2025 Kickoff!! NeurIPS Recap and SOTA

For our annual NeurIPS event, we're once again bringing back our thoughts/impressions/knee-jerk reactions from the December NeurIPS conference.

Talks:

"The State of LLM Patterns & Agents" - Sam Witteveen

In this talk, Sam will look at both recent academic and industrial advancements of how people are using LLMs/VLMs. He will cover what patterns and paradigm have survived over the past 18 months and also look at the state of new techniques and frameworks for building LLM apps and agents. If 2025 is the year of AI agents what exactly is new?

"The End of Pretraining" - Martin Andrews

Now that 'reasoning models' have started to appear from the big labs, NeurIPS clearly separated into the main conference (with papers from the 'before times') and the workshops... There was no clearer watershed moment that Ilya Sutskever's talk in which he likened text tokens to the limited supply of fossil fuels. Martin will reflect on this, and how 'reasoning' will enable AI to continue to flourish - and talk about Deepseek's R1 model release.

---

Talks will start at 7:00pm and end at around 8:45pm, at which point people normally come up to the front for a bit of a chat with each other, and the speakers.

As always, we're actively looking for more speakers - both '30 minutes long-form', and lightning talks. For the lightning talks, we welcome folks to come and talk about something cool they've done with keras, PyTorch, JAX and/or Deep Learning for 5-10mins (so, if you have slides, then #max=10). We believe that the key ingredient for the success of a Lightning Talk is simply the cool/interesting factor. It doesn't matter whether you're an expert or an enthusiastic beginner: Given the responses we have had to previous talks, we're sure there are lots of people who would love to hear what you've been playing with. If you're interested in talking, please just introduce yourself to Martin at one of the events.