🚀 Introducing DeepInfer, our in-house inference framework, now fully open source!
Powering DeepSeek-V4.1-Flash at scale with high throughput and low latency. We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable.
🔗 github.com/deepseek-ai/De…
1/4
0
0
21
I absolutely love this so I had @grok make an app that does it. try it out here: recipe-reduction.grok.me
Here's the alton brown aged eggnog recipe
Here's the alton brown aged eggnog recipe
i HATE how an engineer invented the correct way to write recipes like 20 years ago and nobody else ever used it
6
2
54
@w_s_bitcoin @cryptoquick I told him he made an agreement and he should honor it and he blocked me.
3
0
20
progress
Inspired by @GeoffreyHuntley i have one agent making a lisp harness/repl/thing and another one doing llm-generated smalltalk object swapping
0
0
2
“I hope nvidia releases a new spark”
Congrats to early DGX Spark owners. Same 128GB box was $3,999 at launch. Founders Edition is now $6,950, and the new 64GB starts at $4,999 with half the RAM.
blogs.nvidia.com/blog/local-ai-…
1
0
13
Getting 2 and 20 on tokens
I think we're looking at tokens as a function of salary when we should be looking at salary as a function of tokens. If that engineer is a good steward of tokens he should be managing a $50m portfolio of token spend and charging 2% management fee + 20% carry on that investment.
1
1
20
something new im starting: when I switch out the model I'm using for the main chat thread for my hermes agent, I ask it to make a self portrait of itself (well, two portraits: one as a cartoon and one realistic).
Here's deepseek vs the new qwen-next for my agent
Here's deepseek vs the new qwen-next for my agent
flipped my main inference box from deepseek-v4-flash-0731 to the new qwen3.8-flash-next. so far it seems really good
0
0
4
interesting that you can lease an M5 Ultra Mac Studio (256GB) for just a bit more than a claude max sub
43
20
663
Prompt i gave hermes using deepseek-v4-flash-0731 (with grok 4.5 for vision):
‘’’
Procedurally generated an svg of a face, black and white, use vision to check your work, tell me when you’re happy with it
‘’’
This is what it made. lmao
‘’’
Procedurally generated an svg of a face, black and white, use vision to check your work, tell me when you’re happy with it
‘’’
This is what it made. lmao
3
0
10
@InQuisitive138 @steverabinow @stephanlivera @mattkratter I don’t know if I’ll see you later so I’ll just leave this for you early
0
0
1
