A Conversation With Jacob Wenger of Shortwave.com

Missing the big wave versus catching it can be a matter of inches and an extra push to make or break the ride of a lifetime. Shortwave co-founder Jacob Wenger may not be a surfer, but he and his colleagues look to have caught the big wave after years of hard paddling.

The wave in this case is the "AI moment" embodied in the ChatGPT buzz; Shortwave's particular ride is putting AI to work on that tough, tough problem called email.

Shortwave's tag line is "intelligent email, powered by AI." The Shortwave app (mobile and desktop) looks like an email client, but acts more like an email assistant with (as Jacob says) some opinions of its own. Among the assists — intelligent, automated handling of sets of messages; message summaries and 'magic tags' to tell that smart assistant how you want things handled.

Jacob shared the driving vision behind Shortwave -- creating a common communication platform not owned by anyone. He also delved into some of the pragmatic decisions involved in moving forward. Why build on top of Gmail? Why not become an inbox provider?

As Jacob detailed in this engaging conversation, email is simultaneously one of the best places to apply AI, and one of the hardest. It's probably the most personal data-collection for each of us — but it's built on standards that are older than Jacob!

If you want a glimpse of the future, grab the Shortwave email client (Shortwave.com and mobile). Ask Shortwave to summarize a long message thread and go...Wow.... Check it out at Shortwave.com!

TRANSCRIPT

A Conversation With Jacob Wenger of Shortwave.com

===

Matthew Dunn: Good morning, Dr. Matthew Dun here, host of the future of email, my guest today I'm very excited to speak with Jacob Wenger, co-founder, head of product at Shortwave. Jacob,

Jacob Wenger: welcome. Hey Matthew, thanks for having me. I'm, I'm happy to be here and looking forward

Matthew Dunn: to our discussion. Oh man. And Shortwave, I, I'll let you introduce it cuz I'm so excited.

I probably won't get.

Jacob Wenger: Awesome. Cool. Yeah. So Shortwave is an AI email app designed for stress-free productivity. You sign in with your Gmail or Google Workspace account and we give you a brand new interface for managing your email. Our apps are designed based on an opinionated approach that we call the Shortwave method.

And it kind of have, has two parts. The first is a like flexible triage system that allows you to treat your inbox like the to-do list that it really is. You can pin items, you [00:01:00] can snooze them you can rearrange them using drag and drop and combine them and gives you a lot of flexibility to do that.

The second part is a highly customizable and composable set of rules that allow you to customize your workflow. So you can do things like block senders. You can group together related emails into bundles. You can schedule the delivery of threads for specific times. And the end result here is you get a, an email app that allows you to get through your inbox quicker and with less stress.

And then we've layered on a bunch of the latest AI tech to really bring things to the next level. The coolest is probably the a summarized feature that allows you to capture the key points from a set of threads. So there's a lot more to it, but that hopefully gives a kind of overview of the kind of things that.

So,

Matthew Dunn: and, and I have to hand you the comment right out of the gate. I've been using Shortwave for I think a month now, and it, it, it, it's jaw dropping. Like, like [00:02:00] wow. And I've used a lot of email clients. So the fact that I keep coming back short wave, like that's really something, like one of the things I love.

I jump into the inbox from hell in the morning and there's all the promotion stuff clustered together, and I could basically go, pow, shut up, Legette, go away. Yes, yes, exactly. It's huge time saver right there. Yeah,

Jacob Wenger: the, the bundling feature and being able to group related emails and then just like mark them all done in one click is, is, is great.

And it also just makes your inbox look a lot smaller than it actually

Matthew Dunn: is. Right, right. But without, without losing stuff. I've tried some other email clients that maybe were opinionated in, in different ways and, and had their strength to be really fair. But when I end up feeling up fighting, fighting the client, or I'm fearful that it's hiding something that I really do need.

Yes. Then, then it's like sorry. This relationship isn't gonna work out here. One of the things I have to ask you [00:03:00] right out of the gate you, you and your team, not a huge team, what does it look like? A huge team from the product. You guys have been working at this for a number of years. And here comes the media called Chad, g p t.

Putting AI at front and center of every discussion from coffee to cocktail party. Good thing, bad thing for you guys.

Jacob Wenger: I, I, I think it honestly is, is a great thing for us. Yeah. I think we, we are actually really well positioned to take advantage of, of this moment. So I think chat, G p t and the, the, the generally the, the wave and ai mm-hmm.

That is based on all this text, I think is email is a perfect fit for it because your email archive. Probably the largest set of writing and knowledge base of any app that you use. And so as an email client, we are able to use all of this data and surface it in really interesting ways. Now using this new technology and anyone who is, who's going to start an email app who says, oh wow, this is the right [00:04:00] technology, I'm gonna start building an email app.

What they're gonna find is it's gonna take several years to get to the point where we are. And this was something we learned the hard way. Even though we knew looking at other competitors who had tried this before, it took them three, four, or five years to get to market. There is just a such a wide surface area for email applications and the bar is so high.

That it just takes a long time to actually build the fundamentals and we feel like we have a really solid foundation and now can layer on the new AI stuff and really take things to the next level. So

Matthew Dunn: un, un, un unpack some of the pieces of that. I'm, I'm, I'm not, Surprised by what you said, but I play in, in somewhat the same sandbox and have some, some years in email.

I know what a complicated sucker is. Why is it so? Why is it so hard? Why so many years? Just to get to start.

Jacob Wenger: Yeah, I, I, I think their email is one of those apps where lots of, lots of applications, email clients, like Gmail as an example [00:05:00] take the approach of we're gonna give you a bunch of features and then you can build your own workflow on top of that.

And you see this in the way that the app is designed, you know? Yeah. They have six different versions of the inbox and they have a ton of different label features and combining those in different ways. Everyone ends up with their own ad hoc system that they develop themselves. And so if you want to get people to move to your application, you can't just support 20% of the features and assume that 80% of users can come and use your app.

You have to build this like. Crazy patchwork of features so that you can support everyone's custom workflow. Mm-hmm. And we are trying to build an actual workflow and, you know, we, I mentioned the Shortwave method. Try and have an opinionated way to get through your inbox that we think is best. Yeah. And as opposed to just building a bunch of features, but we still need all of the features that you have in these other emails.

Right,

Matthew Dunn: right, right. So you've gotta build sort of, you gotta build two, two products, so to speak, or two architectures and then, [00:06:00] and then get them. Merge in a usable, you know, usable, comprehensible set of Yes, exactly. Set of

Jacob Wenger: control. There's one other piece of this, which is email is a, is a very old protocol and there are a lot of dragons hidden in just like taking an email and displaying it.

Well, yes. And it took a a lot of effort from us and in the early days I would go to like, open up my inbox. I'd get through two emails, then I'd see an email that was just busted, an HTML email that maybe displayed poorly, log a bug, and then get back to my inbox two emails later. Another thing that's screwed up.

Someone's signature isn't showing up. Yeah, yeah. And just getting to a point where you can reliably see emails in a way that displays 'em as nicely as Gmail. Yeah. It takes a lot of effort. Yeah.

Matthew Dunn: Yeah. Okay. Well, you, you, you, you stole my thumb. I was gonna, I was gonna say grafting on top is such a relatively old and kind of patch's.

Architecture. Ah, right. Yes. There are key [00:07:00] protocols involved in email that are well over 20 years old. Yes. And you had to work with

Jacob Wenger: them. Yes, for sure. And I think there's no way around them. Yeah, exactly. There's no way around them. And I think the, the backwards compatibility and being able to work for any email, any email service out there is like a critical aspect of, you know, the federated nature of email and is, is really important towards like our long-term mission to be able to create like a common communication platform that is decentralized and secure.

And not owned by any central organization.

Matthew Dunn: Oh, bless you, son. I love that. I love to hear that. So I'm gonna go, I'm gonna go back in time and embarrassingly long ways to tee up part of this conversation. I worked on email stuff at, actually, we, we share a company, a common Microsoft. Oh, okay. But in my case, it was the nineties, early, early days, like land, email, pre.

And was involved at a few different points. Companies changing email [00:08:00] systems and the emotional attachment that people have to their email even back then was unbelievable. Like the level of upset when you change someone's inbox experience or what it looks like when they're start, you're like, dang, this is like, this is counseling, not technology.

It's, it's surprisingly person. For for sure. And you had to work with it, right?

Jacob Wenger: Yes, definitely. I, I, I notice this talking to users every day. And it is, I, I think it, it makes sense to me in the sense that there are people who spend hours and hours of their day, both in their work life and in their personal life.

On email. Yeah. And it's not surprising me that people have such a strong attachment to this app that they use and is so critical to communicating with people and getting their work done. Yeah.

Matthew Dunn: Yeah. Well you, you look at, you look at look at the rise of mobile and I would argue the first killer smartphone was probably the blackberry.

Much as I like, you know, like the iPhone [00:09:00] sitting on my desk here now. But what was the killer after the blackberry? No question. Right. It's like everyone's like, oh, thank God I don't have to log this big heavy laptop around and, and dial up in the hotel room. He mean I can get my email in my pocket. Like it was, it was, that was transformative.

And I'm sure one of the top X number most frequently opened apps on smartphones now still has to be email with, with, I'm sure his text as an, as an interesting vector on that. Yeah. It, it is, it is one of the things that's. As you said, everyone makes their own and ends up making their own UI architecture in Gmail.

Why? It's like, it's mine. It's my, you know, my email that possessive word comes up all the time and all the decisions about what you keep or don't what you, you know what you folder here or don't like, it's all on you, right? It's all personal, so it is bold of you. To have an opinion.

Jacob Wenger: Yeah. Yeah. I, I think like you, [00:10:00] you need to have an opinion because you need to stand out.

It is it is a very crowded space and lots of, lots of startups have come and gone before us with varying levels of success. And if you don't have an opinion, then no one's probably gonna care. Yeah,

Matthew Dunn: well, you're, you're right. Delve into that for a second, cuz there are the, the road is kind of littered with the carcasses of email clients, isn't it?

Yep, that is,

Jacob Wenger: that is very true. Why, why, why go after? Why go after email?

Matthew Dunn: No, I mean, one, why, why is it such a hard space? And then two, are you crazy whyt you go after email.

Jacob Wenger: Yeah. So I'll, I'll, I'll answer them in maybe the, the opposite order. So, so why email is we, we started the company actually with, with the mission that I, that I mentioned earlier around creating this common communication platform that's decentralized and secure.

Mm-hmm. And we've, we've been concerned about, Like the centralization of our communications, whether that is in the work life, things like Slack or [00:11:00] Microsoft teams or on our personal life, things like iMessage or Snapchat or even TikTok. And all of these are owned by either large corporations or in the case of something like WeChat, a government.

And we think that this is generally bad for society and that if we continue to have all of our private communication, In these centralized services that at some point we are really going to regret this. And so we kind of surveyed the landscape and said, if we wanted to build a replacement for all of these Yeah.

What technology should we use? Yep. And the thing that kept coming up was email and email is already this large, decentralized, federated communication platform that's not owned by any one company. The problem is, is just the user experience is not very good, and there's a reasons why people are using these messaging apps.

It's because their user experience is just so much nicer. They have typing indicators and emoji reactions and real-time messaging, and so we thought, It is easier for us to take [00:12:00] email and move it towards being a really good messenger than it is for us to bootstrap an entire new network that is as big as email.

And so that's how we kind of ended up going after email. Okay. Okay. I'm with you. And then as far as why it is a, a hard space I think it's, it's always a hard space when you are competing against some of the biggest companies in the world. That's cool. Google and Apple both have really high quality email applications.

Like they are. They are not perfect for everyone, but for the average person they do a pretty good job and they are fast and reliable. And so the bar here is quite high. And then the other piece here is, as I mentioned earlier, there's just a large surface area and it takes a large, strong technical team a long time just to get to the foundational.

Matthew Dunn: Yeah. Yeah, that, that, that one certainly makes sense. I mean, you already, you already touched on some of the, I mean, almost trivial, but not trivial front end problems. Like, oh, this email is broken. It doesn't render. You gotta fix that. If, if someone's gonna keep [00:13:00] using the client, cuz they'll, they'll throw you out in a heartbeat.

And you're also, I mean, you touched on it, but you're. You're competing with big companies, giving it away. Not a trivial lift. Not a trivial lift. Go back to I love what, I love everything you just said about, about the why of starting this, cuz I've said much the same things myself about why I, I really dig the email space.

Like no one owns it. That's a big deal. It's better for all of us than no one that no one owns it. But you, you had to make. Judgment call against that. And pick an inbox to start with. And you picked Gmail. Yes. That had to have been some long discussions.

Jacob Wenger: It it, it definitely did. And this was a, a decision we made very early on at the company.

Yeah. One, one option we thought about was being our, our own provider and going the route of something like, Hey and this has, this has its benefits. It does have, allows us to kind of more [00:14:00] control the, the privacy and security. We're not relying on another company to provide some of our service.

But it also has a downside, which is it's really hard for people to get switch email clients. It's even harder for them to switch email accounts. Yes, yes. And completely start from scratch. Yes. And with a brand new account and not be able to bring their existing email. Is something that we thought would just make it too hard to kind of get the company bootstrapped.

Yeah. And so we decided to, to build off of Gmail long term we wanna support a lot of other providers and Oh, really? Still interesting. Be our own provider. Allow you to self-host it like you can, something like Microsoft Exchange. Mm-hmm. But those are kind of years out still for

Matthew Dunn: us. Yeah. Yeah. Well, let's, let's delve into the Gmail thing a bit and one of these days.

Someone from a Gmail team on here, because I've been kicking them in the shin so hard that maybe they'll pay attention. Like I have a couple different G Suites accounts. I use that client as well. It's not that I'm [00:15:00] not a fan on a bunch of levels, but something like 60 plus percent of inboxes are with one company and it bugs the bugs, the jeers out of.

To see that for the reasons that you already cited. It's like, wait a minute, hang on a second. Too much centralization is bothersome to see and yes, totally agree. Yeah. And, and to Gmail's, to Google's I dunno if it's to their credit or there's a longer game I'm not seeing, which is probably the case as well, but the touch on email.

In the, what, 15 years or so? The Gmail's been around lighter than everyone expected at first. Like there were concerns about. You know, blue links at the top of your email that you didn't ask for hasn't happened, right? Unwanted ads showing up inside the message body hasn't happened. They've played a pretty straight up game as far as standard support with AMP for email.

Longer discussion. Some people say they're [00:16:00] extending the standards. Some people say it's a bad idea, me among them. But they, they've, like, they're playing in the, in the email. They haven't sort taken the ball away despite having 66% of the field and said, it's our game now. But they do keep gaining market share of inboxes and I keep watching that going cause I worked at a monopolist.

You did too. Mm-hmm. Back when it really was a monopoly. And I don't like monopolies, even though I got, you know, got to benefit from being part of one. Like, ah, dang it. I think competition's a healthier thing. Like, where do you see it going? Where do you see the inbox control?

Jacob Wenger: Yeah, so let me, let me actually start a little a a little further back.

So, I think the, the reason Gmail was, was so innovative when they came out is the fact that they, they actually offered a bunch of really cool features with their initial version back in, I don't know, 2004 when they initially came out. You know, they gave you a larger archive so you could store a lot more.

They gave you the ability to [00:17:00] like thread messages together, which was like revolutionary at the time. Mm-hmm. And I can't even imagine using email without it now. Right. And they also gave just like a faster interface that was using the latest. JavaScript and web tech. And so they really came out with a bang and really impressed people.

And over time they just like have not really innovated from that space. And the email of 2004, it looks a little bit different today, but the fundamentals are still, fundamentals

Matthew Dunn: are still the same. Yes, that's true. That's very good.

Jacob Wenger: And the, the, the biggest thing they've done is, is introduce Google inbox you know five, you know, I guess longer than that, seven, eight years ago.

And then killed it a few years after that. Yeah. And that was really you know, them designing. Inbox from, if email was made from a mobile first perspective, and I think they got a lot of things right in there. They brought some of it to Gmail, but most of the fundamentals from that app like never really made it.

And I think a lot of people have been missing it since this. Yes, yes. Yes. And the, the reason I think they've continued to grow is [00:18:00] that like the Google Suite is actually continuing. Better, you know, the introduction of docs brought a lot of people into Google Workspace, which is what's called now.

And it happens to be that all these things are bundled together and as we've seen in the tech industry, like bundling things together is a great way to get adoption across a wide set of products. Mm-hmm. And so I think Google has like done a good job of. Building up their suite and getting users to kind of adopt a bunch of features.

Even if they came for docs, they end up using Gmail as well. Mm-hmm. Where, where I think things are going and part of the reason why we, we are actually building on top of Gmail is, is distribution. I think that in order for, in order for us to really like, push the protocol level changes we wanna make to email and bring us to that world where it is this like decentralized, secure protocol we need to have market.

And no one's going to care if we come out with this self-hosted application if no one knows about us. Right. And it doesn't work with their current email. Yeah. And so our strategy is we want [00:19:00] to try and make as big a splash as we can, try and gain as much market share, gain as many people using our application, and then use that size to actually drive innovation in the space.

Otherwise, we're just gonna kind of be crowded out by these larger

Matthew Dunn: players. Yeah. Yeah. Wow. Both, both, both of you to actually. Say it so clearly. And I think there's a whole bunch of, you're more qualified than I am to talk about it, but I think there's a whole bunch of sense strategically to what you just laid out.

And I actually think this Sputnik moment of Chad g p t is really gonna play to your favor

Jacob Wenger: because I think that this is, this is a moment. Sorry to, sorry to interrupt. This, this is a moment where I think. Foundational apps that we use on a daily basis. Things like email and notes and to-do apps are all going to fundamentally change with this new technology.

Yeah, and like mobile, I think it has the opportunity to like unseat the [00:20:00] existing players. Yeah.

Matthew Dunn: Yeah. Un, un, un unseat and, and existing players with a notable exception at, on, at this particular moment. Existing players don't typically. As fast as startups. The Astro is there. You worked there? I worked there.

I gotta take my hat off to the boys in Redmond. Holy free holies. Are they moving fast on the AI stuff for

Jacob Wenger: for sure.

Matthew Dunn: Wow. You know? Yeah. Like, wait a minute. You have what? All across office. This just happened yesterday. How'd you pull that off? Like, you know, and I mean, investment, pre-planning, a whole bunch of other stuff.

Smart company, great leadership, like hats off to 'em. It is making Google look flat footed, in my humble opinion. Yes. Gmail, as you said,

Jacob Wenger: looks just like it did. Yeah, yeah. No, Microsoft has, has moved quicker than, than I expected and I think is delivering really interesting and innovative stuff. I wouldn't count Google out just yet.

No, no, no. Not at all. Not at all. [00:21:00] A lot of expertise here, but yeah, it definitely feels like they are, they're behind at this point. Well, and

Matthew Dunn: one of the things they, it does it temporarily and, and I suspect there's probably some gnashing of teeth, cuz I know Google has a standing investment of, you know, decade plus in ai.

A lot of the things. Gmail do the magic it does deserves to label ai. But that's not the, you know, they didn't, they didn't get the public label, so they're probably going, we were doing this already, right? Yeah. And I wouldn't, I wouldn't count 'em out. Very, very, very smart very long term thinking company.

But you already said it. Essentially they have an installed base, and that's a mighty big tale to drag around when you're trying to. Like you, you, you, you tweak a pixel on Gmail and a billion people are gonna squeal at you. You can't just try stuff. It's a scale problem too. Like I don't, I'd love to know the cost of running that bad book.

Jacob Wenger: Yep, yep. For [00:22:00] sure. Cost. And I think also just The, the, the difficulty in getting a billion and a half people to switch to a new paradigm. Yeah, yeah. And I, I, we have a, a much smaller user base and making fundamental changes to the app can sometimes cause a lot of people to be very upset and tell you how they feel.

And you kind of have to push through that. That's a lot easier when it is a couple thousand people who are complaining versus a couple hundred of million people who are Yeah,

Matthew Dunn: yeah, yeah. Absolutely. At the scale of the company you've got now, you could sit down and have a conversation, make a decision, and go.

It doesn't take a committee, cor Correct. Which is a, it's a big, big deal. I, I wanna take this in a slightly different direction. Just stay on the AI sprain for a second. Put Microsoft in the, like, agility that they're showing aside for a second. Some aspects of the most visible AI [00:23:00] models kind of inherently require scale and, and investment, right?

Training Chad, g p t was a billion dollars or something like that. Like, I'm not gonna do it. You're not gonna do it. But a day later, a month later, a year later, You get to build on top of it or leverage it, which is kind of amazing and kind of magical. And I, I wonder how the, the ones that are spending a billion to train the models are gonna leverage that and, and, and recoup it without someone, you know, based on open source or based on, you know, being smarter and more agile at 11 than than a hundred.

It's just gonna out-maneuver 'em like a PT boat around a battleship. What do you

Jacob Wenger: think? Yeah, I I think likely, whatever I will say will be outdated in a month, given how quickly things are moving. But I kidding. I'll give you where my, my mind is now. Yeah. I, I do think you'll end up seeing multiple layers of providers.

So you'll have people [00:24:00] like Open ai or Philanthropic or Google who are building their own large language models and offering. A, an API that gives you this like, amazing technology that understands reasoning and language and all of these really cool things. Then you'll see a, a second layer where you have specific companies who are training those base models for specific use cases.

Yeah. You'll have one that is really good at being a lawyer, one that is really good at being a doctor. Yeah. One that maybe is really good at writing emails. Mm-hmm. And then below that you'll have a third layer, which. End applications that will either be using the, the second layer customized ones mm-hmm.

Or the top layer based models to build app functionality. And right now Shortwave is kind of in that third layer. Yeah. We are using open AI's APIs to build a handful of features. And you may see us move into more of the second. Where we are actually offering other email clients a app like services that [00:25:00] kind of allow them to do things that are a lot easier than they could do with the base layers.

Matthew Dunn: Nicely put, nicely put that that makes a lot of sort of tech evolutionary sense and it, it, that structure you described is cognizant of. Enabling layer of API-driven development that frankly we didn't, we didn't have 12, 15 years ago, like Salesforce just to pull something outta thin air, right?

Salesforce trying to pioneer you know, cloud-based SaaS stuff. We're starting this from scratch or, or doing it with, with nasty old IOP protocols that weren't nearly as easy as APIs. And now that's kinda like, that's how the world drives. And, and open AI is obviously playing to that already. Yeah. Well, well put, I like another

Jacob Wenger: description.

Another, another area where you've, you've seen this is in like cloud computing. You have a bunch of the, the large tech companies like [00:26:00] Google and Amazon who offer. Like raw compute. Sure. And then you have the, a layer of companies who offer kind of infrastructure as a service. So you have things like Firebase where, where I used to work that is kind of like abstracting away a lot of the like, low level fundamentals of cloud computing and giving you really simple building blocks for building applications.

Yeah. And then you have the third layer, which is apps who are using Firebase to actually build their applications.

Matthew Dunn: Yeah. Right, right, right. And then finding the, finding the various. Business models at those layers is, is part of the trick shot. One of the things that I know Microsoft's has gotten complimented on by Ben, I think Ben Thompson and Stratec among others, is, you know, investing in open AI w was a good investment in, in Azure, right?

Because now no explosive amount of you know, activity at Azure data centers from the company that they helped. Sort of kickstart [00:27:00] and, and kick over at that, at that base layer. And you can only have so many winners at at that level. Zero of what? Of the stack you just described, right? There's the, yes.

There's only gonna

Jacob Wenger: be so many. Yeah, I, I hope there is, there's not one. I hope there are, there are multiple I think the, the AI and how good they can get is, is unclear and unclear if there's going to be one. AI that is, is, is, is just significantly better and unable to catch up to that. So good. But my hope is that there is competition at, at every layer there.

I think that's good for, for consumers.

Matthew Dunn: It is, it is. And I mean, we could, we could hang out on this Zoom call for many hours, try to unpack and speculate frankly for sure, on, for sure on where the AI wave is going. Tech side of that aside, it is a fascinating cultural thing to watch the range of responses to the Sputnik moment of Chad, g p t, everything from, oh my God, I'm gonna lose my job to.[00:28:00]

Dang, this is exciting. Let's go build stuff, right? It, it really spans the spectrum in a, I haven't seen this kind of energy honestly, since 95, 96. Everyone realizing the internet was not a fad, right? No. No. Ai, not a fad, and. It's like pieces of this have been part of our life for a long time. You started building an AI enabled email client bef way before Sputnik showed up.

Yep. And okay. At the same time, the, that cultural moment means everything in terms of public focus and recognition. What do you make of the call to, to call HALT for six?

Jacob Wenger: I, I, I don't think there's any stopping this technology. I, I don't know how you'd possibly enforce, enforce a hold like that. And I think that we, we need to get comfortable with figuring out how we can coexist with the, this technology, just like any technology [00:29:00] advancement before us.

And I think like, yes, there's the, the, the chance of job replacement, and I'm sure that is going to happen, but there's. Lots of opportunity for people to use these tools to make them better at their current job and to learn new things. One of the things I've found super fascinating about just talking with chat G p t is how well I'm able to take a new concept that I wanna learn.

And use it as a way to kind of Yes. Tell me what I need to know. Yes. And even, even if you don't know, even if it doesn't always give you accurate information, right. It can help you understand the boundaries of a concept, and then you can go and dive deeper and go to first sources mm-hmm. To like actually learn the thing you wanna learn.

Matthew Dunn: Yeah. I, I like, I, I've got a, I've got a family of family of teachers, ed teachers and educators. And so I keep an eye on that. As well, and, and there's a pretty good degree of freak out about AI in the world of education, but as a user with a chat with two chat G p t tabs open here all the time, [00:30:00] like this is so cool, having an on-demand tutor with a very vast knowledge base and doesn't always get it perfect.

But I'm not asking to get it perfect. I'm saying like, help me, like, help me get this one thing and then I'll, I'll take it from there, and then I'll come back and ask you again. It's like, it's remarkable. It's just remarkable. I, I had a. I was faced with fighting with a, a BigQuery sequel query and I was like, how hell am I gonna do this?

I thought, oh, hey, ta G P T, how would you do blah, blah, blah. And it was like six, eight step iterative conversation. I'm typing short English sentences. I'm getting back fruity, well designed sequel queries that helped me understand how to get there. I actually shared that with some other friends who were CTOs.

I'm like, look at this, like this would've taken me four or five hours longer than, than getting some learning help. Didn't use it for production code. That wasn't the goal. The goal was, let me get my head around it. So I can actually do a really good job of, you know, using it and [00:31:00] taking it from there. Woo.

Jacob Wenger: It's remarkable. It is, it is. It is very impressive. I, I have used it for the same, same purposes trying to, to build up sequel queries. And the thing that has me really excited is, it's great to do this all in chat, g p T, but what if this can be done in the place where you're writing the sequel query?

Yes. Or where you're writing the email and like, yes, yes. Like talking to a chatbot is great, but really bringing this. Productionizing it and putting it in the applications we use is really where I think like the huge productivity gains are gonna come.

Matthew Dunn: Well, one of the, one of the one, I think one of the cultural impacts, and this, I'm trying to pivot this back to talk about short way more.

One of the cultural impacts is likely to be, look, if you, if you couldn't do some form of, let's call it software, anything from making an Excel spreadsheet on up. Okay. The growth of the digital world has to be frustrating because, I know what I want this box in front of me to do, but I don't know how to tell it to do that.

And now we're seeing glimmers [00:32:00] of just tell me what, you just tell the box what you want and it'll do what it always should have done. Like figure out how to do it right, like using your client. Yeah, very good. How's that

Jacob Wenger: for a pivot? Yeah, no, I totally agree. Like being able, we, we interface with everyone on a daily basis using natural language and now having the like capability to do this with our devices and have them like understand the intent.

And not have to be so prescribed in the way we interact with them. Yes, is, is really going to change how we interact with software. And that's as, as a product person. One of the things that like has me most excited about this new technology is how it is going to change the patterns and the ways in which we actually interact with our software.

I imagine we will be using voice a lot more. We will be writing in English way more than we do in in our current applications. And I think if you look. You know, different companies out there, like Notion, I think does a great job of like building an interface that is [00:33:00] novel and new and allows you to interface with the AI in a way that like doesn't require you to understand exactly how the AI is built or how to best communicate

Matthew Dunn: with it.

Right, right, right. I, you might have cut the news a day or two ago. Male Chimp. That, that they've now gotten a, an AI assistant to help write emails built right in the platform. Right. And mail champion owned by Intuit, so small to medium business client base. And I thought good on them. Right? Because the guy putting emails out at a small, medium sized business has a zillion other things to do If he can get that out the door faster and grammatically more correct.

Cool. Is it gonna write as well as if he sat down and wrote it? Mm. That's a different question.

Jacob Wenger: Yeah. I, I, I think that what you'll end up seeing is I would expect pretty much every, every provider in the email space to add some sort of like generative AI feature set. Yeah. Yeah. And it's, it's just a matter of time and I [00:34:00] think it'll happen pretty quickly given Yeah.

How, like, revolutionary the technology is. But I think you, you touched on an interesting thing, which is like, how well will it do it? We, we ourselves have been playing around with building you know, the ability to generate replies, drafts, and generally what we found is it will generate a reply that sounds like English and is well structured, but it doesn't necessarily capture the.

Or answer the question in the way that you would want it to be answered. And there are different methods for prompting it to say the thing that you wanna say. But really what you want is you want it to have an understanding of past emails you've sent and information in your knowledge base, and then put that in your response.

So if someone says, Hey, do you have the link for the mox for this particular feature? And you can say, yeah, the Figma link is, and then have it auto complete with a link to your actual. And this is the part where I think like we're really gonna see the huge gains in productivity in and being able to like actually generate [00:35:00] replies that understand your data.

Matthew Dunn: What un understand. Yeah. And your, your data even more than devices, understand and can harness your data and take the workload out of the stupid workload outta doing that. Yep. For you, I. You guys handle it already. Like I've always thought the reverse, the reversed email thread was possibly one of the dumbest things we've ever done in civilization.

Like I'm supposed to read this from the bottom up and understand what the hell's going on now. Right, right. Can you like pull the pieces out of there and go, yeah, he needs to know about this and the link is such and such and the, you know, the so-and-so doesn't have Figma permissions or whatever. Right.

Get all of that contextual stuff so I can, I can focus on actually. Reply that I have to write. Yeah, yeah. We're looking at the, we're looking at this early moment, I think of the, you know, visible AI revolution at, at sort of the [00:36:00] surface and cosmetic stuff. And I think what, what you've been chasing for a number of years, it's clear from your, from your comments are, are sort of more fundamental structural.

What do we actually do when. Getting work done with email. Right. It's more than just writing a couple paragraphs.

Jacob Wenger: Yes. For, for sure. I think we've, we've spent a lot of time trying to understand like why are the reasons people are opening their, their email and what are they really trying to, to get done, and how can we, how can we guide them on like a happy path, even if.

Everyone comes from their own workflow, how can we guide them on a workflow that will like, allow them to get in and out of their inbox and understand the priority of the things in their inbox? Gotcha.

Matthew Dunn: Gotcha. So, parting question, cuz with, with the, the job you guys are tackling, I do not wanna chew up a ton of your time.

Where would you like to see Shortwave as a company, as a product in, say,

Jacob Wenger: I think in, in a year we're gonna be making, we're gonna continue making a larger [00:37:00] investment in the AI side of things. So I would like to have an email application where you can essentially talk to your inbox and ask it questions in a way that you can talk to J G P T and you can say, you can use it to conversationally to draft email.

You can use it to say, Hey, what is my frequent flyer mile on Delta? And it'll pull that information from your email archive and just be able to have all your information easily accessible from a natural language interface. So I think what you'll see is you'll see a larger doubling down on AI features mm-hmm.

And making the most out of this like crazy moment that's happening.

Matthew Dunn: Crazy moment indeed. Is the is the competitive demand. Expertise with the set of technologies involved. Getting going, going bananas as well.

Jacob Wenger: I, I think one of the, the cool things about this wave of AI is that people who are not tr don't have an AI background.

Like I have no AI or ML background are able to kind of [00:38:00] use this new technology because it's all just English. Yeah. Like you just interface with it by talking to it. And if you can express yourself clearly and understand how the pieces. Can play together. Yeah. You can actually build these really complicated AI systems without having any understanding of neural nets or the training aspect of it.

That's the sort

Matthew Dunn: of layer 1, 2, 3, that you were talking about earlier. Right. You become a. Consumer of user of without having to build the

Jacob Wenger: of. Yes, exactly. And thankfully, our, our team is a, a team that has a lot of experience building distributed systems. And we are, we've been able to pick up a lot of the new technologies fairly quickly and figuring out how we can best do like build these really, like, mind blowing experiences.

Matthew Dunn: Nice. Nice. Well, I have to say short wave is already pretty mind-blowing experience. So here's the wrapper. If hopefully a bunch of people listen to this go, Ooh, I gotta try this client, where do we send them? Yeah,

Jacob Wenger: so you can go sign up@Shortwave.com. [00:39:00] Like Shortwave radio.

Matthew Dunn: It's like Shortwave radio. I love that reference, by the way.

I may have thought of when I first saw, but some of us are old enough to have listened to a Shortwave once upon a time. Not you. If

Jacob Wenger: you, if you get to, to inbox zero in, in Shortwave, you'll, you'll get a little illusion to to that. So I'll leave that as a teaser.

Matthew Dunn: Yeah, ve teaser for someone else. Inbox zero and Iron.

Haven't seen each other in a long time, but I'll keep trying. Jacob, what a delight to speak with you. Thanks so much for making the time.

Jacob Wenger: Thank you for having me on. It was great conversation.

Matthew Dunn: All right. My guest has been Jacob Winer, co-founder, ed Shortwave. Go check him out. It's amazing. Thanks Jacob.

We're out. See you.