This CEO thinks software devs will be ‘just fine’ amid AI worries

00:00 Speaker A

What should an investor look for in some of these software companies? How do they know if they are durable?

00:04 Speaker B

You know, I guess, you know, I mean, we’re trying to get these cycles across mobile, cloud, etc. I watched as. I think you’d like to see which companies are adapting to this. That’s why I think it’s important for companies that stick their heads in the sand to go bankrupt. But I think most of the companies that are being penalized now are actually responding well and will do well.

00:23 Speaker A

Do you see themselves, do you see them reinventing themselves?

00:26 Speaker B

I think so too. Yes. Yes.

00:27 Speaker A

Talk to us about what you’re working on at Grammarly.

00:29 Speaker B

Yes, because Grammarly people are strangers, we are one of the most popular communication assistants in the world. Approximately 40 million daily active users generate over $700 million in revenue. I think this is one of the world’s best kept secrets.

00:43 Speaker A

I didn’t even realize it was this big now.

00:44 Speaker B

Yes, this is a much bigger job than people think. But actually, the thing that people misunderstand the most about this is that they think it’s about grammar. This is a completely reasonable assumption. But what Grammarly actually does is the OG AI assistant. It is an assistant working right next to you. We see millions of unique surfaces a day. So every desktop application, i.e. web application and mobile application, we can observe what you are doing. We can unobtrusively annotate and make changes on your behalf. Right now we’re only doing this for grammar, and the big change is that we’re about to do this for everything. So anyone can run agents like Grammarly on the same platform.

01:21 Speaker A

How will this change the user experience?

01:23 Speaker B

Yeah, maybe a simple example, let’s say I’m writing I’m a salesperson, I’m writing an email to a customer. Today, Grammarly makes you feel like your high school grammar teacher is sitting on your shoulder, marking everything with his red and blue pen. Now you will feel as if other agents are with you. So your sales coach may be saying, you’re about to recommend the wrong product. You may have support, uh, this person had an outage yesterday, you should acknowledge that. Your electronic device or digital assistant may be saying that you said you would meet tomorrow evening at 7, but your daughter has a recital at that time. And all of this can help you, just as we’ve helped millions of people with their grammar. Now we’ll be able to help millions of people with everything else they do, too.

02:08 Speaker A

As someone who actually worked on it, I can say that this is an amazing technology. Are you afraid of Anthropic’s next update?

02:14 Speaker B

I mean, we’re a big customer. So, Anthropic’s customer. Yes, we are customers of Anthropic, Open AI. We process over 100 billion LLM interviews a week. So if you’re judging us as just an AI property, we’re probably one of the biggest in the world. However, we do this in a way that delivers it directly to the user so that the user does not have to think about it.

02:35 Speaker A

Are you worried about them making a product like yours obsolete?

02:39 Speaker B

I don’t think so. I mean, what we’re doing is a little bit different in terms of bringing AI directly to the surface of humans. I think they’re working hard on models, they’re working hard on what we call chat interfaces. But if you think about the main metaphors in AI, there are a lot of people focusing on conversation, we want to talk to an AI bot. There is too much focus on what we call doing: task automation. We’re working on something we call assists. So if I give you the statistic that I’m doing 100 billion LM searches a week among 40 million people, that means for the average Grammarly user, we’re doing a few thousand AI searches a day. So, if you’re a really good chat user, if you’re a really good cloud user, you can do this maybe a dozen times a day. But we do this every time you type a character, open a new document, open a new application, we call all artificial intelligence systems on your behalf. And I think this scale and this level of integration is a very different paradigm for how to think about AI.

03:37 Speaker A

What about superhuman? Now these businesses have been merged. For example, what is the new generation superhuman?

03:42 Speaker B

Yes, we brought together four different products. So the original Superhuman product, what we now call Superhuman Mail, Grammarly, my old product Coda, and the new product we call Go. Go is Grammarly’s platform layer and allows anyone to build Grammarly-like agents. We brought the four of them together and then decided to rename the company and we had a lot of ideas about what to do. We decided to remove one

04:10 Speaker A

It looks great. No fault of Grammarly. I mean, Superhuman sounds good.

04:12 Speaker B

I think yes, it sounds good. Obviously this is a starting point. That’s it. It looks great. It should sound nice. It’s a broad topic, which I think is really important, so you can cover a lot of things. But what we actually love about it is the word human. Because if you think about what most people working on AI are really worried about, you’d think they’re trying to replace humans there. We have the exact opposite perspective. Grammarly has always been the product that we work with, but at the end of the day, you write the article, you submit the blog post, you submit the article. Oh, that’s all, we’re there to help you. I like to say that we spent the last 16 years turning people into superwriters, and now we can spend the next few decades turning people into superhumans.