In short
AI Today Podcast Episode Summary
Episode Title
Amplifying AI Writing: Lex's $2.75M Funding Round
Overview In this episode of AI Today, the hosts analyze Lex's recent achievement of raising $2.75 million in seed funding, led by True Ventures. The discussion centers around the implications of this investment for Lex's AI writing tool, its position in the market, and the broader landscape of AI in content creation.
---
Key Points Discussed
Lex
The Company
- Background:
- Lex is an AI-powered writing tool spun off from a subscription media service called Every, co-founded by CEO Nathan Bechez.
- Aims to redefine writing platforms by integrating AI to enhance user experience.
- Purpose:
- Targets a digital-savvy writing community that has not fully adopted AI.
- Focuses on being a supportive assistant to writers, enhancing their processes rather than replacing them.
Investment and Competitive Landscape
- Funding:
- $2.75 million seed funding led by True Ventures.
- Raises questions about its competitive differentiation from established tools like Jasper and ChatGPT.
- Market Context:
- Discussion of the ongoing Writers Guild strike against AI content and the concerns from writers about job security.
- Nathan Bechez argues that AI in writing represents a progression and needs to be embraced by writers.
Features of Lex
- User Interface:
- Aims for a clean user interface, free from traditional clutter of word processors like Google Docs and Microsoft Word.
- Functionality:
- AI assists when the writing process stalls; users can prompt the AI to fill in content.
- Features include asking questions, rewriting, and generating headlines.
- Future plans include rephrasing tools and summarization capabilities.
Concerns and Critiques
- Skepticism About AI's Role:
- The hosts express doubts about AI's ability to truly assist writers without replacing them.
- Concerns about the lack of originality and differentiation from existing tools like ChatGPT.
- User Data Privacy:
- Bechez claims that Lex is not using user content for training AI models.
- Host raises concerns about transparency and the potential for data misuse.
Business Philosophy and Future Outlook
- Startup Strategy:
- Lex aims to maintain a small team and control expenses, adhering to a "hire when it hurts" philosophy.
- Plans to charge users around $20-30 a month, although pricing details are not finalized.
- Initial Success:
- Gained 25,000 users within the first 24 hours after a YouTube launch.
- The hosts acknowledge the potential market opportunities for AI writing tools but remain critical of Lex's long-term sustainability.
---
Conclusion The episode concludes with a mix of optimism and skepticism regarding Lex's strategy and the future of AI writing tools. While the ability to enhance writing efficiency is acknowledged, the hosts question the uniqueness of Lex's offerings in a competitive landscape dominated by established players like ChatGPT. The debate about the role of AI in writing continues, particularly against the backdrop of ongoing concerns from the writing community.
---
Key Takeaways
- Lex aims to be a supportive tool for writers rather than a replacement, but skepticism remains about its effectiveness.
- Funding and initial user acquisition suggest a promising start, but the long-term viability and differentiation are questioned.
- Concerns about data privacy and ethical use of AI highlight the ongoing dialogue within the AI community and among users.
Written by AI. May contain mistakes. Listen to the episode to check what was said.
Transcript
Automatic transcript. May contain errors.0:00Lex is the AI powered writing tool and it recently announced a$2.75 million seed funding round led by True Ventures. So TrueVenture is a very solid venture firm. A lot of people are asking, why did they invest into this company when this is very similar to a concept like Jasper? There's not a lot of wow factor to an AI writing tool, essentially. That's just what ChatGPT does by default. Why would people use something like Lex? So essentially, the platform emerged as a spinoff from Every, which is a subscription media service focused on technology and productivity. And this was co-founded by Lex's CEO, Nathan Bechez.
0:40And really, it is targeting a digital savvy writing community that wasn't wholly adopted AI in their work yet. So Lex aims to redefine what a quote unquote modern writing platform means by integrating AI into functionalities and kind of giving a really clean user interface. So Nathan Bechez explained that the integration of AI into writing tools represents a natural progression in the long arc of advancements in the practice of writing. This is really interesting, right? This is coming at the same time as, you know, the Writers Guild is holding a strike against AI content. And there's all of these, there's a lot of writers complaining about AI.
1:20And so Nathan Bechez thinks that, you know, this is kind of the next step that writers actually need and actually need to embrace. So although AI technology has often been perceived as a potential threat to writers, Lex really kind of seeks to serve as an assistant that enhances and streamlines the writing process rather than replacing the actual writer. This is interesting. I think it's a happy concept, to be honest. I don't buy it. I think AI completely replaces the writer. I mean, you could still have a writer that gives like a three sentence prompt outlining exactly what they want it to write and in the same tone and style and whatever.
1:59But I think like beyond that, beyond some like dropdown menus, which really doesn't even have to be a writer. It could be anybody. I think AI will completely replace it. So that's their theory. That's mine. Let's see who, which one actually plays out in the end. But the platform essentially brings to the table several features that are targeted at power users. So those who are engaged in intensive writing activities, the AI is activated when the user's writing process slows down or halts. So if you ever get stuck, the company guides, quote, just hit CMD plus enter or type plus plus plus and GPT-3 will fill in what it thinks should come next.
2:38So I hope they come up with other tools because if it's just like, oh, gee, I can't think of what to say next. Like, dude, Google is going to integrate this into Google Docs. This is already in Gmail. Like, also, this is, I don't know why they specifically say that, like, GPT-3 is going to fill in what it thinks should come next. Why can't they get GPT-4 on there? Whatever. I don't know. Unimpressed up until this point. But let's keep going because maybe they have some redeeming qualities. Within the Lex interface, users can also utilize AI to ask questions with, you know, within comments, requests, rewrites for making something more brief or even generate headlines.
3:15wow generate headlines make things more brief what is this sounding like literally just like chat gpt if they got money for literally doing exactly what chat gpt has but just like kind of slightly rebranded it and gave it a skin for like writers they will have pulled off the ultimate uh the ultimate heist so congratulations to them so while the ai generated outputs can occasionally produce um inaccurate content that's uh beshez who's their ceo finds it quote really helpful for overcoming writer's block. The company also has plans to introduce more features that are going to rephrase sentences and generate summaries.
3:51What do you, oh my gosh. So this is the thing that's crazy to me is the fact that they're planning these new features that are gonna rephrase sentences and generate summaries. Like what do you mean planning them? Like that literally already exists in ChatGBD. Like you could just make it so you select it and have a dropdown button that you push go. It will take you, literally you could have a developer implement that in two hours max. Like, what do you mean in the future you're going to add those features? That takes two seconds to add. Like, I literally just know developers that could do that in two seconds.
4:22So, again, not super impressed, but let's dive deeper. So, concerns about user data privacy, I think, were also addressed by their CEO, Bechez. He said that Lex is currently, quote, not using any user content for training. So, should the company choose to fine-tune its models in the future, it promises to be transparent and careful about the types of data used for training, ensuring they meet most users' privacy needs. Okay, this is the most ludicrous thing I have ever heard. And this makes me so, this is hilarious. Honestly, congratulations for pulling off such a heist. But it's funny. They're like, we're not using any user content for training.
5:00Yeah, no duh, because you literally just said that you're using GPT-3 to complete. And anything that uses gpt3's api doesn't get added and doesn't get used for training um so the fact that they're like should the company choose to fine tune its models in the future it promises to be transparent and careful about the types of data used for training blah blah blah pretty much what happens here is when you're doing this uh because i've literally had this whole all of these things i've had happen in in startups i'm doing now ai startups but pretty much what happens is when you're grabbing the api call from like open ai right you give it to the user you can still save it all on your own server, essentially everything that's coming.
5:37And so you could theoretically use all of the back and forth, all of the questions and answers that are happening to fine tune, to essentially do fine tuning, right? So like ChaiGPT even has, like a lot of people are just doing it with different third-party tools, but ChaiGPT just announced that they're doing their own in-house fine tuning of GPT-3. So essentially what you could do is like a tool like this, when it generates like an automated response or like a recommendation for somebody. It could say, was this helpful, right? ChatGPT already does this. So they're already obviously using this, but it's mostly on the free version.
6:12But anyways, you could say, was this helpful? Then you mark that down in your own database and you could actually make your own fine tuning specifically that you know, like this isn't just ChatGPT's fine tuning. This is like specifically for writers and for what writers are working on and specific modules like that they're doing. So maybe you have like a headline generator and you fine tune the headline generator from your own, you know, the people saying whether a response was good or not, whatever. So the fact that they're saying like, if we want to do it in the future, we'll let everyone know.
6:40So like they're collecting the data just like anyone else. They haven't integrated it just like anyone else. They're literally just grabbing this from, from open AI's chat QBT through an API. I think it's just pretty funny. I mean, well, well played by them, I guess. In any case, the platform kind of breaks away from the traditional clutter often seen in legacy word processors like Google Docs and Microsoft Word, which tend to carry over historical design elements intended for paper-based publishing. So I mean, I will give them that, right? It's slightly cleaner, but essentially by focusing on what a digital-first writing tool should be, Lex says that they offer an experience designed for contemporary writing, unhindered by pagination or unnecessary features.
7:24I think that's also code for like, we made this really fast and we didn't want to add a lot of features so it's minimalistic so in terms of its journey from concept to viable startup lex found immediate traction after uploading a simple youtube video it managed to attract around 25 000 users within its first 24 hours i mean congratulations that is that that's big that's important right 25 000 users in the first 24 hours i will say i've talked to a number of other founders who like a tiktok video went viral about their product and they got like 100 ,000 users or something. So I'm assuming this is the same thing that happened with YouTube.
8:00I mean, that's cool. That's good. Is it anything different than ChatGPT in the long run? Apparently no, but they got some users, so that's good. I think the initial success drew investor attention, leading to its recent seed funding. So despite the capital infusion, the company doesn't plan to expand its team aggressively. According to Bechez, they'll quote, keep the team really small until it is painful, adhering to a hire when it hurts philosophy to control expenses. Okay, actually what he's saying there is we don't need to hire any more of a team because our product is so simple and we just wanted to raise money because we got 25 ,000 users and it made us look really good so we knew we could raise money.
8:40That's pretty much what happened. As for pricing, although the cost details are not yet revealed, they haven't even picked their pricing yet, the platform's use of commercial large language models suggest that it's not going to be as inexpensive as some might hope. However, Bechez hinted that Lex's paid tier would likely cost no more than 20 or 30 bucks a month. So Lex's emergence comes at a time when AI-powered digital services are increasingly commonplace, but I do think it differentiates itself by being a tool designed to help writers rather than replace them. And by differentiates themselves, I mean the branding differentiates, differentiates.
9:20So the technology is literally the exact same thing, literally with an API. They got lucky and had a YouTube video go viral, got 25 ,000 users and a happy birthday. They got a couple million dollars. Am I salty? Maybe. Um, is this a useful product? Probably is Chad GPT useful product? Definitely. So, I mean, I don't see a major difference, but anyways, I hope, uh, this, I hope this podcast episode wasn't too salty of an analysis of this platform. But I mean, honestly, this is my opinion on this platform, on Lex, on these types of startups, on these AI rappers. Are they bad businesses? Absolutely not.
9:56I know lots of indie hackers that are making like$100 ,000,$200 ,000 a month off of these kind of tools. So can they make you money? 100%. Do I think they're bad businesses? No. Like, go for it. But like if you're if you're saying like you're this revolutionary product and should you raise venture funding for them? That's where I think like probably true ventures that just gave them three million dollars may not get an incredible return. But maybe not. They already got, you know, one lucky hit with the twenty five thousand users. Maybe lightning strikes twice in the same place and they're able to pull off something bigger.
From the publisher
In this episode, we delve into Lex's recent funding achievement of $2.75M, exploring how this investment will enhance its AI writing tool and revolutionize content creation processes.
-
Invest in AI Box: https://Republic.com/ai-box
-
Get on the AI Box Waitlist: https://AIBox.ai/
See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.
