Is This AI The End Of Content Creators?

Manav: In this video of Emerging Founders with Manav, we'll be interviewing Loïs. He's the founder of Argil AI, and they're building AI avatars that look insanely humanlike. He's one of the most interesting people I've actually talked to — he was a filmmaker, he's really good at fundraising, and he went through Y Combinator. He tells all these stories in the podcast.

I'm really excited to have you on the show, Loïs — I really want to use your product myself.

Loïs: Thank you for having me, Manav.

Manav: You're one of those multifaceted people who's proven himself across multiple domains — you were in the movie industry, you've helped founders raise a lot of money, you've hired a lot of engineers, you went through Y Combinator. It's really interesting talking to someone like you, because you're not just building, you're doing many different interesting things. How did you end up here — how did you get into so many different interests?

Loïs: I think there's a common thread across everything I've done, which is storytelling. Building a startup, getting clients — that's telling a story about a product and the problem it solves for them. The creative work I did before, writing a novel, making movies — that's also about telling a compelling story. Even helping founders fundraise, which was actually the first "master class" I did, back in 2020 — helping a founder raise money is fundamentally about helping them craft a story that investors can follow and believe in. So the common thread has always been storytelling, just through different mediums, whether that's building something or actually writing something.

For my novel, I wrote a science fiction novel — it took me nine months, but I'm not the kind of person who can put in a hard day of work and then come home and switch into a writing mindset. I've always managed my time by carving out intense, focused periods — a couple of months to a couple of years — dedicated to a single project, which is what we're doing now at Argil.

I also think there's a window, roughly from 18 to maybe your early thirties, where the opportunity cost of trying things is genuinely low. When I was very young, one of my goals was, "by 30 I'll be retired and a millionaire" — that evolved a lot over time. At some point I realized, at this stage of life I can work hard, but I can also do a bunch of creative things I might not have the energy for later — writing a book, making a movie, things you don't necessarily have the time or energy for later in life. Making the most of that window feels really important.

Manav: I want to know the founding story of Argil — what problem are you solving?

Loïs: I've been doing a lot of video content over the past several years, either very creative work, like when I was working on my movie, but also video master classes and TikTok accounts. The recurring issue with video has always been production cost. When you really think about it, video is a very elite medium — it's expensive to get into, regardless of what you're doing. I'm not even talking about Hollywood or TV — even any YouTuber with a sizable audience is spending real money making videos. There's a real barrier: cameras, lighting, either the skill or the money to pay for post-production and editing.

Video ends up being basically the only medium where you can't really run a lot of cheap experiments, because everything you produce is expensive enough that it has to work — if it doesn't, you either abandon it or eat the loss. We were working on a bunch of AI projects about a year and a half ago, and at some point it became obvious that the biggest problem I personally wanted to solve was: how do I make more TikTok videos, more educational videos, without spending a ton of money?

We decided to dig into that model. Looking at the market, we saw some players already doing avatars, but mostly for use cases like enterprise learning. My focus was always that avatars weren't only going to matter for internal enterprise use cases — they'd eventually be used for social media too. At some point in the future, it'll make no sense to pay for a $5,000 camera setup if you can generate essentially the same content by typing it into a platform. If you accept that this is inevitable, you can work backward to what we're building today — the first foundations of that technology and the platform to get there.

I'm genuinely bullish on this idea, even though a lot of people point out downsides — "you're replacing humans," or some companies thinking "we won't even hire creators anymore, we'll just use avatars to pump out short-form content." But from what I've actually observed, it's more job-trading than job-destroying — nobody's really being replaced. The companies that never did video before start doing it with avatars, but they never had the budget for real video production in the first place. The companies already doing video mostly just enhance their existing workflow with it.

Manav: A bit of context for people watching — there are existing tools like HeyGen and Synthesia, but they mostly target enterprise companies, and their avatars look pretty robotic. What Loïs has built with Argil genuinely shocked me the first time I saw it, because it looks so organic. How did you achieve that — that's really impressive.

Loïs: I think one of the things you think about when starting a company is: what real edge or "secret" do you have about the industry? A lot of the founders behind the companies you mentioned are very strong technical profiles — but I don't think any of them were content creators before starting their companies. So I'd say it's less a pure technological breakthrough and more an approach — we decided that the avatars need to be good enough to actually earn views, likes, and subscribers on social media. Every decision we make flows from that standard: which avatars we choose, the features we build — you can control body language, body movement, natural gestures — and even how our editor works, which is built to output a fully edited video in about two minutes, captions, backgrounds, everything. The whole pipeline was designed around one goal: the video actually has to perform. It's a combination of a lot of different components.

Manav: I've actually used the product, so I know how the setup works — can you explain to people how someone signs up on Argil, and how long it takes from starting the process to getting their avatar up and running?

Loïs: It depends on the quality level you want. Content creators who are already producing at a high level, with good material and a proper setup, will often spend an hour or two making an absolutely flawless avatar, because they're replacing their actual on-camera self with it, so it has to be extremely good — they'll basically replicate their real setup and use every body-language gesture in the training footage, so it carries over into the avatar. But for the average person who just wants an avatar for some general use case, you really just need about three minutes of footage of yourself talking to the camera, with decent lighting and decent audio. You can also create two or three versions of your avatar with different outfits or backgrounds. Overall, starting from scratch, it can take about 20 to 30 minutes to get your first avatar training started on the platform, then a couple of hours for training to complete, and after that, generating an actual video only takes a couple of minutes.

Manav: Will you be integrating a voice model directly into the software, or does that always have to be done separately, like through ElevenLabs?

Loïs: Right now you can generate the whole package on the platform — if you have an ElevenLabs account with specific voices you want, you can plug that into Argil directly. In the future we'll probably build our own voice model, but right now it's very convenient to use their infrastructure, since they already support so many languages.

Manav: I'm really excited to see how you guys progress, because this is a big challenge — filming every day is such an inconvenience, so much setup, so time-consuming, and for a lot of people that technical barrier alone can be overwhelming enough that they procrastinate, or don't post consistently, and that inconsistency ends up being the reason they burn out and quit, because they don't see results fast enough. We all know podcasting, YouTube, TikTok, Instagram — it's such a long-term game.

Loïs: Exactly.

Manav: MrBeast didn't see real results for something like seven years — if he'd quit, he wouldn't be the biggest creator in the world right now. It's great that you guys are building this. I want to know what's next for Argil — what are you focused on, and where are you right now in terms of go-to-market, is it still early stage?

Loïs: Right now the platform is live, we already have a lot of users, a lot of paying clients, ranging from a couple dozen dollars a month up to thousands of dollars a month depending on size, since we also have an API for high-volume use cases. A good example is a job board that wants to turn every one of its job listings into a video, at scale.

The vision I have is that, in the future, creating a video should be as easy as posting a tweet. If we want to bring in a new generation of content creators who so far haven't been willing to do video — because of the effort, or because they don't like how they look on camera, or any number of reasons — but who'd happily post on LinkedIn or Twitter because there's no real barrier there, we want to remove that same barrier from video. The goal is to keep making the model more and more realistic, and keep building out the platform to generate pre-edited videos that genuinely perform on social media, creating that loop — I make a video on the platform, post it, get a bunch of likes — and eventually we want to track that performance data to help you draft future videos. Longer term, we want the platform to act as a genuine assistant, helping you figure out what content should even exist.

The way I see it, in about two years, the average local plumber or carpenter is going to take a picture of a job they just finished, and it's going to turn into a really compelling video, posted across all their social channels, bringing in local clients. Everyone has an incentive to create video, and we want to help people actually do that.

Manav: This is probably something you think about a lot — how are people going to differentiate real from generated content going forward? Is there going to be some kind of watermarking situation on videos? It's not just your platform, there are so many AI-generated video platforms now — how do we differentiate in the future, or will that need even go away?

Loïs: If I had to predict, I'd say for the next two or three years we'll still need to watermark videos, because we're at a stage where people feel a bit cheated or put off when they realize something's an avatar. A lot of what I personally do is explain to people that I'm not cheating anyone — I'm still creating the content, I'm just saving the energy and money of filming myself, but I'm still trying to deliver as much value as I would if I'd filmed it myself. We need to go through that transition period, but especially because younger generations are so much more comfortable with AI avatars and AI generally, I'd guess that in three or four years, the need for watermarks will basically disappear, because the audience simply won't care whether the content is AI-generated or not. I think that'll apply pretty broadly, honestly, across everything.

Manav: I completely agree with that — I think AI is just raising the bar on the quality people expect. It's funny, right now there are AI-generated ads that are actually outperforming human-made ones, just because they stand out more, which, in my opinion, is a little crazy.

I was actually going to ask you for a prediction, but you basically already gave me one, so I'm happy about that. Thank you so much for coming on the episode, I really appreciate your time — I know you're an extremely busy person, and I appreciate you sharing your thoughts. There are a lot of great nuggets in this conversation, and I'm really excited and rooting for you guys — the quality is genuinely impressive, there's no other company doing this at the level you're doing it, and it seems like you're on a great trajectory.

Loïs: Thanks a lot, Manav.

Manav: Thank you, Loïs — see you soon.