Companies want AI to be easy
Nvidia bets on agentic AI for enterprises
Why the US is restricting AI chip exports
See 2025 clearly with BILL + free Ray-Bans
When you’re at the starting line of 2025, you need a pair of Ray-Ban Smart Glasses to stay focused!
We’ve got you covered with a free pair on us when you take a 30-minute demo of BILL Spend & Expense.1
Choose BILL Spend & Expense to:
Automate expense reports so you can focus on strategy
Set budgets across teams to better control expenses
Get real-time insights into your company spend
Access scalable credit lines from $50K to $5M1
Ready to hit the ground running and see clearly in 2025?
1 Terms & Conditions Apply. See offer page for more details. The BILL Divvy Card is issued by Cross River Bank, Member FDIC, and is not a deposit product.
***
Companies want AI to be easy
For businesses adopting AI, simplicity is king—even at a premium. Companies are prioritizing AI tools like OpenAI's ChatGPT and Anthropic’s models that work seamlessly out of the box, even if they come with fewer security features than traditional enterprise software. This preference has fueled explosive growth: both OpenAI and Anthropic saw revenues soar by more than 500% last year, outpacing a growing array of free, open-source alternatives from Meta Platforms, Mistral AI, and others.
Just a year ago, enterprise software leaders like Databricks and Snowflake were betting their customers would gravitate toward open-source models. These models offered greater flexibility for customization and security while delivering performance nearly on par with proprietary solutions like GPT-4. The idea was that businesses could fine-tune models to align with their specific needs, such as adapting to niche industry jargon.
However, the reality is that most businesses aren’t keen on the complexities of customizing large language models (LLMs). "It’s a little bit too hard," explained Naveen Rao, head of Databricks’ generative AI division. "You have to do a lot of data engineering and data cleaning to get a model that works really well. That’s a big step for most customers to take." This shift in sentiment has prompted Databricks and Snowflake to largely abandon efforts to build their own open-source generative AI models.
Still, open-source AI continues to find its champions. Meta’s Llama models, for instance, have been downloaded over 770 million times, highlighting what Meta spokesperson Ashley Gabriel described as "the transformative impact of open-source innovation." But adoption has been uneven. Llama has struggled to gain traction with Amazon Web Services customers, and the latest version saw fewer downloads than its predecessors. For many, OpenAI’s polished solutions remain the more convenient choice. Established software providers are embedding AI features into existing platforms, banking on customer loyalty and ease of integration....
***
Hire Ava, the AI SDR & Get Meetings on Autopilot
Ava automates your entire outbound demand generation process, including:
Intent-Driven Lead Discovery
High Quality Emails with Waterfall Personalization
Follow-Up Management
Free up your sales team to focus on high-value interactions and closing deals, while Ava handles the time-consuming tasks.
Book a demo to see how Ava can 10x your outbound.
***
Nvidia bets on agentic AI for enterprises
At the Consumer Electronics Show, Nvidia CEO Jensen Huang and CFO Colette Kress took the stage to delve into one of the event’s hottest topics: agentic AI. Huang described the technology as an advanced form of AI capable of reasoning step by step, retrieving and analyzing information, contextualizing it, and generating answers. “Agentic AI can work on things for thousands of tasks” Huang said, underscoring its potential for complex, sustained work.
Enterprise AI has long been a strategic focus for Nvidia, which collaborates with Fortune 100 companies on AI initiatives across a range of industries. During the session, Huang highlighted Nvidia’s latest innovation, AI Blueprints, designed to help businesses build agentic AI applications that automate workflows. “The mental model for AI and enterprise is really AI agents,” Huang explained, envisioning these agents as a new layer of technology that operates tools on behalf of users. “That layer has never existed before,” he said. “That’s why AI is a growth industry.”
Huang revealed that Nvidia is also partnering with software companies to develop AI agents that leverage existing tools and offer them as services to their clients. He emphasized the transformative impact of AI assistants in the workplace, noting, “There’s a billion knowledge workers. Everybody’s going to have AI assistants.” He issued a stark prediction for businesses: “Starting next year, if a software engineer at your company is not assisted with an AI, you are losing already, fast.”
During a fireside chat at CES, Kress expanded on the potential applications of agentic AI in enterprise settings. She cited examples like automating call center operations, enhancing fraud detection and testing, and improving risk management. “As Nvidia continues to help enterprises design their future, agentic AI is going to be an important part of that for many of them,” Kress said.…
***
Why the US is restricting AI chip exports
On January 13, the Biden Administration announced new restrictions on the export of advanced computer chips used to power artificial intelligence (AI). The move follows two significant developments over the holiday season that sent shockwaves through the global AI community and raised fresh concerns about technological competition with China.
First, OpenAI unveiled its latest model, o3, which achieved an unprecedented 88% score on a challenging set of reasoning tests where no prior AI system had surpassed 5%. The breakthrough prompted Francois Chollet, a former AI researcher at Google and longtime skeptic of near-term “artificial general intelligence” (AGI), to acknowledge a paradigm shift. “All intuition about AI capabilities will need to get updated,” he remarked.
Second, Chinese AI company DeepSeek stunned researchers by releasing an open-source AI model that outperformed all American open-source language models, including Meta’s Llama series. The achievement defied expectations, as U.S. officials and industry experts had long believed China lagged behind in AI development due to a U.S.-led embargo on advanced AI chips. DeepSeek’s success, despite restricted access to cutting-edge hardware, signaled a significant leap in Chinese AI capabilities. These developments have reignited debates about the timeline for achieving AGI—a level of AI powerful enough to conduct economically valuable work and make autonomous scientific discoveries. “I think AGI will probably get developed during this president’s term,” OpenAI CEO Sam Altman told Bloomberg.
At the same time, DeepSeek’s progress highlighted China’s potential to compete directly with the U.S. in the race to AGI. For some U.S. policymakers, the realization underscored the urgency of curbing China’s access to AI-enabling technologies. “Restricting China’s access to AI is essential for U.S. national security,” one official said, echoing a sentiment that has gained traction in Washington.…