Since ChatGPT was launched in November, a new small industry has mushroomed, belying the broader slowdown in technology. not a week goes by without someone the unveiling A “generative” artificial intelligence (AI) is underpinned by “foundation” models—the large and complex algorithms that give ChatGPT and other AIs their intelligence. On February 24, Facebook’s parent company Meta released a model called LLaMA. It was reported this week that Elon Musk, the billionaire boss of Tesla and Twitter, wants to create an AI that would be less “awake” than ChatGPT. A catalog maintained by Ben Tossell, a British tech entrepreneur, and shared in a newsletter. Recently, among others, Ask Seneca (which answers questions based on the Stoic philosopher’s writings), Pickaxe (which analyzes your own documents), and Issac Editor (which helps students write academic papers) has evolved to include
ChatGPT and its companion chatbots can be talked about a lot (and they can be talked about: ChatGPT now has over 100m users). But Mr. Tossell’s newsletter hints that the real action in generative AI is moving into all kinds of less chatty services enabled by the Foundation model.
View Full Image
Each model is trained on text, images, sound files or any other pile of data. This allows them to interpret, react to, and create narratives using natural language as well as art, music, and any other type of content found on the Internet. Even as the venture-capital (VC) industry follows a massive hangover after the tech crash that recently paid off for a few years of bubbly, entrepreneurs experimenting with generative AI are attracting investment. I don’t have any problem. In January it was reported that Microsoft poured another $10bn into OpenAI, the startup behind ChatGPT, on top of an earlier $1bn investment. A spreadsheet created by Pete Flint at VC firm NFX now lists 539 generative-AI startups. Excluding OpenAI, they have collectively raised over $11bn in capital to date (see chart). Mike Volpi of Index Ventures, another VC firm, calls it the “Cambrian explosion.”
Several factors are driving this. Although the Foundation model has been around for a while, Mr. Volpi points out that it took a consumer-facing service like ChatGPT to capture the imagination of the world and investors. This happened just as the cryptocurrency crash left venture capitalists frustrated and the empty metaverse looking for the next big thing. Furthermore, even more than web browsers and smartphones before them, foundation models make it easier to build new services and applications on top of them. “You can open your laptop, get an account and start interacting with models,” says Steve Loughlin of Accel, yet another VC firm.
The question for venture capitalists is which generative-AI platform will make it big. For now, it’s a matter of head scratcher in technical circles. “Based on the available data, it is not yet clear whether generative AI will have a long-term, winner-takes-all dynamic,” wrote colleagues and partners at Andreessen Horowitz, another VC firm, in a recent blog post. Many startups provide me-too ideas, many of which are a feature rather than a product. Resource-intensive foundation models may also end up as low-margin commodities over time: although proprietary models such as OpenAI’s GPT-3.5, which powers ChatGPT, are still leading, some open-source ones lag far behind. are not.
Another source of uncertainty is the legal minefield on which generative AI is tiptoeing. Foundation models often make mistakes. And they can derail. The chatbot that Microsoft is developing for its Bing search engine based on OpenAI’s model has insulted more than one user and professed love to at least one other (Sydney, as Microsoft’s chatbot is called, has since been reined in). Generative-AI platforms may not enjoy the legal protections from liability that shield social media. Some copyright holders of web-based content on which existing models are being trained voluntarily, without asking permission or paying compensation, are already protesting. Getty Images, photo repositories, and individual artists have already filed lawsuits against AI art-generators like Stable Diffusion. News organizations whose articles have been plagiarized for information can do the same.
OpenAI is already trying to manage expectations, downplaying the launch of GPT-4 later this year, the highly anticipated new version of the foundational model behind ChatGPT. The VC type’s appetite for generative AI is unlikely to wane. For more risk-averse investors, the safest bet at the moment is on providers of the sufficient processing power required to train and run the foundational model. Share price of Nvidia, which designs chips useful for AI applications, is up 60% so far this year. Cloud-computing services and data-center landlords are also wringing their hands. Whichever AI platform comes out on top, you cannot unfairly sell picks and shovels in a gold rush.
© 2023, The Economist Newspaper Limited. All rights reserved. From The Economist, published under license. Original content can be found at www.economist.com
catch all technology news And updates on Live Mint. download mint news app to receive daily market update & Live business News,