Venture capitalists are rushing to invest in artificial intelligence start-ups as growing hype about generative AI fills the void left by failing cryptocurrency and blockchain ventures. The leap in developing programs that can write scripts and create art in seconds has driven a surge of investor interest, creating a rare bright spot in a start-up landscape dominated by tumbling valuations and job cuts. In five days, ChatGPT surpassed 1mn user and was praised by Elon Musk, a co-founder of OpenAI who left the board in 2018. Moreover, the freedom to tinker with such a powerful AI has sparked start-up ideas for countless investors and entrepreneurs.
Venture capital investment in generative AI has increased 425 per cent since 2020 to $2.1bn this year, according to data from PitchBook, even as the broader tech market falls. One entrepreneur said that after discussing fundraising with just three investors, he had been inundated with offers from more than 20 others, securing funding after a week of whirlwind meetings. Jasper, an «AI copywriter» for marketers, raised $125mn at a $1.5bn valuation. In contrast, one of the companies behind image-generation tool Stable Diffusion, London-based Stability AI, raised $101mn in a move that saw it reach unicorn status. This week another Stable Diffusion developer, Runway, raised $50mn.
When Cristóbal Valenzuela co-founded Runway four years ago, investors told him generative AI «is not a thing». In September, Sequoia partners co-wrote an investment thesis using GPT-3 software, saying AI could produce final drafts of writing better than the human average, generate code on a commercial scale, and make drafts in images and gaming in the next two years. «Generative AI is well on the way to becoming faster and cheaper, but better in some cases than what humans create by hand,» Sequoia's analysis said. Due to the high costs of running the programs and storing data that the AI programs learn from, large language models have been the reserve of groups such as Microsoft, Google and Facebook.
But OpenAI has made its tech available through an application programming interface, offering any company access to its capabilities. In 2019, it became a for-profit enterprise, followed by Microsoft's $1bn deal, including using its Azure platform to conduct experiments. Under the deal, Microsoft gets a first shot at commercialising early results from OpenAI's research. Microsoft's focus on OpenAI came as part of a push to claw back an edge in AI following heavy investment by Google in using the tech for search and speech, as well as acquiring UK AI company DeepMind for about £400mn in 2014.
«One of the big historic advantages corporates had is access to large, typically proprietary data sets. So they've been able to use these data sets to train up large, larger and larger models,» said Stacey. The cost of running ChatGPT is estimated to be a few cents per chat, according to Sam Altman, boss of OpenAI. Bloc Ventures, a UK deep tech VC firm, focuses investment on the tech that enables high cloud levels, seeking to reduce the costs and energy used in generative AI.
«Companies are chasing net zero, and the luxury of having chat-bots we can talk to through AI is burning a hole through the earth in a data centre,» said David Leftley of Bloc Ventures. New York-based Runway is doing the primary AI research to build models and turning them into a suite of image generation and collaboration tools already used by companies including Publicis, Google and CBS. «There's a lot of companies building on top of existing APIs. Our bet long-term is you need to own your stack, you need to own your technology, to allow you to more quickly and more easily change if needed,» said Valenzuela. He said AI start-ups such as Runway could outmanoeuvre the Big Tech companies.
By making ChatGPT open-source and available to the public, OpenAI can collect more data to train its large language models and iron out bugs. One limitation of the tech, and similar AI tools, is «hallucinations», where the program gives an inaccurate answer and struggles with simple maths. OpenAI says GPT «sometimes writes plausible-sounding but incorrect or nonsensical answers», among other limitations, leading many to say it needs human intervention before being embedded in businesses. «There are a lot of questions around how commercially viable these models and capabilities are,» said Lisa Weaver-Lambert, Microsoft private equity, data and AI lead.
Comments
Post a Comment