Legion Holiday Events, Open-Source Models, Closed Source Models and Where NEAR Fits In.

Captions as YouTube publishes them, usually machine-generated, so product and people names can be misheard.

1. 0:00

   Good morning folks. At 12 p.m. eastern time today we have the X Space with the Aftermath team to talk NEAR Intents value proposition and their product plan for the future. Hope to see many of you there. Also note for those of you interested in the Legion holiday plans, the 26th of December at 2 p.m.

2. 0:15

   Eastern Time, we have our Legion Christmas live stream. It will feature the Degen Arena Santa Claus himself, as well as the opportunity to come up on screen and interact with other Legion members. The video today is actually going to be a step back to focusing on one of the three crucial themes of the AI era we're currently in, and where NEAR fits in that paradigm.

3. 0:33

   We'll do the other two in coming videos. The first theme has to do with models. On the one hand, we have closed source models and on the other hand we have open source models, closed source models or models like Gemini, grok or ChatGPT.

4. 0:45

   These models are created using billions of dollars of compute and training costs from private companies, who keep these models closed and charge for use. However, these models are also data extractive and do not enable end users to privately benefit from the model intelligence.

5. 1:01

   The owner of the model can see all of the inference and engagement with that model. Open source models like DeepSeek, Kimi K2, Qwen and Llama are available for anyone to run should they have the compute, but are not usually monetizable.

6. 1:15

   They also tend to have their own implicit biases within them, specifically if the models have been made in China. NEAR enables open source models to be run privately for users by leveraging trusted execution environments. this is an important note because the success of Nia's enterprise products is tied very closely to performance improvements in open source models.

7. 1:35

   Every new open source model in the market is to the advantage of NEAR. Now, why does the model matter? Because every AI agent that you hear about runs using one or multiple closed source or open source models. The agent's capacity and performance, as well as its privacy and ownership, is tied to the model or models that it's been built off of.

8. 1:55

   At root, AI agents are only as capable, private, and user owned as the models they're built on. And that's why private inference with NEAR is actually a very big deal. A genetic commerce on NEAR Intents has the opportunity, for the first time, to be fully private and user owned agentic commerce.

9. 2:13

   Stick around. J1 Championship season kicks off tomorrow and we'll dive into the other two crucial themes of the AI era as well.