Compute Finance @computefinx

@chipflare @TBU12345678 @insane_analyst This is my fav space for opportunity currently. Share prices are up across the board but the fundamental opportunities look like they can go much higher + margin improvement from volume

0 0 1 26 0

View Details

Compute Finance @computefinx

6 days ago

@bubbleboi @labubu_trader It’s all about EMIB-T. EMIB’s issues are routing power around the bridge. EMIB-T solves that and should actually be a real step forward vs EMIB which mainly just catches them up to CoWoS

0 0 6 2K 1

View Details

Compute Finance @computefinx

7 days ago

@andrewmccalip How do you go about building something like this? Mind open-sourcing it?

0 0 0 821 0

View Details

Compute Finance @computefinx

2 weeks ago

@EnoReyes This is amazing and likely implies a roadmap of further cost savings over time as the orchestration model is able to find more opportunities for prompt caching as it builds data context on each user

0 0 1 16 0

View Details

Compute Finance @computefinx

2 weeks ago

@EnoReyes Is this done via a lightweight orchestration model that runs over the top and manages the project in real time, mostly pre-programmed policies that partitions the project into different model pipelines ahead of time, or some hybrid version of the two?

1 0 0 13 0

View Details

Compute Finance @computefinx

2 weeks ago

@EnoReyes Man this is amazing. Tell me more. Huge fan of this

2 0 3 159 0

View Details

Compute Finance @computefinx

3 weeks ago

@ivanburazin @daytonaio How does this change the capex formula as the business scales though? Higher ratio of local NVMe sounds expensive

0 0 0 340 0

View Details

Compute Finance @computefinx

3 weeks ago

@ohlennart Because it isn’t. It’s somewhere between an overstatement and outright lie

0 0 3 881 0

View Details

Compute Finance @computefinx

3 weeks ago

@esidery @ShamsCharania If the Clippers offer pick 5, a future 1st, Mathurin, and salary filler - does that get done?

5 0 0 633 1

View Details

Compute Finance @computefinx

4 weeks ago

@ivanburazin How is memory management handled though? Is the prior context and environment compressed at all, or does it just balloon the underlying model memory usage?