This is an excerpt of Sources by Alex Heath, a newsletter about AI and the tech industry, syndicated just for The Verge subscribers once a week. Reinforcement learning (RL) is the next frontier, Google is making significant strides, and the party scene has truly escalated. These were the main themes at this year’s NeurIPS in San Diego.
NeurIPS, or the “Conference on Neural Information Processing Systems,” started in 1987 as a strictly academic event. It has since grown, fueled by AI hype, into a major industry gathering where labs come to recruit, and investors look for the next big AI startup.
Although I wasn’t able to attend NeurIPS this year, I wanted to catch up on the conversations happening in San Diego over the past week. So, I reached out to engineers, researchers, and founders to get their insights. Among those I spoke with were Andy Konwinski, cofounder of Databricks and founder of the Laude Institute; Thomas Wolf, cofounder of Hugging Face; OpenAI’s Roon; and attendees from Meta, Waymo, Google DeepMind, Amazon, along with several others.
I asked everyone the same three questions: What was the buzziest topic from the conference? Which labs feel like they’re thriving or struggling? Who threw the best party?
The consensus was quite clear. “RL RL RL RL is taking over the world,” said Anastasios Angelopoulos, CEO of LMArena. The industry seems to be rallying around the concept that fine-tuning models for specific applications—rather than just scaling data for pre-training—will drive the next AI breakthrough. When it comes to lab momentum, it’s evident that Google is in the spotlight. “Google DeepMind is feeling good,” Wolf from Hugging Face noted.
The party scene was, of course, non-stop. Konwinski’s Laude Lounge emerged as one of the week’s favorites, attracting notable figures like Jeff Dean, Yoshua Bengio, Ion Stoica, and around a dozen other top researchers. Model Ship, an invite-only cruise with 200 researchers, boasted “an unprecedented commitment to the dance floor for a conference event,” according to organizer Nathan Lambert. Roon had a more reserved take on the festivities: “You can learn more from Twitter than from literally being there… mostly I felt like ‘this is too much.’”
Here’s what attendees had to say about NeurIPS this year:
What was the buzziest topic among attendees that you think more people will be talking about in 2026? Which labs feel like they’re surging in momentum, and which ones feel more shaky? What was the best party you attended or had FOMO over? Yes, some attendees even considered keynotes to be parties. It seems academia still holds its ground at NeurIPS after all.
