This is a fragment Sources: Alex Heatha newsletter about artificial intelligence and the tech industry, distributed once a week only to The Verge subscribers.
Reinforcement learning (RL) is the next frontier, Google is on the rise, and the party scene is completely out of control. These were the results of this year’s edition of NeurIPS in San Diego.
NeurIPS, or “Conference on Neural Information Processing Systems,” began in 1987 as a purely academic event. Since then, the event has grown with the AI buzz and turned into a massive industry event, with labs flocking to recruit and investors flocking to find the next wave of AI startups.
Unfortunately, I wasn’t able to attend the NeurIPS conference this year, but I still wanted to know what people were talking about in San Diego over the last week. So I asked engineers, researchers, and founders for their takeaways. The list of responses below includes Andy Konwinski, co-founder of Databricks and founder of the Laude Institute; Thomas Wolf, co-founder of Hugging Face; Roon from OpenAI; and participants from Meta, Waymo, Google DeepMind, Amazon, and several other places.
I asked everyone the same three questions: What is the hottest topic of the conference? Which labs feel like they are growing or struggling? Who had the best party?
The consensus was clear. “RL RL RL RL is taking over the world,” Anastasios Angelopoulos, CEO of LMArena, told me. The industry is coalescing around the idea that tuning models for specific employ cases, rather than scaling the data used for initial training, will drive the next wave of AI progress. It’s clear from the question about lab dynamics that Google is having a moment. “Google DeepMind feels good,” Wolf from Hugging Face told me.
The party flow was naturally relentless. Konwinski’s Laude Lounge turned out to be one of the week’s scorching spots, with visits from Jeff Dean, Yoshua Bengio, Ion Stoica and a dozen other top researchers. Model Ship, an invitation-only cruise with 200 researchers, had “a level of engagement on the floor that you don’t see at any conference,” one of the cruise organizers, Nathan Lambert, told me. Roon commented dryly on the whole scene: “you learn more from Twitter than from being on it directly… first of all, I felt that ‘this is too much’.”
Here’s what attendees had to say about this year’s NeurIPS:
What was the hottest topic among participants and which one do you think more people will be talking about in 2026?
Which labs seem to be gaining momentum and which are more tentative?
What was the best party you’ve ever been to or had FOMO at?
Yes, some people thought speeches were parties. I think that academia lives on in NeurIPS after all.