Generative AI has already shaped up to be a fiercely competitive consumer-facing space, one in which AI tech valuations soared, as well as one that creates new demands from talent like how to write good prompts for a large language model.
Though the tech giants have generated most of the early headlines, what has been missed that you, as a marketer, need to be checking out is what’s going on in both costs (dropping dramatically so access is easier) and in the part of the marketplace that is seeing the most interesting developments if the least press attention–the smaller emerging players that you might not have have heard of have lots of new products to sell.
In 2022, investors poured $4.5 billion into 269 generative AI deals globally and YCombinator’s class of 2023 AI startups is the biggest ever. Through the rest of 2023, we should be seeing a faster pace of investments in new players. This will be even more important if, as the latest predictions have it, the economy skirts a recession and, as a result or cause, digital advertising grows.
There are many startups, some of which haven’t gotten the attention they deserve as they are doing incredible things for marketers and disrupting the marketplace with fierce competition.
Scale AI, a new ecommerce startup, is offering a full-stack AI platform for any enterprise. Shopify is also using generative AI in commerce for product descriptions. Alison is an a16z funded startup out of Israel that analyzes creative work, such as typeface characters, colors, sounds and text. It then gives the members of the creative team a data-driven creative brief prompt that they can use to feed into a generative AI image production platform like Midjourney or Stable Diffusion. A new app called Roll takes the work out of editing video (and recreates dolly shots and other cool effects) with AI. A startup called Jasper just raised $125 million to help marketers auto-generate promotional blog posts, photos, press releases and other marketing materials. Replikr uses generative AI to create customer-service avatars, and Musico creates original copyright-free advertising jingles. A video editor called Veed automatically takes the umms and ahhs out of video and picks your best takes for you. And GoCharlie.AI has a generative AI marketing assistant designed to create promotional social-media posts. And that’s just the tip of the iceberg.
AI models are already impressively capable when it comes to generating music and mimicking human voices. In music, generative AI is likely to increasingly become a valuable tool for songwriters and composers, creating novel compositions that can serve as inspiration or encourage musicians to approach their creative process in new ways. We are also likely to see it being used to create real-time, adaptive soundtracks – for example, in video games or even to accompany live footage of real-world events such as sports. AI voice synthesis will also improve, bringing computer-generated voices closer to the levels of expression, inflection and emotion conveyed by a human voice. This will open new possibilities for real-time translation, audio dubbing and automated, real-time voiceovers and narrations.
Of course, this also changes the game for advertising on media properties. With gen AI, instead of getting a search result and then navigating to the site that delivered it, a synopsis of what a consumer wants to know will be summarized right on page or in the chat window. There’s no need to click for more—it’s all there. But that same consumer might click on related ad links that are strategically presented instead.
Efficiency in scale and the ability to generate everything from translations without human intervention to automated ad personalization and dynamic responsiveness to the on-the-fly effectiveness improvements are possible through so many new offerings.
At the same time, across the industry, competition has meant that the most important costs of AI–training models, running AI models (i.e., AI inferences) and hardware needed to power AI computation–have been dropping dramatically and they continue to fall rapidly.
As has been our experience throughout the tech revolution, as the technology improves, it becomes more efficient, with drastic reductions of required compute power. Just for one example, AI training costs have been dropping at an annual rate of 70% since 2020. A model that cost $4.6 million to train in 2020 is expected to drop to $30 by 2030!
It made sense at first that the main search and advertising players would be among the first to use gen AI because the technology wasn’t cheap. But now with the vibrant competition all around, it’s anyone’s game as the prices have continued to drop across the board–and will be, by 2030, a small fraction of what they once were.
Gen AI is following the usual pattern of new emerging capabilities, opening up new areas of fierce competition in tech from which the market has shown, again and again, the ability to generate economic advantages as the spinning wheel of winners and losers goes round and round, not needing tech-wanna-be bureaucrats to interfere as the tech world changes all on its own.