Why AI Leads to More Work, Not Less

**Nathaniel Whittemore** (0:00)
Today on the AI Daily Brief, why AI power users are actually working more? Before that on the headlines, is this the best AI video model yet? The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI.
All right, friends, welcome back to the AI Daily Brief Headlines edition, all the daily AI news you need in around five minutes. One of the ongoing conversations when it comes to the China-US AI race is not just how far behind China is, but when and if it will cross the barrier of actually being able to innovate ahead of the US rather than just catching up quickly. For some, the release of a new video model suggests that that threshold has been crossed. TikTok parent company ByteDance has surprised with a new video model that absolutely seems to push the state of the art. The model, called SeedDance 2.0, was released without fanfare on Monday. The early demos are fairly incredible. Menlo Ventures' Didi Das presented a series of examples in a thread capturing numerous styles. And there really is range here. There is a Pixar scene, a product launch video with not just coherent text and animations, but actually impressive graphics, a Goku cartoon scene and many more. Didi wrote, China's ByteDance just dropped the most advanced video generation model in the world. See, Dance 2 has native audio gen, drastic step up from VO 3.1 and Sora 2 in quality, supports multimodal input, 2K resolution, goes beyond cinematic video and can do product demos as well. And it's really hard to tell it's AI. In addition, it appears the model is capable of generating 15 second clips with multiple cuts. Ray Diao, a former Google senior engineer commented, What actually sets this apart is native audio visual cogeneration. Competitors handle audio and post production, but ByteDance generates it alongside the video. And this is one of, if not the first time, that Chinese models have added sound. Watching some of the videos, the perfect lip sync and immersive sound are part of what make the new model stand out. After taking the model for a test drive, 36KR wrote, The original sound experience is truly different from added voiceover. It shows that AI is not just creating pictures. It understands what's happening in the picture and knows what sound should be made in that environment. This is quite interesting. Their review noted amazing character consistency, fantastic physics and the ability to prompt the model to storyboard across multiple cuts. They dinged the model for dialogue quality, but overall, it was a minor gripe. Alongside the model, ByteDance included their own interface to make getting started much easier. Previous Chinese video models have typically been API only, making access a little more limited. Right now, there are examples just absolutely flooding the internet. And I wouldn't be surprised if this meaningfully increases the timeline of when we see the next Vio or Sora. Now, moving to a totally different topic. Back to the US and to data center politics. The White House is pushing AI firms to sign a pact on community protections around data center development. Politico reports that the Trump administration is seeking commitments from tech giants on principles for the AI buildout. They obtained a draft document spelling out the agreement, citing two anonymous administration sources. The pact is designed to ensure that data centers do not raise household electricity prices, strain water supplies or undermine grid reliability. Primarily, the tech companies will be pledging to bear the full cost of infrastructure upgrades and new power generation required to support their data centers. The administration is said to be planning to roll out the agreement in a splashy White House event which is yet to be announced. Now administration officials speaking on the record said that the draft pact was outdated but declined to provide details of changes. Over in our SaaSpocalypse watch, monday.com is the latest victim. Monday fell by 21% on Monday after the company issued weak guidance as part of their full year earnings report. Revenue guidance was between $338 million and $340 million for the current quarter, falling slightly short of analyst expectations of $343 million. Income also fell way short of expectations. Like we've seen from some of these other stocks that have been hit, it's not that the report itself was terribly bad. Revenue grew by 25% over the past year. But their 2026 revenue forecast was cut by a third since their investor day last summer, and the company withdrew 2027 guidance entirely. While co-CEO Aaron Zimman tried to comment that, quote, we don't see any impact currently from any AI company, and we're shifting our product regardless to be more AI native, investors clearly weren't buying it. Overall, Monday's stock is down more than 45% this year. Now it doesn't help that for many, monday.com is the poster boy for a company that's set up for AI disruption. Indeed last week, as part of her coverage of the software crash, CNBC reporter Deirdre Bossa tried to recreate the platform to demonstrate that vibe coding isn't quite there yet, and was shocked to discover that Claude Cowork managed to deliver a functional duplicate that suited her needs in under an hour. I continue to believe that while the magnitude of the sell-off may be exaggerated, markets in this case are sniffing out something fairly important. Does that mean though that SaaS is dead, or that the model of what a SaaS company is, is changing so radically we simply don't have the shape of it yet? That's certainly more in line with the example of Databricks. This week, Databricks announced their next tranche of fundraising, and released some serious revenue numbers alongside. The company's revenue run rate is up to $5.4 billion, which is up 65% year-over-year. The fundraising round saw Databricks gather $7 billion in fresh capital across debt and equity. CEO Ali Godzi framed this round as a rebuke of the death-of-software narrative, but only for companies willing to make the AI-first transition. He said, Everybody's like, oh, it's SaaS, what's AI going to do with all these companies? For us, it's just increasing the usage. Of their $5.4 billion in ARR, a quarter is attributed to Databricks' AI products. The company started making the transition in 2024, recognizing that they needed to build an agentic stack on top of their database product. To that end, they went on an acquisition spree targeting companies that specialize in agent-compatible data discovery. Databricks now has two core product lines addressing the AI transformation and is making the bet that companies won't rip out their SaaS in favor of an in-house-coded solution. At the same time, they are betting that agentic UXs will completely change the SaaS business, eliminating the need for clunky frontends or technical skills to query a database. He believes that the big risk for SaaS companies will be clinging to their legacy UX while everyone else goes agentic. One really crazy statistic that was released, again shared by Deirdre Bossa, she writes that 80% of databases on Databricks' platform are being built by AI agents, which means, as she points out, AI is building more enterprise software than humans are. So does all this excitement around Databricks mean that they are going to be added to the list of potential AI IPOs this year? CEO Ali is reading the room, and the answer seems pretty clear. Now he said it's not a great time to go public.