SPEAKER_00: And we found it to be really magical where when AI, again, is combined with the chat experience, doesn't make sense for employees of a company to go to one place to do their AI chats and another place to do their human chats and should all be in one place. I think the AI becomes powerful enough eventually that you don't even have to ask it a question. It'll just jump in and provide help as needed like a virtual teammate would. And then we have to make sure that the quality of those answers is good enough that it doesn't become a pest. I think we will get there probably pretty

SPEAKER_02: soon. I mean, I call that promptless, by the way, right now, AI has to be prompted.

SPEAKER_04: Promptless makes a lot of sense.

SPEAKER_02: Yeah, it'll get good enough that you don't even have to prompt it. And that's another reason why you want the home for AI to be in your enterprise chat app is because when you go off to ChatGPT or something else that's separate, it doesn't have the context to just chime in. Again, to conversations

SPEAKER_00: you're already having with your co-workers and add value. That's one of the reasons why we call this

David Sacks: glue is we think of it as the hub between your humans, your apps, and your AI.

SPEAKER_09: This Week in Startups is brought to you by Squarespace. Turn your idea into a new website. Go to squarespace.com slash twist for a free trial. When you're ready to launch, use offer code twist to save 10% off your first purchase of a website or domain. LinkedIn jobs. A business is only as strong as its people and every hire matters. Go to linkedin.com slash twist to post your first job for free. Terms and conditions apply. And dot tech domains. Don't miss the jam session with Jcal contest for your chance to join Jason on air to showcase your startup and potentially take it to the next level. To apply and get more details, go to jamwithjcal.tech. Brought to you by dot tech

SPEAKER_18: domains. All right, everybody. Welcome back to this week in startups. We have a great treat for you to my bestie. David Sachs is here back on the program. People don't know this, but 14 years ago, David Sachs was, I think the 10th. No, it's number five. Oh, number five. Oh, he remembers. I guess that's in year. Twist number five. Twist number five. And here we are 1914 years later. 005. 005.

SPEAKER_24: Exactly. We both had less gray hair and there was no all in podcasts back then. We were just trying

SPEAKER_18: to build products and services. And here we are today to talk about Sachs' latest product. He's back to his roots working on a product called Glue. Glue is a Slack killer and his co-founder is Evan Owen,

SPEAKER_27: who's with us on the program. Evan, welcome to the program. Thank you. All right, Sachs, let's just

David Friedberg: start with this very basic question. You're a venture capitalist. You're an all-star podcast co-host

SPEAKER_18: and why build another product? Like it's exhausting to build startups and here you are exhausting yourself, going back and starting a company. So talk to me about like this itch that you keep

SPEAKER_37: going back to scratch. Well, Jason, why have you done, I don't know, 1200 episodes and counting of

SPEAKER_00: podcasts, right? I love it. But actually we talked about it when I was on twist number five, I think I was on there to promote Yammer. I think we had just launched that. Yeah, you'd started with Genie.

David Friedberg: You were kind of doing a genealogy site, but then you internally were using Yammer to build that site

SPEAKER_00: and you realized, wait a second, this product's better. Yeah, we kind of pivoted. Yeah, that was the era of social networking and we kind of pivoted from doing a family social network to doing a corporate social network. And that was the beginning of, I'd say, intra-company feed-based collaboration. The paradigm then for company communication was feeds. Then it evolved to kind of IRC, channels, chat. And now we think it's moving on to another paradigm, which we call threads, which we'll get to in a minute. But back to your question, you know, we sold Yammer four years after it got founded. We launched in 2008, sold the company to Microsoft in 2012. It was a very fast ride. And I still felt like I had a lot of ideas for the space that we never got a chance to implement. And then there was a lot of changes. Like I mentioned, chat and kind of channel-based communication really took off. I had a bunch of thoughts about that, why it wasn't really that great. And now we're in the era of AI. There's more thoughts about that. So I had a lot of, I felt like I had unfinished business in the space. So did Evan, because he was the head of engineering at a company called Zinc, which was also doing messaging. And they sold that company to ServiceMax. And so we got together, Evan started as an EIR at Kraft, and then we incubated this company. And we've been working on this project sort of quietly for the last couple of years.

SPEAKER_44: Yeah. And Evan, you took the EIR, entrepreneur residence position, and started building this

SPEAKER_18: inside the company in stealth. And let's take a look at the product. And maybe you could highlight for the audience, who is almost certainly using Slack or Discord. A lot of startups these days using Discord because it's free. I guess it's got a gaming route. So some young people seem to like that. And then the industry standard is kind of Slack or Teams, right? Microsoft Teams. So maybe you could fire it up. And for those people who are listening only, you can watch this on YouTube to search for

SPEAKER_27: This Week in Startups, or we're going to describe what you're seeing on screen.

SPEAKER_50: All right. So this is, this is blue. Okay.

SPEAKER_51: First thing you'll notice is there's actually a lot of familiar elements to our UI. We definitely recognize the popularity of team chat. There's a lot of reasons why people love Slack, as well as maybe have some annoyances. But what we've tried to do is figure out where the problems are and improve those rough edges, and then just embrace the kind of standard things that everyone expects from work chat everywhere else. The main thing that we wanted to solve with blue is the issue of channel fatigue, which is this problem where when someone needs to add you to a thread, they have to add you to the whole channel, which might have many other threads that aren't relevant to you. And so over time, you start ending up in more and more channels that, you know, just kind of pile up in your sidebar, and then you have to check to make sure there's nothing for you. So what we've done with blue is inverted that and made the thread the most important unit of conversation. And so these are my threads in the inbox here. A thread is conversation with a name, quite simple. But where the power comes in is you can add multiple recipients and a recipient is a group or a person. And so you can add multiple groups to a thread, a thread doesn't have to live in just one place, it can actually be shared between groups.

Chamath Palihapitiya: So here you have invite link API is the thread that you're working on, but you added the developers and

SPEAKER_18: the API team. So those are kind of like in the old days when you might have groups in your outlook or

SPEAKER_58: your email. And so now this is one contained thread about one item of work that you're trying to get done at your startup or your company.

SPEAKER_00: Right. Yeah. So this thread here is called invite link. It's been sent to the design and product team. They're working on a feature where they want to be able to send an invite link to a whole group and then they can all click and sign up. In any event, in Slack, if you were to post that to say the product team and the designers needed to see it, then you're going to have to add all the designers to the product channel in order for them to get access to this one thread. Well, over time, every single employee ends up in every single channel just to get that one thread they need. We've basically jailbreak the threads from the channels. So you have the thread, you just address the people in groups that you need, and you don't get all this noise that you don't need.

Chamath Palihapitiya: And then if the thread dies out because you've solved the problem, it moves down your list.

SPEAKER_65: Yeah. You just archive it. Yeah. So it becomes archived and...

David Sacks: You just click the little arrow there that Evan just moused over and boom, it's gone.

SPEAKER_66: Just like in your Gmail or your superhuman, you just archive it, put it away.

SPEAKER_18: It still comes up on search, I assume. And the challenge with the oppression of the Slack channels becomes once you're added to so many of them, there's so many discussions occurring, as you've pointed out, Saks, that are not relevant to you. So then you're left with two choices. Have the number 752 sitting there above the channel creating anxiety for you that you feel like they

SPEAKER_58: need to clear it, or you leave the channel. I find myself looking at 50 channels that all have

SPEAKER_34: discussions on them. I'm like, this is now no longer helpful, right?

SPEAKER_00: Right. So in this case, the threads inbox, so you have one inbox, which is really nice for all of your threads and all of your new group chats and your new DMs. There's just one place you have to go. The threads that you're a participant in, where you've either commented or you started or you've just chosen to follow, they're going to be in your inbox. The rest of the threads you can find by going to the groups or by going to the feed. And we've created a feed. This is kind of like very Yammer-esque, where you can use this to scroll through. And just like, when you have extra time, you can just scroll through this to see everything that's happening in your organization.

SPEAKER_02: The pulse of the company, right? Yeah. This is the pulse. This is about discovery, but there's no obligation to have to like, go through all this stuff because you know that all the important stuff's going to go in your inbox.

SPEAKER_44: Yeah. And the prioritization is happening based upon whatever was updated last, which is helpful

SPEAKER_58: because it's top of mind. You could just start with whatever the most recent active discussion is.

SPEAKER_74: Yeah, the sort is most recent.

SPEAKER_44: But I see you have a sort, you know, a little funnel up there. I'm sure you can change it to other.

David Sacks: You could choose between recent or unread. Those are the two sort choices right now.

SPEAKER_66: Much more efficient.

SPEAKER_78: All right, startups. You ever notice how successful businesses are constantly evolving? They're constantly adding new features and you get that product velocities? Well, that's why I am all in on Squarespace and you should be too. My pals at Squarespace keep adding new features and revolutionizing their platform all the time. Week after week, month after month, year after year. And they've added AI. Squarespace Blueprint AI is amazing. It's their guided design system for building a new website. It's fast, it's custom, and it's built specifically for your business needs. They've put over a billion design combinations right at your fingertips. And so you get an online presence that's as unique as your brand. We used it to build our Founder University website. The process was so simple. Basically, the AI prompts you with some questions and then the Blueprint AI took control, tailoring everything to our needs from website structure to color schemes from design elements to my favorite, fonts. And boom, within minutes, we had a sleek new website design ready to go. Easy, easy, lemon squeezy. AI is changing everything. And this Blueprint AI tool is something to behold. Check out squarespace.com slash twist for a free

SPEAKER_82: trial. And when you're ready to launch, go to squarespace.com slash twist to get 10% off your first website or domain purchase at squarespace.com slash twist.

SPEAKER_58: And so the great feature that we were playing with in the all-in instance was, of course, AI. And I

SPEAKER_84: think this kind of steals the show. It has incredible potential. What's the vision here for AI, Evan?

SPEAKER_85: You know, I think we're all becoming used to chat GPT style conversations with AI. And one of the

SPEAKER_51: reasons is because AI really needs a contained focus in order to give you the best answers. It needs to know what you're talking about and not be distracted by other things. And of course, that's why people also benefit from a focused context. But what we discovered in the past year or two, as chat GPT started taking off, is that what we were building here actually dovetailed perfectly into the way that we interact with AI through conversation. And so what we want to do is really add AI right into your workspace, almost like a teammate. What would it be like if you could interact with this AI instance that also has visibility into all of your conversations, that knows everything that you know or could know about what's happening in Glue? And then also, even what's outside of Glue as you plug in integrations and send attachments and whatnot, then Glue AI can learn and benefit from all of that information in one place. So that's the vision.

SPEAKER_64: What's the number one data source people want to add to your HubSpot, Salesforce? What comes up often?

SPEAKER_85: We're hearing a lot of Google Drive, of course. We're definitely hearing, you know, on the

SPEAKER_51: developer side, you know, your project management tools or GitHub or whatnot. But definitely, I mean, look, there's benefit to every single team. So what we want is to provide those connections and pull in information so that, you know, your sales team and your marketing team and development team, whatever tools they use, can funnel in to Glue AI and benefit from it. Okay. Let's ask it a question.

SPEAKER_96: Yeah. Sure. Let's see here. Find out what my designer has been up to. Our designer here, Jason Lang.

SPEAKER_51: Oh, wow. Of course, Glue AI knows about people. And so I'm just going to, it knows now that it needs to do a search. So it's going to find some conversations and then give me an overview of what it found that Jason has been working on.

SPEAKER_64: And so this is like the killer feature for managers. Like this is the wow moment. You are constantly

SPEAKER_86: wondering, what are people up to, especially in a remote world or a distributed world? And you ask

SPEAKER_58: people for reports, but it turns out what they're typing and how they're interacting in their messaging platform speaks volumes. And here we have that he's working on new user experience in the onboarding

SPEAKER_86: flow, the old UI audit and invite link functionality and summarizes it quite nicely. So that to me is just

SPEAKER_106: killer for management. It's killer. Yeah.

David Sacks: You can also ask it about like who in your organization has expert. By the way, the AI gave

SPEAKER_00: the, it auto generates a thread name. So if you're wondering where the thread names come from, you can edit them, of course, but that's a really nice feature because it then uses those thread names as sources. So you saw in the previous answer, the AI cites what threads is getting the information from. So you can just click to go to that. And you can also at mentioned threads. So it's very easy for referencing any event. Here's Evan's asking a question about who in the company knows about iOS

David Sacks: or Android development. Let's see what it comes up with.

SPEAKER_27: So it looks like Marco Sven, Sven, has experienced dealing with UI users on iPad devices. Hmm. And that's from another thread request to join group profile bug. So it's, it knows that.

SPEAKER_00: Yeah. Yeah. I've asked it before, like who it crapped has expertise in fill in the blank. And it'll give me the person on the investment team who's done the most work in that area.

SPEAKER_18: It's amazing. This is asking questions is great. I guess, David, you and I are probably looking at this saying, when is it going to tip over into proactively participating? So have you thought

SPEAKER_66: about that where it starts alerting you, you're having a conversation about APIs and it says, you

SPEAKER_00: know, we should probably have Sven involved. Oh, look at this. So I knew that Evan was actually knew the most about API development at glue, but look at all the, the threads that cited as sources. Pretty awesome. Yeah. So, you know, there's all these like really expensive, complicated corporate internet products to, you know, identify expertise in the company. They're all based on people filling out profiles. I get really out of date really fast. Yeah. What's great about integrating AI with the chat history is it, it stays current, it stays fresh. And so it's really, we found it to be really magical where, when AI again is combined with the chat experience, you know, it makes sense, right? I mean, chat GPT is a chat experience. That's the way that people have learned how to interact with AI. It doesn't make sense for employees of a company to go to one place to do their AI chats and another place to do their human chats. It should all be in one place. But to go back to your question, Jason, I think the AI becomes powerful enough eventually that you don't even have to ask it a question. It'll just jump in and provide help as needed, like a virtual teammate would. And what do we have to do to get there? I think it's mostly just about speed and cost and quality. Like the, we have to run every, um, every post in glue through the AI and then ask it if it, if it has something relevant to say, and then we have to make sure that the quality of those answers is good enough that it doesn't become a pest. So I think we will get there probably

SPEAKER_02: pretty soon. I mean, I, I call that promptless, by the way, right now, AI has to be prompted. Yeah. I think eventually it makes a lot of sense. Yeah. It'll get good enough that you don't even have to prompt it. And that's another reason why you want your, the home for AI to be in your enterprise chat app is because, you know, when you go off to chat GPT or something else that's separate, it doesn't have the context to

SPEAKER_06: just chime in again to conversations you're already having with your coworkers and add value.

SPEAKER_18: The interesting thing would be to start building roles. So you have people in your organization who already have roles and you're watching their behavior. So you already have developers, designers, maybe a sales team, maybe SDRs, you know, if you created a role of SDR and you had a

SPEAKER_58: virtual, you know, an AI SDR working with your sales team, you could just ask it like, you know, Hey, who, who have, what customers haven't we talked to in a while? Or it just might insert, Hey, here are some customers we haven't talked to in a while who might be likely to churn. And it just knows that churn is something customer success people focus on because it was watching them,

SPEAKER_127: facts. And I haven't previously asked questions about that. Right. Cause some boss came in and said like, who's in danger of churning, you know, uh, how should we avoid that?

SPEAKER_00: Yeah. I think it's a pretty simple thing we could do is just let every group in glue have its own custom instruction for, uh, glue AI. So it could just tell glue AI the way it wants it to interact with that group. Uh, that'd be like a really easy thing to do. I mean, this is all on the product roadmap. It's, you know, again, once you're thinking in this way of how do we make human chat and AI chat seamless, then you start getting into that, you know, into those features.

SPEAKER_18: And if you're studying the behavior of the sales lead or the tech lead or the CEO of the company, it's going to know, Hey, over the last five years, this person's asked these kinds of questions,

SPEAKER_58: et cetera. When that person's working on something, they could get anticipatory management and questions. Hey, your manager's probably going to ask about X, Y, and Z. And just, again,

SPEAKER_132: they're promptlessly inserting themselves to help. Right? Yeah. And a big piece of this is also the

SPEAKER_00: integrations, right? Because if you, if you're, you want to integrate with, um, Salesforce or HubSpot or wherever your customer data is, all of that stuff, then the AI has context and is more valuable. So that's one of the reasons why we call this glue is we think of it as the hub between your

David Sacks: humans, your apps, and your AI.

SPEAKER_69: And I noticed that you can plug and play different AIs into this, Evan. I saw you have Claude, you got ChatGPT 3, 5, 4, 4, 0, whatever. Are you seeing that any of these models are outperforming

SPEAKER_58: the others particularly well? Or do you think we're going to be in a world where these models quickly become commoditized? And do you think you're going to maybe use an open source one as like the standard

SPEAKER_34: one in here? How do you think about all the different models that are available? Because they're moving very quickly. They seem to leapfrog each other every couple of months.

SPEAKER_51: Right. I mean, that's what we've found really. And I think that's, that's why we've taken the stance that we have is that we want our users to have the best model possible. And the best way to do that is to give an option. So yeah, definitely some of these work better than others. We've also noticed some work better for certain types of queries. And so by default, we have an automatic setting, which lets us kind of pick and route depending on, on the question, the context. It does seem like over time, these models will get commoditized to some degree. We just want to be ready for wherever that goes. Hey founders, hey startups. When you're hiring for your small

SPEAKER_140: business, you want to find quality professionals who are right for the role. But I understand it can be tough to find the time and resources to find that perfect hire. That's why you want to check out, you guessed it, LinkedIn Jobs. LinkedIn Jobs gives you the tools to find the right professionals for your team faster and for free. And LinkedIn isn't just some simple job board. It helps you hire professionals that you cannot find anywhere else. There's over a billion members on LinkedIn now, and they're in more than 200 countries. And I know many of you are leveraging different time zones to get work done, different geos to find talent. And LinkedIn is the best place to get eyeballs on your job posting. You know that even those who aren't actively searching for a job will see your ad. Why? Because LinkedIn knows who might want your job. You know, they got a lot of AI and machine learning going on over there. So even if they're not actively searching, they're going to see your role. And you know what? Those are some of the best candidates. And did you know that 70% of LinkedIn users don't visit the other leading job sites? If you're not looking on LinkedIn, you're looking in the wrong place. And the best part is when you use LinkedIn Jobs, 86% of small businesses get a qualified candidate in 24 hours. So hire professionals like a professional on LinkedIn. LinkedIn Jobs helps you get the right people faster. And guess what? The first one's on your boy, J-Cow. Post your first job for free. LinkedIn.com slash TWIST. That's right. LinkedIn.com slash twist to post your job for free. Terms and conditions do apply.

SPEAKER_18: Saks, let's talk about inserting a product into a very competitive landscape. How do you strategize

SPEAKER_58: getting people off of Slack, which is, you know, it's got deep, deep roots into some companies right now, deep API calls, et cetera. How do you start to get people to convert? What's the strategy there?

SPEAKER_84: I'm sure a lot of founders who are listening are curious what your strategy is.

SPEAKER_00: Well, we have a Slack import feature, so you can just export. Slack has a generic export where you can export all of your Slack data to a file and we can import that file. And so we, yeah, we can import all of your previous chat history, the public chat history, the stuff that you want to keep. And we can also import the channels that you want to keep so we can retain your memberships in those key channels and that, but you don't have to take all of them. We'll, we'll give you a selector so you can select which ones you want to take and which ones you want to leave. So yeah, the Slack

SPEAKER_45: import's a big deal. That means that it's sort of backwards compatibility. That means you're not

SPEAKER_150: like starting from a cold start. Yeah. You get the benefit of, you know, I have, I don't know,

SPEAKER_58: 10 years of Slack right now. It's been really interesting as a venture capitalist to go back and say, Hey, this company broke out, or we made a mistake investing in this company. Can I see the discussion of that? And so now in our firm, you have to put your deal memo into Slack. You have to advocate in Slack. People have to ask you hard questions in Slack. And we now can go back

SPEAKER_84: and look for who championed this company that broke out. Right. And that's incredible for decision making and corporate development. So that's just really awesome to be able to import your history.

SPEAKER_152: Yeah. The more context you can give the AI, the better. So yeah, you'll just, you import it. Boom. The AI has all this context. You can start asking questions right away.

SPEAKER_58: I mean, it's almost worth doing that just to start asking questions as the manager, even if you didn't bring everybody over just that alone, to be able to start interacting with it

SPEAKER_84: makes a lot of sense. Now, a lot of people in my portfolio have been asking, Hey, how do I get an invite? It's glueai.com, I believe is the domain. No, just glue.ai. Oh, glue.ai. Got it. Yeah. So the domain is glue.ai. You can sign up there, but what's the pace at which you're going to add people to this? I know lots of people are waiting to get in and then talk a little bit about

SPEAKER_00: your onboarding strategies. We're doing white glove onboarding, basically superhuman style. I don't know if you remember, superhuman made everyone sign up for a wait list and then they did an onboarding call, 20, 30 minute onboarding call to make sure that you understood the product and because it is kind of a power user product and make sure you're set up correctly. So for now, we're doing that where you can sign up to the wait list. The product is public, but you have to sign up for the wait list. And then we're doing demos as quickly as we can. Demo slash

SPEAKER_06: onboarding calls. We have a, how many people are we up to now on the wait list? It's gotten kind

SPEAKER_02: of nutty. 5,700 last I checked. So I think what'll probably happen is at some point here, we're just going to throw the floodgates open. But for now, we're just trying to make sure that

SPEAKER_45: every single team that signs up has the best experience. Evan, as a product manager and a

SPEAKER_66: product designer, what have you learned from doing this superhuman onboarding process? We were

SPEAKER_84: the first investors, ironically, in superhuman. And we invested in Raul's first company. So when he did this onboarding thing, we're like, what are you doing? You're slowing down all this friction. And he said, yeah, but our churn is like nil. And we don't let people use the product we know are going to churn because who's going to pay a dollar a day for email if they're not serious? If you're in your email box for 10 emails a day, you shouldn't buy our product. So we sort you out. We don't let you buy the product. It literally dissuade people from buying the product who don't need it because they don't want to have churn. They don't want to waste people's time. So tell me what you've learned. I'm assuming you've done some of these onboarding yourself. Yeah, we want our customers,

SPEAKER_103: our users to be successful and to love what we've built. We also want to learn from their experience

SPEAKER_51: and what's working and what's not. And so by picking the teams that are the best fit, you know, upfront and walking them through it, making sure that they are getting all the value from it, it just makes total sense. And then we're able to immediately roll that feedback as well. From day to day, each set of demos is getting better from what we're learning the previous time. If we had just let the floodgates completely open, everyone would have had that maybe mediocre experience that the first few demos had. And so we're just kind of continually making that better. And we do want to open it up. But, you know, early on, especially when when users don't know necessarily what to expect or exactly what Blue does, it gives us an opportunity to to share and dig deeper with them.

SPEAKER_86: I think connecting a couple of data sources also makes the thing infinitely more interesting. So if

SPEAKER_58: you can, if you can interview them and find out what their data sources are, and just right there on the phone, if they've got the admin privileges to get them to actually authenticate Salesforce or Notion or Coda or Zendesk, whatever their jam is, that's going to really help, huh?

SPEAKER_00: Yeah, just to build off that, here's the thing about Glue or any product like it is that there's nothing to really do there by yourself, or not much, right? You need some data there to be able to ask the AI questions and get interesting answers. You need co-workers there to be able to have team chats. You need to connect apps to get the benefits you were just talking about. So if someone just signs up and then they're just there, it's like they have their experience kind of like, okay, now what? Now what do I do? So the point of the onboarding call is to get them to the point where they create a couple of groups, they, you know, tee up five to 10 employees or co-workers to invite, and maybe they

SPEAKER_02: connect a couple of apps, and now they're ready to go, and they can actually have a good experience. Now, before doing that, they may have a bunch of questions or concerns, they need someone to answer them. And so I think that's the purpose of this call is get them comfortable enough where they're willing to do the little bit of work they need to do to get this thing jump-started. Remember on Facebook, they always said that if, as a result of going through the sign-up flow, they can get you to friend 10 people, and they kind of had, you're stuck for life. Yeah. So this is kind of like that, right? Like if you just sign up and that's it, you're probably going to a trip. But if we get you to sign up, create a couple of groups, you know, invite five, 10 co-workers, and then connect an app,

SPEAKER_00: then we think you're going to be very successful, and you're going to see why this is better than Slack or any other tool. But if it's just you solo, you're not going to see it.

SPEAKER_86: How are you thinking about pricing for this? This has been a big debate in the industry, SaaS pricing, per-seat pricing versus per-server pricing. And, you know, it's a little bit of

SPEAKER_58: like, I don't know how to say resentment, but resentment of how expensive Slack can get. People are wondering, like, should I really be spending a half million or a million? They have like million people spending millions of dollars on using this. What do you think about the pricing

SPEAKER_173: and your strategies there? Because I didn't see any outward facing pricing yet.

SPEAKER_06: We have a pricing page and it's, yeah, it's $7 per user per month after a free trial period.

SPEAKER_00: And the free trial right now is like super generous because we're just launching. So we're going to let, everyone who's coming on board now is going to have like three months free, but the long-term, you know, free trial period will be a month. In any event, seven bucks per user per month is slightly underpriced to Slack. I mean, obviously that's deliberate. Yeah. And then there'll be a separate plan for enterprises that want to scale this much

SPEAKER_58: bigger. Yeah. I think, um, the trap of Slack has been large groups for your non-corporates. It's just not worth paying for it. Then they let you use it, but you lose your archive and here your value

SPEAKER_86: is the archive and the power of it. And you have to charge a little bit because it's not cheap to use

SPEAKER_18: these LLMs. Am I correct? Or is it, is it going to be expensive for you to keep indexing this? And if people start using it, or is it the pricing now of these LLMs and inference down enough that

SPEAKER_51: it's not that expensive? There's definitely included costs there. We, you know, we were thinking through, Hey, how should we price this in comparison to say using chat GT, right? This is cheaper. Um, what we realized is we just want to include some blue AI usage with every single seat. And I think that the costs are coming down quickly enough that, that we're going to be able to do that. Uh, but we also can, you can actually just plug your API key in if you have an open AI API key or a anthropic API key. And then of course, you know, lift your, your limits that way. We're thinking of other potential models to be able to give people more LLM usage, but we just want to make everything work out of the box for that base price. Whereas I believe you need an enterprise Slack plan in order to get Slack GPT. So a lot of people don't even know how it works or have used it.

SPEAKER_176: I've never used it. Never heard of it. Um, right. And the enterprise is that starts to get absurdly

SPEAKER_58: expensive. Now you're hundreds of dollars per user per month. I think it's a non-starter for a lot of

SPEAKER_182: folks. So what about windows?

SPEAKER_00: But it's just something important there, which is if people abuse it, like for example, let's say you're a big company and you're trying to use glue as your workaround to not pay for chat GPT. Yeah. You're going to hit a limit. You have to put in your API key. So the bottom line is you'll be able to use whatever LLM model you want, whether it's anthropic chat GPT, et cetera, et cetera. And then you just give us the API key and then you'll be able to access it through glue. Uh, on top of that, there'll be some basic amount of free usage or let's say included in the basic plan so that everybody

SPEAKER_186: gets access to glue AI right out of the box. Okay. I am passionate about innovation and tech, but I love hearing from the founders themselves. And now I've got an exciting opportunity for you.

SPEAKER_82: I'm hosting a jam session with J Cal it's powered by my friends over at dot tech domains. And it's over the summer. I'm going to have five founders get on a call with me and I'm going to give them my unfiltered feedback. And the best part is these mentorship sessions, these jam sessions, where we go back and forth and talk about your customers, your product, your team, your problems, your victories, the things that are going well, the things that aren't going well in these jam sessions, we're going to take them and we're going to publish them on this week in startups. Think about how great that is. You're going to get to have a jam session with me. That's awesome and fun and educational and hopefully helpful. And you get to share it with the world. So all you have to do, if you want to be part of this insanity, this awesome, beautiful insanity is you need to have less than 2 million funding. We want to do this for startups. And you got to have an awesome dot tech domain name. So apply and get more details jam with J Cal dot tech jam with J Cal dot tech. It's in the show notes. And I'm partnering with my friends over dot tech domains because all these great innovative startups are starting to use dot tech. I see it all over the place rabbit dot tech, Aurora dot tech, one X dot tech. And of course, founder Fridays dot tech. If you're using a dot domain name for your website, I want to hear about it. Apply for the jam session with J Cal today, jam with J Cal dot tech jam with J Cal dot tech jam with J Cal dot tech.

SPEAKER_86: Yeah, I think people have a lot of privacy concerns or training other people's models on what's going on in their business. How do you think about that sex?

SPEAKER_06: Well, we don't use your company data to train our model. We're not training models. I don't,

SPEAKER_00: it's like had that, uh, kerfuffle just a few days ago where people are tweeting about saying, why is Slack using our company's data to train their model? And you had to email the company, Salesforce to opt out of that. They didn't give you a setting for that. People were complaining about it. We're not training models. I mean, the model training is being done by open AI. It's being done by anthropic. We're using the best foundation models and that list is going to keep growing. And there are things we do to optimize your query so that we can give you really high

SPEAKER_02: quality answers, but model training is not one of them. So we don't need company data for that purpose. I'm kind of curious what they're using it for.

SPEAKER_18: I would suspect that if you were ambitious about this, um, and I know you are that there's going to be the end game here is you're studying roles in organizations. That is extremely valuable over time. If you believe that certain roles could be automated, I think watching people work, some amount of their work could be automated. So that's, I think eventually where people's mind

SPEAKER_27: would go to, and that could be super valuable for the company, your company, or the individual

SPEAKER_00: company. Yeah. I think it's an interesting thought. I mean, it's not where our heads are at today, but I think we could get there for sure. Um, it's just, I think our view right now of, uh, AI models or model training is that there's so much innovation happening right now at the model layer among all these foundation companies that are literally spending billions of dollars to build out the infrastructure, to create better, faster, cheaper models. Why would we ever try to compete with that in any way?

SPEAKER_02: We just want to ride that wave, take advantage of it. And, um, you don't need to lay your own fiber

SPEAKER_196: or build your own data centers. It's not a competitive advantage. Totally. And look how good

SPEAKER_02: the results are already for glue AI. It can tell you who has expertise in your company. It can tell you what people are working on, what their priorities are, how people are feeling in the company. It's crazy. The kinds of results we can get. And by the way, even though we're using chat GPT 4.0 for those, um, prompts, you can't get those answers. If you go to chat GPT, even if you're on there, you know, teams or enterprise plan, because it doesn't have the context. So again,

SPEAKER_06: that's that magic of connecting the chat history with, uh, with the AI model.

SPEAKER_18: Maybe we talk a little bit about how building a company 14 years later, after building Yammer has changed the amount of resources, the place, the product velocity, how has all that changed in

SPEAKER_06: the last 14 years? I mean, everything's faster, right? So, I mean, how many people do we have on the glue team? The second eight, Evan? Yeah. Seven. That's wild. Yeah. We would have needed

SPEAKER_00: 20, 25 people. Now we're just at the point now where, because we're going to market, we are going to need to do some hiring. So if people out there are interested, we're going to be hiring across the board because we do need to kind of race and seize this opportunity. So certainly developers, um, we're going to need a lot more of them, but we're also going to start with, you know, we're going to start adding some marketing and sales roles, things like that. And that comes with operational roles as well, customer support. So yeah, we're going to start scaling things up

SPEAKER_02: quite a bit, but yeah, we've had, you know, an eight person R and D team has been building this

SPEAKER_150: out. Incredible. When you think about that, like eight person teams building like fully polished,

SPEAKER_127: really good, stable software and the, and the velocity is only going up. Uh, people have been

SPEAKER_02: asking my one thing, you know, we haven't worked on this for a while. So, you know, there are products

SPEAKER_00: that you just, you know, ship in six months. And, you know, it's an old saying about if you're not embarrassed by the MVP, then you've done too much. That was actually not our view. Uh, our view is that

SPEAKER_02: we cannot be embarrassed by the launch version because the table stakes are really high in the space. I mean, like we said, people have, I think channel fatigue, but they still think slack on the whole is a good product. So the bar is not low. The bar is pretty high. Same thing, even for Microsoft teams, you know, it's a Microsoft product. There are issues. However, it's still a pretty good capable product. So again, the table stakes in the space are pretty high. So we knew we couldn't come out of the gate with just kind of the MVP. So we have been working on this for a long time and tinkering to get it exactly right.

SPEAKER_86: It makes sense if you are, you know, starting a new category and, you know, you come out with a

SPEAKER_58: micro blogging service 15 years ago. Yeah, sure. It can be a little clunky and have some fail wells, but today's day and age, the, the consumer's expectation is pretty high, huh? Evan?

SPEAKER_51: Yeah. Yeah, absolutely. I mean, I've, I've seen similar things, you know, companies like Figma, companies like linear, that's kind of blowing up right now. Those are very mature spaces with really good products in them. And so the, the bar for what you need in order to attract users is, is pretty high. So yeah, I think it's a growing sentiment amongst founders that just over time, even in general, across the, the, the software industry, user expectations are just continually

SPEAKER_66: rising. So yeah, definitely. How do you guys look at, um, building in person versus remote teams and the pluses and minus of those? There's a, there's a thing going around Silicon Valley. Like somebody

SPEAKER_86: said to me, I would invest in a venture firm if they only invested in people who were back to office.

SPEAKER_218: So the mandate of the venture firm was to find startups in an office because they felt it was

SPEAKER_24: such a compelling advantage. Um, what, what do you think? I mean, you're building a tool that helps remote teams, obviously, but what, what are your thoughts on that in 2024, people being in the same

SPEAKER_86: space, cohabitating and building products versus remote and being able to hire anybody anywhere in

SPEAKER_98: the world? Lots of trade-offs. Uh, you know, I gotta say I, I managed a hybrid team at Clearbit, uh,

SPEAKER_51: previously, and that was prior to, you know, any of the, you know, work from home period that we went through, but you know, we had an office, it was hybrid. And when I went into the office, the energy was so infectious and the amount of sort of focus and the, the way communication load is just something that I, I personally feel like you, you have a hard time getting remotely. So we have some remote team members right now, uh, at glue. They're incredible. We're also planning to open an office, um, very soon as we, as we continue to grow. So I think we're trying to be pragmatic about it, but honestly, you know, having that face-to-face interaction, especially for certain teams, um, we found

SPEAKER_86: is really crucial. Product sales team sacks, you know, you probably want to get them and get that infectious energy going. Where do you think the industry is heading? Do you see something in your portfolio or your intuition tell you where this is all going to wind up and who has the advantage?

SPEAKER_224: Look, I'm a fan of, uh, of having an office. When I was running companies full-time, I used to manage

SPEAKER_00: by walking around and you'd kind of run into people, ask them what they're working on. They would tell you that five minute conversation by the water cooler or getting a snack or something that can make a huge difference. Um, I would have developers tell me all the time, I'm working on X, Y, Z. And I'd ask them, why are you doing that? And they'd explain, and somehow that someone had gone down the wrong path. And I was like, you don't need to do that. That's not even important. You know, you'd solve problems that way, uh, when you're a founder. And so there's a lot of benefits. There's also the, the serendipity of other people getting together, the information sharing,

SPEAKER_02: having product and sales near each other, like overhearing each other's conversations to some

SPEAKER_00: extent. So I am a fan of having an office at the same time, even back, you know, a long time ago when we were doing hammer, we did have a remote, a remote sales and service office in London to service

SPEAKER_02: Europe. So just realistically, you need to have some field offices also, uh, dev dev offices. I mean,

SPEAKER_00: it was never possible to hire all the developers you needed in the Bay area. And so dev went hybrid again, even before the whole COVID work from home thing. So like Evan said, I think it's worth being pragmatic about it, but I guess the one model I don't believe in is the fully atomized where you've got a hundred people in a hundred different places, you know, basically all working from home. I think that is really hard to manage. And it only gets harder to manage as you scale up to 500 employees. Now, look, there's a couple of companies that I guess have done it. And there are the names that everyone always cites, but there's a lot more companies you've never heard of that were unable to pull that off. Yeah. And this is just, in my view, not something that needs to be reinvented, like offices work for a good reason. So I, I liked the idea of having an office. And then I also liked the idea of having making room and accommodation for like amazing developers who we just couldn't hire in the Bay area. Yeah. There's somebody with the unique skillset. There's some 10 X

SPEAKER_191: developer. You can make an exception if they're only going to live in Lake Tahoe or something, and you just don't, they already live there, you know, they already live in Europe or whatever.

SPEAKER_00: Yeah. Right. And so development's always been like that. You know, when I think about sales and

SPEAKER_02: service centers, especially if your sales is more telesales, as opposed to field sales, I think it's like nutty not to have that in one location. People tend not to do it in the Bay area

SPEAKER_172: because it's too expensive, but you know, they open them in Denver or Arizona Lake or yeah, exactly.

SPEAKER_124: Somewhere where a home does not cost $4 million. Like if you have a home that costs 500,000 and a

SPEAKER_191: person who's a sales executive could actually afford a mortgage. Yeah. That makes more sense.

SPEAKER_00: You know, my point is just, there are definitely better and cheaper places to do it in the Bay area, but I think it's crazy not to have those people in one place because then you can put all

SPEAKER_02: the training resources in that place. You can have apprenticing. So a new employee starts, they just shadow your best, most experienced employees. It just, it's so much easier. So I don't get the

SPEAKER_00: idea of having, like, if you've got a hundred customer support reps, I don't understand how you can, I'm working from home. It just makes no sense to me.

SPEAKER_124: The apprenticeship is something I keep coming back to. I'm just wondering what's going to

SPEAKER_27: happen to this next generation. We saw like our kids who miss school, you know, young people who miss college and then young people who are not getting the experience of mentorship in an office. It's something's very much lost in their careers. Also, I think it makes people weird and lonely and a lot of this mental illness, anxiety, all this crazy stuff people are having, I think it's because they're not socializing enough. The office kind of creates that joie de vie, that

SPEAKER_210: culture creates a lot of culture. You need it. I, you know, again, I don't know that you have to

SPEAKER_00: have one office, but having hubs. Again, you know, when I was doing Yammer, we had the hub in the Bay Area. That was our main office. We had a office in London. We had a small office in New York. We actually had a small office in, I think it was Sydney in Australia. So, because we were doing sales

SPEAKER_02: there and they're on different time zones, but in any event, clustering employees so they can feed off each other's energy. I think like you're saying is really important. I honestly think looking at how

SPEAKER_18: Europeans live and how now Chamath lives as a European or bestie, like maybe two weeks vacation is not enough for people to feel like they have a life. So there's like, I think this like Fridays off or, you know, whatever remote Fridays, it feels like there's some compromise between, you know, ownership and

SPEAKER_24: employees where people can still, you know, go skiing a little bit or take a more than a week off, you know, maybe

SPEAKER_00: work remote half time for three weeks. Well, if you think of, if you think of remote as a benefit to the employee, I think it's fine to think of it that way is that it is a benefit to the employee. And sometimes you can get a much more senior employee than you otherwise could hire because you're willing to give that benefit, but let's not pretend like it's a benefit to the company. And I think that's, that's the way a lot of companies act is this is like somehow some wonderful new management technique. Yeah. I think it just makes their lives harder. Can it be a competitive advantage for people

SPEAKER_66: who choose to have the discipline of having the team in an office?

SPEAKER_00: Yeah. And like you said, the, the younger employees who need mentorship, who need apprenticing, who just need to shadow someone in the first month of the job so they can learn the job.

SPEAKER_02: Yeah. Way harder to do when everyone's remote, when you're in an office, much easier to do. And again, it's different when you graduate to a place in your career where you're very experienced, you know, the job, you know, again, maybe you can negotiate to be remote

SPEAKER_00: because your value is so high, you know, again, companies need to just need to take that into account when you've got a more kind of entry level employees wanting to go to an office and one who

SPEAKER_251: into something remote. I mean, there's a big difference there and which one's going to add

SPEAKER_86: more value in my opinion, who's going to learn faster. You know, who's going to contribute more. I think that's what it comes down to. All right. Listen, everybody go to glue.ai,

SPEAKER_18: sign up for the wait list. Please don't ask me to try to push you up the wait list. They'll get

SPEAKER_127: through it as quick as they can. I got enough people trying to, you know, ask me for favors. Uh, and, uh, hopefully we can, uh, invest in the company again, wet my beak.

SPEAKER_84: You will be able to wet your beak. Yes, I need to. Um, Evan, congratulations. I know you put a lot of hard work into it. So, uh, congrats. Um, and, and on that huge wait list, I think it's absolutely fantastic. And we'll see you all next time on this week in startup. Bye bye.