DeepSeek V3 - what will you build?

[justin.woo]

Last week we announced that we have Deepseek V3 in the SambaNova cloud.

I know others in the community have shared what they have built.

But… I would love to hear what others are building with DeepSeekV3 or even what they plan to do!!!

P.S. Here is a video of DeepSeekV3-0324 →
"Making the coolest HTML game" - GALACTIC DEFENDER

[me11]

We have max token limit of 8192. Which limits us to build on top of this api. Is there any way to bypass this.

[tonyd]

@me11

We have max token limit of 8192. Which limits us to build on top of this api. Is there any way to bypass this.

Not quite a bypass, but hang tight. SambaNova staff told me that increased context length for DeepSeek V3-0324 is coming soon:

@tonyd

DeepSeek themselves cite the new DeepSeek-V3-0324 model as supporting a 128k context length (source). Does SambaNova have plans to expand the context window size from the current limit of 8k (source) if/when DeepSeek-V3-0324 goes from Preview to Production?

@vasanth.mohan

Yes, we are making updates week by week to the model, so expect continued updates to increase the context length

From this thread I made where I asked the same question