In partnership with
From Hype to Production: Voice AI in 2025
Voice AI has crossed into production. Deepgram’s 2025 State of Voice AI Report with Opus Research quantifies how 400 senior leaders - many at $100M+ enterprises - are budgeting, shipping, and measuring results.
Adoption is near-universal (97%), budgets are rising (84%), yet only 21% are very satisfied with legacy agents. And that gap is the opportunity: using human-like agents that handle real tasks, reduce wait times, and lift CSAT.
Get benchmarks to compare your roadmap, the first use cases breaking through (customer service, order capture, task automation), and the capabilities that separate leaders from laggards - latency, accuracy, tooling, and integration. Use the findings to prioritize quick wins now and build a scalable plan for 2026.
Hey, josh here. This one surprised me.
Yann LeCun Just Walked Away From Meta—And It Tells You Everything About AI's Civil War
Listen, when a Turing Award winner leaves a company after 12 years, you pay attention. When that person is Yann LeCun—the guy who basically invented modern AI—and he's walking away from Meta to start his own thing? That's not a departure. That's a statement.
Here's what happened: LeCun announced he's leaving Meta to launch an independent startup focused on what he calls Advanced Machine Intelligence. Meta stays on as a partner, which sounds cordial enough. But peel back one layer and you see the real story.
In April, Meta's Llama 4 flopped. Zuckerberg panicked. By June, he'd dropped $14.3 billion—10% of Meta's annual revenue—on Scale AI and brought in its 28-year-old CEO to run a new "Superintelligence Labs." LeCun, the Turing Award winner who built Meta's legendary FAIR research lab? He got put under the new guy's authority.
Then came October's massacre: 600 AI positions cut, mostly from FAIR. The lab LeCun built was gutted while Meta doubled down on a $600 billion bet on scaling large language models.
Here's the kicker: LeCun thinks that entire bet is wrong. He's been saying it loudly—LLMs are just pattern-matching machines that can't reason like humans do. A four-year-old absorbs more real-world data through their senses than these models train on through text. His alternative? "World models"—AI that learns by understanding how the physical world actually works.
So now we've got one of AI's founding fathers raising what could be $200 million to prove the entire industry is heading in the wrong direction.
The question isn't just whether LeCun is right. It's whether a 12-year legacy and a Turing Award matter when a CEO decides the company needs to move faster.
Apparently not.