Meta Platform One of the US high-tech giants leads the race to invest in the basic large-scale language models that generator AI and AI breakthroughs rely on, and has also adopted an open-source approach to AI development, allowing them to share technological innovation widely.
These innovations are occurring rapidly. “It’s not just every day, it evolves many times a day,” says Clara Sy, Meta’s Head of Business AI.
Available to developers around the world, Lama LLM has been downloaded over 800 million times. Earlier this week, Meta Chief Product Officer Chris Cox said that the upcoming open source Lama 4 AI will support Power AI agents, the latest trend in generator AI.
AI agents do not respond to prompts alone. They will be able to make new levels of reasoning and action – surf the web and handle many tasks that will be useful to consumers and businesses. According to Shih, Meta’s AI is already being used by over 700 million consumers, and her job is to bring the same technology to businesses.
“Not all businesses, especially small businesses, have the ability to hire these large AI teams, and now they’re building business AI for these small businesses, CNBC’s Julia Boorstin spoke in an interview with CNBC Changemaker’s Spotlight series.
She hopes that inter-company intake will occur soon and spread widely.
“We’re coming straight away where every business, from very large to very small businesses, has a business agent representing it and acting in that voice in that voice. That’s how businesses today have a website and an email address.”
Large companies across the economy sector have invested millions of dollars to develop clients LLM, but “doing flashy things like small tuning models,” “If you’re a small business, if you own a jewelry store, you own a jewelry store online. You distribute through Instagram.
For both consumers and businesses, the meaning of progress discussed by Cox and Shih is important in everyday life.
For consumers, Sea says, “from studying their AI assistants to planning a trip to planning a social outing with friends.”
Rival Openai recently launched Operator AI for tasks such as travel planning.
Business-wise, Sea has pointed to 200 million SMEs around the world who are already using meta services and platforms. “They use WhatsApp, they use Facebook, they use Instagram. They both get customers, but they attract and deepen each relationship. Soon, each of those companies will help express them, automate redundant tasks, speak out loud, find more customers, and provide concierge services to almost customers.”
The more AI there are, the fewer people will have to do, at least in traditional role definitions. Sea says everyone has to prepare now for the change that comes. “No one job has been completely converted by the Internet, mobile or social media. I think it’s clear at the same time as AI, that there are certain occupations where AI can dramatically change jobs.
Her advice for every individual and every company is to “learn, experiment, understand, and almost define what (AI transformation) looks like for a particular job.”
“I had to learn to search just like I had to learn email in 1990. I think today, everyone needs to learn AI, regardless of where they live, or whether they have,” she said.
Good news? According to Shih, AI makes learning easier. “You can talk to it. You can literally talk to it.”
She does it herself. “What I do often is if you have a long research paper, an AI research paper, or development, you can ask AI. You can use meta AI. You can break it down and ask it.
A full interview with SHIH is available on CNBC Changemaker and CNBC’s YouTube channel.
