent effort, and I think we’ll get it out of there as time goes on.

But I don’t feel like there’s any fancy formula. I don’t feel like – probably other people do, that are more experienced. For me, it’s just - I looked at it, I’m like “It seems like a good business. It’s built in the way that I would run a business.” Very low expenses, very high profit margins… It’s recurring revenue… I’m looking at all those kinds of things and I’m like “Why would I say no?” other than maybe I’m too busy. But it has money to pay people to work on it, so it doesn’t matter if I’m too busy, really. It really just matters “Can it afford to pay people to work on it?” And then it’s just a matter of “Do you have the money or the financing options or other things like that to make it happen?” And again, because of my fortunate past, I’ve got some of that stuff lined up, and so I was able to finance this… Because I never have financed anything, from a software standpoint… And I was like “Let’s figure out how to do that. That sounds fun.” And I learned a lot over the last two months, of like how to do that, little weird gotchas from the bank… At the very end we had to like change ownership percentages and things like that, or other people were going to have to personally guarantee beyond me… Which I did not want. I wanted it just to be my risk, and stuff like that.

[40:08] So yeah, that’s kind of like the longer, meandering answer, is like - in general, SaaS apps that are, let’s say, under a million in revenue, are probably worth two to four times profit. Or like seller discretionary. So it’s like, anything that the seller would pay themselves, that plus the profit - they’re probably worth two to four times that. If you get bigger, if you get to like 1 to 5 million, then you can get – I mean Box Out has had offers very comfortably in the six and a half range, that we’ve turned down. We had some stuff in the nine times range, where we’re like – of revenue, not even profit; just of revenue. And we’re like “Okay, well - yeah, we would say yes to that”, but it didn’t quite work out.

So I would say like in that 1 to 5 million range, your multiples can go up a little bit. And then if you get in the 5 to 20 million range, or the 20+, you can get into 10x really easily, and higher multiples. If you asked ChatGPT, that’s probably what it would say. It’s something like those ranges, two to four. So like two if it’s not growing, four if it is growing when it’s under a million, and then one to five - there’s some multiples in there. And probably people would argue with some of those, but that’s a general thing that I’ve seen.

Adam Stacoviak: Yeah. I don’t know, I don’t disagree with that sentiment. But when I logged into Arc, I saw some features, I believe, that were in… I think it’s like this Enable Max feature. And I don’t know all this stuff, what it does. And Nick’s nodding his head again, which I like. I like when you do that, Nick. And I mentioned this when we talked to Chris Wanstrath and Andreas, when we were talking about Ladybird, was I think if we can incorporate… I think Safari has done some base-level security things. And I believe, for the most part, Apple has my privacy in check. Maybe for everyone else, maybe not so much for them. Maybe they’re using a lot of my data, but they’re at least not giving it away, to my knowledge. Right to repair is a whole different issue; we’ll can that for now. But I feel like maybe this Max, Enable Max thing that Arc has might be things like what the Pi-hole does, in the browser; different security things you can do, and different features that make sense, that would be a paid feature, and that would be a for-profit company. Or at least a company that wants to make money to sustain, even if it’s not for profit; it could be public good company. I think when you raise money though, we’ve seen that ChatGPT was – OpenAI was once open, and now they’re not open… So there can be a rug pull, even in venture capital land.

The part I may disagree about a little bit - andl the jury is still out on this for me - is that while I’m not an Arc user, I can appreciate a browser trying to be sustainably capitalistic, and deploy features that are paid, that I may want, and may use the browser so I can pay for them because they’re not anywhere else. Or I have to cobble together some self-hosted stuff, which is kind of cool if you’re into that… But if you’re not, then making that readily available to the masses for 10 bucks a month or some fee may be kind of cool. I don’t know that that’s what they’re doing, but it’s something like that. It’s like, ask on page, five-second previews, tidy up tab titles, tidy downloads… These are things that seem to be free, so I could be misspeaking about these being paid features…

Adam Stacoviak: That’s right. Well, there’s some UTM data on there that we’ll somehow find a way to put in the voice there… Yeah, I’m just enamored by what you’ve been able to do with this. I think it’s really just so cool. I think the only last question - maybe this is a good closing question… I’m on your GitHub ,profile and either I’m blind and I don’t see it, or it doesn’t exist, but you’re not taking any sort of sponsorships for this. Now, I imagine you’ve got a job, so you’re okay with that… But I’m sure there’s some way, not so much so you can profit from this, but at least get back money for your time… Is there a reason why you haven’t set up a sponsor page, or allowed the community to do anything to support your efforts?