Don't DIY your AI was the pitch from Salesforce CEO Marc Benioff as 45,000 people flocked to Moscone Center in San Francisco this week for Dreamforce.
All week, Benioff and his team hyped up the company's pivot to powering autonomous agents via a new product called Agentforce. Trailblazers, as the company calls workers across the world who use its software, are now Agentblazers.
"I don't think you can DIY this," Benioff said during his keynote address at the conference. "I think you want a single, professionally managed, secure, reliable, available platform. I think you want the ability to deploy this Agentforce capability across all of these people that are so important for your company and all of these digital and human workers."
The pivot to agents is a bit of a rebrand away from copilots. If an AI copilot is more of a conversational chatbot that helps you manage tasks and workflows to increase your own productivity, an AI agent performs specific tasks for you.
Salesforce began articulating that shift from copilots to agents earlier this year, and Benioff now is positioning Salesforce as a one-stop AI shop.
"We all have struggled in the last two years with this vision of copilots and LLMs and how are we putting it all together? So why are we doing that when we can have Agentforce? We can move from chatbots to copilots to this new Agentforce world, and it's gonna know your business and it can plan, it can reason, it takes action on your behalf," Benioff said during Dreamforce.
Benioff has also taken aim at his competition, particularly Microsoft, over the past few weeks.
"Many customers are so disappointed in what they bought from Microsoft, Copilots, because they're not getting the accuracy and the response that they want," Benioff said during Salesforce's quarterly earnings call in August.
OpenAI, which has gotten most of its billions in funding from Microsoft, was noticeably absent from Dreamforce this year.
Why you'll need an agent. Really
Salesforce's head of AI, Clara Shih, further articulated the company's vision for an agent-filled future during Dreamforce's keynote address.
"In the very near future, agents will be as commonplace as apps and web pages," Shih said, and explained that there are five things that go into making an agent: roles, data, actions, channels, and trust and security.
"We're seeing amazing results. 33% greater accuracy. Two-times greater relevance. And these are according to our customers benchmarks, not ours," Shih continued. "This is the difference between a DIY science project versus a real enterprise grade agent you can constantly deploy into production."
Rules and regulations
California Governor Gavin Newsom jumped into the artificial intelligence conversation during an appearance at Dreamforce. Newsom pulled three bundles of paper out of his suit coat in the middle of a one-on-one conversation onstage with Salesforce CEO Marc Benioff and then signed a trio of laws will regulate deepfakes and election-related content.
Gov. Newsom didn't say whether he would sign another controversial bill, SB 1047, but dropped hints that pointed to concerns. The tech industry is split on its support for the bill, but some in Hollywood recently lined up in favor of it, Politico reported this week.
The governor acknowledged some of Hollywood's concerns when he signed another pair of bills that aim to protect entertainment workers from being exploited by the use of AI and unauthorized use of their likeness, NBC News reported.
Full story: Governor Gavin Newsom signs three AI bills at Dreamforce
In the span of one week, Meta admitted to Australian authorities that it has scraped user data going all the way back to 2007 for AI training purposes and then warned European Union regulators that its data and technology rules were too strict.
Users in the E.U. have the ability to opt-out of data collection because of the bloc's strict user privacy and data laws. In July, Brazil banned Meta from using local user data for training AI systems, AP News reported. But Australian users don't have that option, nor do users in the U.S. where there's no federal data privacy laws.
California passed its own data privacy law in 2018 called the California Consumer Privacy Act. It established rules around giving users transparency, the right to delete personal data and the right to opt-out of letting companies sell their data. But it doesn't explicitly mention AI. California lawmakers introduced a bill earlier this year known as AB 2877 that would have added restrictions for training AI systems with data from users under 16. The bill appears to be stalled in the Assembly.
Data Deals
Salesforce and Nvidia will work together to develop AI-powered autonomous agents and avatars, the WSJ reported. The move aims to bolster Salesforce's new Agentforce platform.
T-Mobile and OpenAI announced their own deal this week to develop autonomous customer service agents for the telecommunications corporation, Axios reported.
Entertainment giant Lionsgate will allow New York-based startup Runway to use its content for training generative AI models, according to the Hollywood Reporter.
Real Estate Shuffle
OpenAI just gobbled up more real estate in San Francisco. The company leased 315,000 square feet of space in Mission Bay, adding to the nearly half a million square feet it already has in the neighborhood.
Full story: OpenAI signs another massive lease in San Francisco's Mission Bay
Funding rounds to know
Salesforce is doubling down on artificial intelligence by pouring another $500 million into its corporate venture fund, known as Salesforce Ventures, which now has $1 billion for its generative AI investments.
I sat down with the fund's Managing Partner Paul Drews at Dreamforce this week to chat about the investing strategy at Salesforce Ventures, whether the continued AI hype is justified and why bigger isn't necessarily better when it comes to foundational models.
Read the highlights from our conversation here: Salesforce Ventures is comfortable with the AI hype
OpenAI is on the verge of closing a new funding round that would net the company $6.5 billion and boost its valuation to at least $150 billion, close to double its previous $86 billion valuation, Bloomberg News reported. Lead investor Thrive Capital is reportedly putting down $1.5 billion. Other investors that are expected to join the round include Microsoft (its largest existing investor), as well as newcomers Nvidia and Apple.
Must reads: Dreamforce Dispatches
Salesforce’s New AI Strategy Acknowledges That AI Will Take Jobs -Bloomberg
Salesforce CEO Marc Benioff says he uses ChatGPT as a therapist -SF Standard
Dreamforce 2024 touts itself as 'world's biggest AI event,' but OpenAI is nowhere in sight -San Francisco Business Times
Oppenheimer’s grandson: AI is ‘probably not’ as big a risk as nuclear weapons -SF Standard
Plane flies over Dreamforce calling out SF conference's main attendees -SFGate
What San Francisco’s streets look like as Dreamforce crowds descend downtown -SF Chronicle
Comedian John Mulaney brutally roasts SF techies at Dreamforce -SF Standard
Must reads: Beyond Dreamforce
OpenAI's safety committee to oversee security practices as an independent body -Reuters
Altman previews big leaps ahead for OpenAI's Strawberry -Axios
The new followup to ChatGPT is scarily good at deception -Vox
Copilot Pages is Microsoft’s new collaborative AI playground for businesses -the Verge
Microsoft, BlackRock to launch $30 bln fund for AI infrastructure -Reuters
Three Mile Island is reopening and selling its power to Microsoft -CNN
YouTube announces AI features from Google DeepMind for Shorts creators -CNBC
Amazon releases a video generator — but only for ads -TechCrunch
Billionaire Larry Ellison says a vast AI-fueled surveillance system can ensure 'citizens will be on their best behavior' -Business Insider
Biden administration to host international AI safety meeting in San Francisco after election -ABC News
A.I. Pioneers Call for Protections Against ‘Catastrophic Risks’ -NYT
An AI chatbot helped Americans who believe in conspiracy theories “exit the rabbit hole” -Nieman Lab
Like digital locusts, OpenAI and Anthropic AI bots cause havoc and raise costs for websites -Business Insider