Top News

Quill gets funding for its privacy-first meeting notetaker

In the years after the COVID pandemic, meeting notetakers have become insanely popular among both users and investors. Companies like Read AI, Granola, and Fireflies have raised millions of dollars in funding. These companies possibly shipped fast and got into the user mindset.

Quill, a newer entrant in the market, took a different approach. It stayed quiet and acquired customers rather than doing any press. CEO Michael Daugherty, who was the chief strategy officer at AngelList, said that the company, founded in 2023, was bootstrapped for the first few years with little money raised from friends and family.

The startup announced earlier this month that it has raised $6.5 million in funding led by Basis Set Ventures.

Quill’s focus is on making the app work on your device as much as possible. For instance, the transcription of meetings happens locally. Users also have a choice of AI providers, where they can choose cloud models or local-first models.

“Our view is that in a world with AI, we want to enable people to have more control and more leverage. And that means control over where their data lives, how their AI is processed, but also control over what they do with AI. And so the first trend is that every knowledge worker, every job is becoming a manager of AI agents,” Daugherty told me over a call.

For Daugherty, the core principle is that there are tons of AI companies trying to get more data and context from users, which could be used for training or ads. In this era, Quill wants to provide more privacy controls to users. The company is also open to integrating its tools and data with other services through integrations like MCP (Model Context Protocol).

The meeting tool offers a generous free tier with local recording and transcription. You have to pay for pro tools like live meeting minutes and AI-generated action items.

Study around AI-generated voices puts models from Minimax, Wellsaid, and PlayHT at the top

Vocal Image, a site that uses AI to train people in better communication, released a study involving over 10,000 participants to test out different text-to-speech (TTS) models. Notably, the study focused on how humans perceive AI voices.

The study noted that the voice model from Minimax, a China-based company, came out on top with a 86.2% of approval score, followed by 85.6% for PlayHt and 82% for Wellsaid Labs. Voice AI giants like Eleven Labs and Deepgram only had 74% and 68.4% approval rates, respectively. From big tech, only Microsoft made it to the top 10.

In terms of model quality, PlayHT scored first place with a score of 85, followed by WellSaid and Minimax, both of which ended up on 81.

Vocal Image’s study said that native English speakers are more likely to detect AI-generated voices than non-native English speakers. For voice companies, this means that the latter cohort accepts AI voices better with a higher approval rate.

The report noted that users in Saudi Arabia had the highest approval rate of 72.5% for AI voices. India and the U.S., countries with over 2,000 participants each, were fifth and sixth in the rankings with 67.1% and 66.8% approval.

Quick Bytes

• Krisp launches listener-side accent conversion, which can convert your accent with on-device processing.

• Speechify launches a Granola competitor for meeting note-taking, but it only works in Chrome at the moment.

• Music tool Moises hires singer-songwriter Charlie Puth as its “chief music officer” to drive its AI products.

• New Zealand-based company NestEdge has been facing backlash for its AI-powered voice calling tool that sounds too real.

• Munich-based voice AI platform for enterprises, Voiceline, raises $11.5 million in Series A funding. The platform helps sales agents improve their performance.

Singals & Experiments

I tried Krisp’s accent-changing tech, and I get the need for it for certain calls, but to me, the tone copying didn’t work. The voice that I tested didn’t feel like my own voice. I might find it useful to convert someone’s voice who is heavily accented, but only if I can hear some words clearly. Otherwise, personally, I don’t mind asking someone to repeat themselves because I didn’t understand them.

Thank you for tuning in. Keep listening.

Have a story about call centers?

Email: voiceaiweek@gmail.com or im@ivanmehta.com

Apple tried making Mini a hit, but after poor sales and two iterations, it gave up. It then brought a big-screen phone in the iPhone 14 Plus for the fourth slot in its iPhone lineup. We saw two more Plus phones after that, but the versions performed horribly over the years.

This year, the company launched the sleek and slim iPhone Air. The phone is just 5.64mm thick (vs 7.95mm of the iPhone 17) and weighs 165 grams (vs 177 grams of the iPhone 17).

The phone looks good, and I am sure it gives you a different feeling when you hold it. When Apple made the announcement, many people thought: I want it.

There has been commentary around the iPhone Air pointing towards the “No one asked for this” board.

Specification-wise, this device doesn’t make sense as a purchase. It has only one camera. The battery life on paper seems okay, but it is lower than that of the base iPhone 17. Apple even advertises the phone with an additional MagSafe-based pack that it launched. We don’t know how the battery life will hold up in day-to-day use and long-term ownership.

Just like early foldables, there is a fear that a device this thin might not be durable.

Logically, it might not make sense for you to buy that phone, but the phone is not for people who are thinking logically. This is an impulse purchase for people who can afford it. This is a purchase for people who want a unique-looking device and don’t care about battery life or having multiple lenses to capture photos.

Apple might also be setting itself up for future phones to be thinner. Plus, it could signal prepping for the rumored foldable, slated to release next year.

The Mini and the Plus series were “logical” choices for the fourth slot in the iPhone line up. A compact phone could appeal to an audience that likes the category. When that didn’t work, Apple tried enticing a separate category with a large phone. People didn’t respond to both categories. Either wanted a base phone or wanted to spend more and get a Pro phone.

With the Air, Apple is playing the FOMO card. It wants you to buy it as a fashion statement. For the company, this is a punt on design after playing two supply chain cards for the fourth slot.

Next year, we might see the foldable as the top-of-the-line phone, along with the base model and two Pro phones. If the Air gives better results than Mini and Plus lines, the company may keep it around. We could very well see iPhone SE styled (now the iPhone 16e) updates to the Air in the future, where the company doesn’t release new phones every year. That might be the reason why its called the iPhone Air and not the iPhone 17 Air.

Companies like Perplexity, The Browser Company, and Opera have been chasing the dream of an agentic browser that will handle or automate tasks for you. OpenAI is reportedly set to launch its own browser in the next few weeks.

While the north star for these companies is to have the browser fulfill complex tasks set by users, right now, they are trying to change one basic and repetitive habit: search. Most of our brains are wired to type in a few keywords and rely on Google to give us matching and relevant webpages.

However, ChatGPT and other chatbots have changed this habit in the last few years. People using these chatbots often write elaborate prompts or queries to ask. And a lot of the time, they get the answer within that chat window, without needing to click on a link.

With Chrome, Google instilled a habit in users to type their search into the address bar (omnibar). That’s what new browsers like Dia and Comet aim to do. Users now have access to a chatbot directly in the address bar, so you just need to type your query and get chatting.

For Google, the key driver was to have people using Chrome drive Google searches. For browsers, it is having user to chat with their own chatbots. These companies are trying to build tools within their browsers to have power users not switch to other tools.

These include mentioning multiple tabs and answering questions, having the browser read the currently open page and tell you relevant information, having connectors with other apps like Gmail to get you information, or letting you code mini apps for repeatable tasks.

While some of these features are well-suited for advanced users, search is now one of the most fundamental actions for all kinds of users in a browser. And if browsers can drive the habit of having users ask detailed queries and follow-ups using their own tools.

Browsers have relied on search deals with giants like Google to drive revenue, apart from offering a few goodies. Tools like ChatGPT have gnawed a little bit of search market share from Google. Browsers are hoping that this is a sign enough to have their own Q&A stack and user base to entice advertisers eventually. Some of them might offer premium features, but subscriptions are unlikely to drive the majority of revenue, leaving only advertising as a viable long-term revenue driver.

Google is already integrating Gemini into Chrome and won’t give up its search and ad market that easily. Unless they are forced to, of course. AI browsers are still a niche category in overall web consumption. But if Google has to sell Chrome, they will become really important all of a sudden. And the way we browse the web will change again.

The company said:

“Pocket has helped millions save articles and discover stories worth reading. But the way people use the web has evolved, so we’re channeling our resources into projects that better match their browsing habits and online needs.”

In another blog, Mozilla said that it will build discovery and curation in Firefox.

Pocket is one of the apps that I have been using for countless years. Saving articles to Pocket became a habit of mind. It was an app that let you hoard articles that you might never even read. It was akin to hoarding books that you would never get to reading, but only cheaper.

The beauty of it was that it made it easy for you to save links from anywhere through apps, extensions, shortcuts, and widgets.

Pocket’s shutdown is another signal of a shifting landscape in the link-based web. Social platforms were already pushing users to post “original” content — content without links.

On top of that, AI chatbots are now increasingly making a dent in search volumes. A side effect of chatbots providing ready answers is that people are not going to certain websites to read articles. They are rather relying on chatbots to summarize everything for them.

Multiple studies say that Google’s AI overviews feature, which summarizes information for certain queries, has an adverse effect on people clicking links to websites. Just today, a travel website owner wrote a letter to the U.S. Federal Trade Commission (FTC) highlighting how Google’s new search efforts have been detrimental to his site.

The media industry has also seen widespread layoffs, with some publishers happy to let AI write content. This means there will be more AI slop and less content written by humans who have developed skills in writing.

In this scenario, discovery becomes important. However, companies will opt to let AI also handle discovery rather than rely on human curation.

Pocket’s role on the internet was to become a tool for people to build libraries. Along the way, the company also flirted with discovery. We can only speculate the core reason behind the app’s shutdown, but Mozilla likely saw signs that people were not using the app for discovery. Heck, if AI is summarizing everything for you, why do you need to discover articles anyway?

There are plenty of Pocket alternatives for users who still want a reading library of blogs and articles. But in this age of the internet, their future is also not secure.

I saw a lot of commentary around the coverage questioning the money raised for a notetaking app. You can have endless debates around whether a $250 million valuation of such an app is justified. But I know that Graonla, or several other competitors, raising millions of dollars is not at all surprising.

Digital notetaking has been around for years. People started using notetaking apps to better organize their thoughts and notes and have them accessible from anywhere. However, to be a meticulous notetaker, you still had to develop methods of capturing the best points while you are in a conversation, and later organize them as well. At times, if the conversation was important, it was not polite or efficient to take out your phone and start jotting down thoughts.

Meetings, the primary space these apps are targeting, also had someone manually take down notes to determine next steps. In either case, if you didn’t have a great memory or didn’t write down the detail the very moment someone said it, it became difficult to follow up or recall what was said.

Because of the COVID pandemic a few years ago, most meetings went digital, and people started recording those. But even then, only a few people used AI transcribers like Otter or Fireflies. Mostly, because you still had to go through transcription anyway.

But a few years ago, speech-to-text models became better, and in turn, transcription was improved. Plus, as the number of large language models available increased, it was easier to go through a ton of text and generate things like summaries and insights. For app makers, this was the perfect opportunity to make users “lazy” and release them from their notetaking duties.

Some notetakers prefer to be in the meetings, such as Otter, Fireflies, Read AI, and Circleback; some just use system audio to transcribe, such as Krisp, Granola, and Limiteless. While most of them are passive notetakers, you can use Granola to add your own notes during or after the meeting and also edit AI-generated notes.

Venture capitalists and founders take a lot of meetings every day, and it is hard for them to remember all the stuff they talk about. A lot of information amid those conversations is “high-value” information that they wouldn’t want to forget. They could have another person take notes, but when you have an app on your phone that can take notes all the time, you don’t need any manual intervention.

With an app like Granola, you can dump as much voice and text information as you can. The promise of AI is that it will help you retrieve that information any time.

The ultimate promise of these apps is that you can recall what you said to some person a year ago about drinking a particular brand of coffee, and if you recorded/transcribed it, the app will have that information for you. The hidden tagline for all these apps is “never forget anything now.” For investors who are looking to bet on something people rely on more day by day, that is like catnip.

I started writing blogs more than a decade ago. For me, the idea was to experiment with English, which is my third language, and get better at writing it. Even after a decade in professional writing, I don’t claim to that I spit words smoothly. I still make errors and typos.

For a large part of my writing career, I have relied a lot of spellcheckers of different kinds to guide me through mistakes. Of the various tools I have used, Grammarly has been one of the constant ones. Largely, because it integrates with most writing interfaces easily. That means I don’t have to look for built-in spellcheck tools in different apps.

Lately, Grammarly has been very annoying. The red lines below words and phrases are still fine because they tell me where I am going wrong. But now the app is also displaying yellow lines to show “Pro suggestions” for writing. I did try Grammarly Pro for a bit to understand if those suggestions made my writing drastically better. But that’s not always the case.

When I write for a specific type of news story or a feature, the tone and flow depend on the subject. I can’t possibly have Grammarly understand that particular nuance. So those Pro suggestions might not help me in all cases.

The yellow lines were still fine. But Grammarly has started displaying blurry boxes to hide suggestions under those yellow lines to entice you into buying the premium subscription. You do get one free “Pro” suggestion per day. But once you avail that, you will see blurry boxes.

It’s almost as if Grammarly is trying to shame you into becoming a paid user. The box almost screams, “I don’t think what you wrote is good, use our AI to be better.”

Grammarly is not immune to the generative AI wave, and has made changes to its product and company structure (read: layoffs) to adapt to the new world. The AI-powered Pro suggestions feel forceful. The company launched a feature to have its AI understand your writing tone and display suggestions accordingly in 2023. But that’s a paid-only feature.

In a world where feature launches are accelerated through AI, growth hacks are almost an expected outcome. But shaming users into becoming a subscriber is irksome, especially when I would be potentially willing to pay if I see a value. All I see right now is coercion and forced gamification.

There are conversations around copyright, ethical conundrums around AI training on existing art, and energy consumption by data centers responsible for running AI models. But for most folks, these are not points of consideration while trying to create a cute image for some social cred.

Users who might not have any association or recall with the phrase Studio Ghibli would just want an amusing new image and move on with their lives.

The phenomenon of people using a tool to convert their own images into a different style is not new. We have seen this with Prisma and FaceTune. But earlier, you couldn’t just type your way to a new style of photo. You had to select a certain filter and do some editing on top of it.

Right now, a lot of the new-age AI tools are firmly in the realm of amusement. ChatGPT has seen massive gains in countries like India because of the image-generation feature. Some AI cartoon apps got bumps in downloads because of the image-generation trend.

People who feel an emotional connection to Ghibli movies or art argued that in the longer run, unique things like Ghibli's style of creation won’t remain unique because technology has enabled everyone to recreate that in seconds, while the original creators had to toil for years to create an identity with art.

Founders of AI companies and tools often talk about taste. They say that even if everyone has the tools to generate images or videos by merely typing in a few sentences, they will need to have taste to create something good and long-lasting. This has been an argument with many generations of technological tools.

In the past few days, writers like Scott Alexander and Erik Hoel wrote about feelings around Ghibli images. When we users generate something in mass, there are always questions around whether it is or not.

Alexander writes:

“If you insist that anything too common, anything come by too cheaply, must be boring, then all the wonders of the Singularity cannot save you. You will grow weary of green wine and sick of crimson seas. But if you can bring yourself to really pay attention, to see old things for the first time, then you can combine the limitless variety of modernity with the awe of a peasant seeing an ultramarine mural - or the delight of a 2025er Ghiblifying photos”

For people building these apps, the measure they would want to consider is more people doing more things on their apps. They might not have thought about true artistic representation or even about art. It’s a tool to amuse people in some way. To make them feel that they also can make a good enough impression of some art they might not have been able to.

I Ghibli-fied a photo of my two cats with ChatGPT, and it got some details wrong. The abstract representation was there, and the generated image would have passed on social media as a cute photo. But some details about how they looked and what color their patches were put me off. I tried to edit my way to get a more accurate representation, but failed. Was it amusing enough? Maybe.

For AI apps, amusement — making people feel like they can ‘acquire’ a skill in minutes — is the biggest strength. Can’t write a good email? We are here. Don’t want to make a presentation? Upload the document and click this button. Want to make a video? An image is enough. The amusement of being able to do all this reels us in, and then after a while, it becomes normal. Users do want to use AI for these tasks. Not because it is good, but it is passable and it is fast.

Do I need to acquire a taste to do these? Possibly not as lo

ng as enough people are paying attention, getting entertained, or gaining information out of it. Taste becomes a throwaway word for saying only a few people will be able to make really good stuff, but the rest of you will have a makeshift version of the best thing. It is also hard to evaluate taste when so many AI tools are generating similar outputs.

AI apps are telling us that they will handle drudgery tasks, so you will have more time to do creative or productive stuff. What do I have to do that is productive or creative? Maybe go to another AI app and get amused.

Browsers are at the core of accessing a large part of the internet, even though there is an app for anything and everything. While the likes of Chrome and Safari rule the usage charts, startups like The Browser Company (which makes Arc) and SigmaOS have tried to innovate by creating a complex user interface that with a steep learning curves at times. Older browser companies like Opera, Brave, and Mozilla (which makes Firefox) keep trying the slants of productivity, customization, and privacy to have their userbase grow.

Why do we need an “Indian browser?”

At the crux of browser development, there is nothing nationalistic. You can make a browser from anywhere and have anyone in the world use it. So why did the Indian government create a browser challenge for developing an “Browser for India”? The competition’s site gives us the answer:

“The developed Indian Web Browser is required to have their own trust-stores, pre-loaded with CCA India Root Certificate, which is a self-signed certificate and is the top-most (root) certificate of the tree, that serves as a trust anchor.”

Other listed requirements on the site list standard features of the browser, but a trust store that pre-loads certificates from the IT Ministry Control of Certifying Authority(CCA) is Indian.

Why do we need this? You might ask.

Root certificates are signs of a secure contention through Secure Socket Layer (SSL). When you see an HTTPS on a web address, that means the connection is secure and data is encrypted between the browser and the servers. Different certificate authorities issue this certificate to domains.

In 2014, major browsers stopped recognizing CCA’s certificate as trusted because there were fraudulent certificates through the National Informatics Centre (NIC) —which provides technology solutions to the government — to Google domains.

The government is possibly pushing on this front for contingency planning for dire situations, such as the world cutting off India from the global internet.

“In a situation where our access to the Internet globally may get cut off, to have it completely operational within the country… Today’s efforts will be key in strengthening the resilience of the Internet within the country,” IT Secretary S. Krishnan said during the event, according to The Hindu.

If an Indian authority owns issuing certificates, supported by browsers, it means the authority could decide if a website is secure or not. It can also choose to issue certificates even when other certificate providers might not.

All this certificate stuff is fine, but what about other browser stuff?

For an average user, the certificate might not matter much unless they see a warning about a site being not secure. They would want a functioning browser that is fast and maybe has some extra features.

For such early stage browsers, it is hard to do a direct comparision with established players. We are going to take a look at the top features describes on the websites of three winners of the Indian government’s browser challenge:

• Zoho’s Ulaa browser: This browser claims to be a centric browser with no tracking, geographical data isolation, different modes with website blockers, smart tab management, notes, and integration with Zoho’s own tools.

• Ping Browser: Support for the Hindi language, digital document signature, parental controls, and “AI integrated features” (not sure what that means).

• Bharat Web Navigator: Integrated chatbots, document signing tool, and crypto asset management tools.

Most of these features are nice to have rather and don’t feel that they would attract users right away.

This is not a bad thing, but browsers do need to build a scalable and fast product that is reliable.

Reliability is the factor where these entrants can falter.

A security researcher who goes by @DotSlashTX on X wrote a long thread where he accused Ping of being a wrapper of Brave browser with a few additional features. He added in the thread that Ping’s alleged Brave forking might be violating some copyright claims, too.

Another researcher pointed out that the nomneclature used in Ping’s Linux install file is amateurish and could break system-level aliases.

Uday Bansal, one of the founders of Ping Browser, acknowledged in a post that the company built on top of the Brave browser. He mentioned that the submission for the Indian browser challenge was different than the enterprise browser the company advertises on the website.

He goes on to say that the browser submitted for the challenge had digital document signing through crypto token, parental control features (the security researcher called this out as just an extension), and an AI summerizer. Bansal said that there were some mishaps in terms of packaging certain submission elements of the browsers.

“On behalf of my team, I sincerely apologise for the oversight regarding repackaging certain of the submission's elements - I truly understand how this looks from the community's perspective,” he said.

Despite this clarification, it is not a good look for a new browser.

There are questions around Bharat Web Navigator’s (BWN) features, too. A software engineer pointed out on X that some of the features marketed by the browser are just demo pages. Plus, the company used older versions of Chromium, which could result in security mishaps if released for all users.

The company behind this browser, Anja, is primarily a crypto consultancy building various solutions. As some of the requirements for the browser challenge included adding “web3 content” support, the company possibly took a punt.

The company told MoneyControl that it stopped the development of the browser in April 2024 while awaiting the results.

India’s IT ministry is looking into both Ping and BWN, according to the report.

Did we really get a desi Google Chrome competitor?

The bigger question is, what’s the point of the whole rigmarole? If the companies developed the browser just for the challenge and are not going to put out new versions to really compete with existing browsers, the whole exercise was a failed one.

As some of the startups like The Browser Company have experienced in recent times, building a unique browser, getting users, keeping them as users, and scaling that userbase is very challenging.

The doubts raised by various folks about the two runner-ups and their approach to browser development don’t give a lot of confidence about their vision to develop and maintain browsers for consumers.

Zoho, which has a multibillion-dollar valuation, can possibly put resources behind its new browser. However, it will also face challenges in scaling the product. The browser has strong integration with Zoho products, so some of its customers might be incentivized to use it. The company will have to put a lot of effort into brand-building around the browser and getting mindshare of users to have them switch away from Chrome, Safari, or Firefox.

The Indian government had started programs like an app challenge, which went nowhere beyond the first edition

At the announcement of the browser challenge winners, the IT Minister made a point of pushing India towards a product building nation. But that would require long-term challenges, evaluation, trust building, and support. The browser challenge didn’t possibly set a right example in that direction.

For the last two days, my timeline has been filled with Studio Ghibli versions of people’s photos, memes, movie stills, and pets, all thanks to OpenAI releasing new image-generation capabilities on ChatGPT with the GPT-4o model.

Other platforms and chatbots, such as Google Gemini, xAI’s Grok, and Meta AI, have been able to generate images. But getting a powerful model like GPT-4o to create and edit images is a big deal, given that ChatGPT is possibly the biggest chatbot around. (Meta might want to contest that.)

However, image generation ability — available to only paid users at the moment — was merely a starting point. When people discovered that you can turn photos into the distinct style of Japanese animation company Studio Ghibli — which has produced hits like “My Neighbor Totoro” and “Spirited Away” — they started converting all kinds of photos. Studio Ghibli holds nostalgic value for those who have seen the movies, and for those who haven’t, it can give you a feeling of looking at something distinct.

The image-generation feature or the event of turning photos into Studio Ghibli-styled images might not matter a lot in the long run. But this event hints at an opportunity and a problem for OpenAI.

The opportunity is one more step towards the company’s vision of becoming a consumer tech company. Stratechery’s Ben Thompson wrote about OpenAI being an accidental consumer company during the early days of ChatGPT. But more recently, Sam Altman has been saying this part out loud in interviews and also to investors.

In a recent conversation with Thompson, Altman shared that his goal would be to get to a billion users for OpenAI consumer products in five years. Altman things, in some parallel to Google or Facebook, users will want to sign in with their OpenAI account anywhere, and that will give other applications more context about the user.

“I really believe in this product suite thing I was just saying. I think that if we execute really well, five years from now, we have a handful of multi-billion user products, small handful and then we have this idea that you sign in with your OpenAI account to anybody else that wants to integrate the API, and you can take your bundle of credits and your customized model and everything else anywhere you want to go,” Altman said.

Viral moments like the image generation feature might not be pivotal to OpenAI reaching its billion-user goal. But it gives the product a somewhat temporary boost in mindshare. People already know about ChatGPT, and some might already use it. Seeing this feature, some of the users might convert to paid users even for a while. OpenAI plans to roll it out to free users, but those plans might be delayed as there is high demand at the moment.

There is an ethical and legal gray area that is to be considered here in relation to Studio Ghibli-styled images. The ethical dilemma is that Ghibli creators made this their own distinctive style over the years. Now a computer tool is allowing users to imitate that in seconds without any compensation to the creators. (OpenAI hasn’t announced any public deals with creators.)

The legal gray area is potential copyright cases. OpenAI is facing several of them already. The problem is that a style of images is not protected by copyright. If OpenAI trained on Ghibli material, we don’t know where the company acquired the dataset from and if there is any precedent for the case.

In any case, OpenAI wouldn’t want to worry about that if enough people are generating these pictures and until there are no lawsuits.

On March 17th, podcaster Lex Fridman posted a three-hour-long interview with India’s Prime Minister Narendra Modi. Typically, Modi doesn’t sit down for long conversations. A recent exception is when he sat down with Indian trading platform Zerodha’s co-founder Nikhil Kamath.

More than the topic of the conversation, what struck me was when voice AI startup ElevenLabs announced that it helped in dubbing voices in different languages. The original conversation has Fridman speaking in English and Modi in English. There are full dubs available in Russian and Spanish, too.

What’s notable is the lack of Indic languages apart from Hindi. That puts the conversation out of limit for some users who consume content in their native languages.

Subscribe now

Indian internet users’ primary activity is consuming videos. It is also the top activity for language-specific internet activity. While Hindi represents a bigger chunk, other languages also form 33% of the country's internet consumption cluster.

YouTube is a massive platform for content in India — both for creators and users. There are language-focused YouTube creators and channels that have millions of subscribers. There is an opportunity to cross-sell content in different languages.

That’s where AI could possibly play a bigger role. After Fridman published his interview, Indian AI startup Sarvam’s founder Pratyush Kumar posted a clip from the conversation in different languages using the company’s AI models. However, these clips were just dubbed versions, and there was no lipsync. Another Indian founder, Pavan Reddy, who is building a lipsync model through Sync, posted a lipsynced version of the conversation.

India is a content-guzzling market, and there are tons of users ready to consume content through different mediums and languages. Popular figures like Modi, would connect with audiences more often if the content is available in various languages. People are eager to watch different kinds of content as long as it is available in their native languages. A research platform called Digital Desh highlighted in its report about Bihar that people watch Hollywood and South Indian movies along with Korean content dubbed in Hindi.

“When it comes to music and movies, 97% of everyone we interacted with watches and listens to everything in Hindi,” the report said.

The opportunity for AI companies is to take one piece of content and modify it for various audiences. Content companies like Pocket FM, Kuku FM, and Stage are realizing that there is value in content produced for the masses and in masses regardless of cultural origins. Previously, it took a lot of money and effort to convert one IP in one language to another. Now, with AI tools, there is a chance of reaching more consumers.

That’s why AI companies that provide tooling to convert one piece of content into different formats and different languages easily become important. For instance, ElevenLabs is a good example, as it helps companies convert text written in one language to audio in multiple languages.

India will always be a country that wants more content. Startups will need to figure out the way to produce (or convert from one form/language to another) more and more content. Either they will have to find ultra-cheap ways to mass-produce content or use AI to do this job. As content increases, AI tools will get more “work,” and we will possibly get to listen to one voice in multiple languages.

Subscribe now