GM Readers! ☀️

Welcome to the 21st issue of Evolving Internet Insights.

Idea we’ve been thinking about: If robots do all the work for us, do we just relax?

Thanks for reading!

Liang and Dan 🙌

🌰 In a Nutshell

• Recently discovered Project Q* is OpenAI’s step towards super intelligence 👟

• Anthropic extends Black Friday and drops their pricing to compete with other LLMs providers and open source models 💧

• Pika Labs, a fledgling six month old startup led by Stanford AI researchers, raises $55M in funding 👶

• Which countries lead in AI tool usage? Which AI tools are the most visited? Find out below 🤓

💾 Byte-sized Stories

This week’s top stories with our insights on top.

1. OpenAI’s Q* (Q-Star) Is a Step Towards Super Intelligence 👟

⚡️ TL;DR: An internal letter written by OpenAI research staff highlighted that OpenAI may be one step closer to artificial general intelligence (AGI). AGI is a form of AI that mirrors or surpasses human level intelligence, ultimately being able to achieve any intellectual task that humans can perform. According to internal sources, the clandestine Project Q* (“Q-Star”) was a major project that set off alarms before Sam Altman’s ouster. To date, Project Q* is an AI model powerful enough to solve math problems it had never seen before – getting closer to what humans refer to as reasoning.

⚡️ So What: After the Altman versus OpenAI board drama the last couple of weeks, some are speculating that the reason Altman was fired was because the board was uncomfortable with his push towards AGI without proper guard rails in place. Project Q* may very well have been one of many examples where the Board disagreed with Altman’s aggressive approach.

⚡️ Zoom Out: Achieving AGI is the current holy grail of AI research. AGI represents an AI system that can apply reasoning to decision-making, mirroring human-level learning and problem solving capabilities. Interestingly, AGI would be able to improve upon itself exponentially and unsupervised. While this pinnacle of AI would mark a marvel in human-led technology progress, many worry that having true AGI poses an existential risk to human society (read: what if an AGI decides to do bad things against humans? “I, Robot” vibes? 🤖🤔).

Read More Here, Here, Here

2. Anthropic Drops Pricing To Compete Against Other LLMs Providers and Open Source 💧

⚡️ TL;DR: Anthropic, one of the leading LLM providers, dropped the pricing of its latest Claude 2.1 model. As competition increases amongst AI firms like OpenAI and Anthropic and the Big Tech companies, Anthropic wants to capture more market share by lowering prices. Also, closed source LLM providers are also facing mounting pressure from open source model providers like Meta.

⚡️ So What: As we covered in ⚡Insights 14, Big Tech companies have yet to make money from their AI products. The WSJ reported Microsoft loses $20 per user per month (and sometimes as high as $80 per user per month) from its GitHub Copilot AI tool. While Big Tech companies can absorb the losses given their other profitable business units, startups like Anthropic have to keep raising money to build and develop costly LLMs that one day they hope to monetize. In September of this year, Anthropic raised $4B from Amazon followed up by another $2B led by Google in October. The key question is can closed source LLM providers like Anthropic create a sustainable competitive moat and monetize their offerings effectively to recoup their substantial R&D investments? In general, we’ve seen competing on price can be a “race to the bottom.” 

⚡️ Zoom Out: Anthropic isn’t only facing competition from other closed source LLM providers. There is a fledgling open source movement in AI (check out 🧠 Brain Food 10 for our deep dive on how AI compounds through open source) that companies like Anthropic have to contend with. The Information reported that some companies are using ChatGPT4 and other higher end models to fine tune open source models that they can use in their business. Clearly, the competition is fierce. ⚔️

Read More Here, Here 

P.S. Our next Brainfood Issue will dive deeper into OpenAI Q* and look at what AGI would be capable of vs the AI of today.

Subscribe to our newsletter to get it straight to your inbox 💌

3. Pika Labs, a Six Month Old Startup, Raises $55M in Funding 👶

⚡️ TL;DR: Pika Labs, a startup building a suite of generative (gen) AI video tools, raised a combined $55M in funding led by Lightspeed Venture Partners and other prominent solo VCs. Pika’s model is capable of making videos in a range of styles like 3D animation, anime, and cinematic. Check out this trailer to get a full sense of what Pika is up to. The company highlighted it has “half a million users, who are generating millions of videos per week.”

 ⚡️ So What: It is one thing to create a generative AI model that can create videos with just some prompting, but it is another thing to create a suite of video editing tools. The former has the initial “wow” factor but loses novelty fairly quickly. The latter, if executed well, could land in an artist's “creative stack.” For example, Pika is building modules on top of its flagship generative AI tool to be able to further fine tune and edit (AI) video outputs.

⚡️ Zoom Out: While many in the creative industries are worried about the impact of AI, Pika is in the camp that its tools will benefit creators. “Our vision for Pika is to enable everyone to be the director of their own stories and to bring out the creator in each of us.” More broadly, new technology tools that are accessible to all of us could form the foundations of entirely new industries. Remember when smartphones with cameras first came out—that eventually led to new user behaviors on modern social media platforms (á la Meta, Instagram, and Snapchat).

Read More Here, Here

📊 Let’s Get Graphic

One visual we couldn’t stop thinking about.

50 Most Visited AI Tools

⚡️ Takeaway: ChatGPT is still the most visited AI tool followed by Character.AI, an AI chatbot platform that allows users to engage with personalized avatars and have more nuanced conversations. After these two, there is a bit of a long tail. Similarly, when looking at the country with the most AI users, the US and India lead in visits to AI tools and then it slowly trails off. While there is a lot happening in this space, it’s still very early.

🐇 Down the Rabbit Hole

Some deeper dives to help you get smarter on emerging tech.

1. The AI Industry’s 50 Most Visited AI Tools: deep dive analyzing the top 50 most visited AI apps, with chock full of trends and data points.

2. Interview with Lan Guan, Accenture’s Chief AI Officer: interview highlighting what one of the largest consulting firms is observing about AI across different industries and companies.

3. How AI Disrupts Business Models: essay by Adobe’s Chief Product Officer explaining how AI disrupts business models, particularly creative and marketing departments.

🗳 We’d Love your Feedback

🙏 Shameless Asks

It takes us days to put this together, sharing it takes you 19.45 seconds 😘

An easy way to support us and the newsletter is through the referral program.

⚡️ Share Evolving Internet Insights with a friend and ask them to subscribe

⚡️ Share on X/Twitter and Linkedin with a short note

⚡️ Share on your company Slack/Teams channels and communities