Generative AI - The State of the Game
We are now six months into the generative AI revolution, so let’s check the state of the game - the winners, the losers, the new players and who is the black horse.
ChatGPT was released on November 30th, 2022, and since then the world has not been the same. Artificial Intelligence has become one of the most discussed topics in the world seemingly overnight.
We are now six months into the generative AI revolution, so let’s check the state of the game.
The Winner
As I am writing this in May 2023, there is one clear winner in the generative AI game - OpenAI.
OpenAI was the first to introduce a Large Language Model (LLM) text-generating chatbot to the public and took the scene by storm.
ChatGPT has now over 100 million users and it reached that milestone in just 2 months. That’s a rate of growth not seen ever before.
According to SimilarWeb, openai.com is now the 16th most visited website on the entire internet and chat.openai.com is more popular than bing.com.
Another thing worth acknowledging is that OpenAI and ChatGPT have become cultural phenomenon. Everyone is talking about it. It is in the news and it is the poster child of the AI revolution.
In comparison to ChatGPT, Google Bard does not exist in Google Trends, apart from two spikes in interest. The first one was on February 7th - the day after Google announced Bard as a response to ChatGPT. The second spike is on March 22nd when Google opened Bard for public beta.
If you are wondering what are those drops in ChatGPT’s line that create the seesaw pattern - those are weekends.
ChatGPT and OpenAI do not have any competitors at the moment. Everyone is trying to catch up.
The New Players
OpenAI’s success created an entire ecosystem of startups building new products and services on top of ChatGPT or using generative AI. And with that success and hype came an explosion in funding.
According to PitchBook, Q1 2023 has seen $1.68B in investments into AI. But even here the biggest winner is, again, OpenAI, which scored a $10B investment from Microsoft at the beginning of the year, bringing its value to $27B-$29B.
Some notable new players that emerged in the generative AI space are Midjourney (a very popular and highly-capable text-to-image generator), Runway ML (which offers generative AI models for images and video, valued at $500 million), Stability.ai (the creators of the text-to-image model Stable Diffusion, which according to Fortune looks to raise funds at $4B valuation), HuggingFace (which creates a platform for machine learning and has recently raised $100 million in Series C) and Anthropic (AI research startup valued at $5B).
We also see the emergence of products and services supporting the development of applications using Large Language Models. Good examples here are LangChain (which is used in building LLM chatbots and AutoGPTs) and Pinecone (a vector database used to store ML models).
According to Crunchbase, LangChain has raised $10 million in a seed round. Pinecone has recently closed a $100 million Series B investment round on a $750 million post-valuation.
The Losers
The biggest loser in the generative AI game is Google. For a long time, Google and its subsidiary, DeepMind, were the leaders in AI. That position has been taken from them overnight by OpenAI and Microsoft. They are now leading the game and Google has to catch up.
Google’s position as the leader in web search and the ad business built around it is now threatened by ChatGPT and by Microsoft, which is incorporating new AI features (powered by OpenAI) into its products. Bing already offers a generative AI image generator powered by OpenAI’s DALL-E. Both Bing and Edge have a conversational chatbot assistant built-in. And soon AI-powered features will be available in Office 365 and Teams.
Google responded by announcing its own chatbot called Bard. At the moment, Bard is marked as an experiment and is only accessible via a waitlist. Google also announced new AI features coming to Workspace, similar to what Microsoft is adding to Office.
What Google does not have in its offering is a business chat app. Microsoft has Teams and it, combined with the entire Office package, is a powerful addition to Microsoft’s offer for businesses. With incoming AI features, Teams can become a true workplace assistant. Google does not have anything like that at the moment. It is a weakness which might be exploited by Slack which announced Slack GPT - its own generative AI features.
Google will most likely present new AI services during this year's Google I/O event which might include the reveal of Magi - an answer to ChatGPT in Bing and other AI products and services in response to OpenAI and Microsoft.
Amazon joined the game with the introduction of generative AI services on AWS. Amazon will offer Large Language Models powered by AI21 Labs and Anthropic, and text-to-image generators powered by Stability.ai.
Meta also joined the party with a rather vague idea of introducing “AI agents to billions of people in ways that will be useful and meaningful” as Mark Zuckerberg said to the investors. But Meta’s biggest “contribution” to the generative AI space came in an unintentional way (we will discuss that later).
The Others
The US companies are not the only ones developing LLM chatbots.
China has banned access to ChatGPT in favour of local products and services. Baidu was the first Chinese company to unveil their chatbot - Ernie - to a mixed reception. If you want to see how Ernie compares to ChatGPT,
published a transcript and a translation of a video comparing both services.A couple of weeks after Baidu, Alibaba showed their answer to ChatGPT - Tongyi Qianwen.
The Emergent AI Superpower
Microsoft has a deep partnership with OpenAI and can benefit from having its brand and products associated with OpenAI, ChatGPT and GPT-4.
Microsoft is in a unique position to offer AI services to a broad spectrum of users and customers.
For everyday users, they offer chatbot and image generation services with Bing and Edge.
Businesses can benefit from integrating AI chatbots and tools into Office and Teams, making their employees more efficient.
If you need to build a custom app using Large Language Models, you can use Azure OpenAI services (which also powers OpenAI itself).
Microsoft also indirectly offers AI assistants to programmers with GitHub Copilot (Microsoft owns GitHub). Copilot uses OpenAI services under the hood.
If Microsoft plays its cards wisely, it can become the AI superpower that offers AI services for consumers, professionals and businesses, while at the same time powering OpenAI - the most popular chatbot assistant right now - and other companies with cloud services.
The Black Horse
But there might be another player on the horizon.
Last week,
published an internal Google document which points out the massive progress the open-source community has done. As the anonymous Google employee writes:While our models still hold a slight edge in terms of quality, the gap is closing astonishingly quickly. Open-source models are faster, more customizable, more private, and pound-for-pound more capable. They are doing things with $100 and 13B params that we struggle with at $10M and 540B. And they are doing so in weeks, not months.
It all started when LlaMa - Meta’s Large Language Model - leaked and got into the hands of the open-source community. Just a week later after the leak, someone figured out how to optimise Meta’s model to run only on CPU and put the code on GitHub. And from there the ball started to roll. People showed how easy it is to have ChatGPT-like software running not only on laptops but also on smartphones and even on Raspberry Pi.
One of the most hyped LLM derivates - autonomous LLM agents, or AutoGPT in short - also came from the open-source community.
A smart move, according to the leaked document, would be for the tech giants to embrace open-source. No matter if they do that or not, we can expect constant innovation coming from the open-source community.
The Future
Where we are now is just the opening phase of the game. Right now, OpenAI and Microsoft have the initiative and everyone else has to catch up.
We will see new AI-powered services and features rolling out in the coming months. Both Microsoft and Google are going to release chatbot assistants integrated in their browsers, search engines and business applications. There are new products or services using generative AI being released every week. Everyone else asks not if but how to include chatbots in their businesses.
One thing is for sure - we are going through a massive transformation in how we interact with computers. A transformation comparable to the introduction of the internet, smartphones, and social media, or maybe even bigger than all of those combined.
Subscribe to H+ Weekly to support the newsletter under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay for a subscription will support access for all).
A big thank you to my paid subscribers and to my Patreons: whmr, Florian, dux, Eric and Andrew. Thank you for your support!
I like your topic choice and writing style, would you be interested in writing a guest post for my newsletter? Let me know if you do in a LinkedIn direct message with the link to your newsletter. I have an idea to pitch you on the topic.
The main reason I offer guest posts is to help new writers on substack or smaller publications get some traffic.