@MrNoveltyGame @d4aniel_k Evolution har ju en extrem fördelning mot ett fåtal kunder. Sen verkar deras reglerade-marknader-mått helt fejk enligt @pokertrader1. Betsson på ATH och Evolution fallande, jämförelsen kan bli mer extrem då.

@Phukettrader Hur som helst borde den ”fighten” ske i MOC väl?

@Pavas_Pareto Vad är VWAP där?

@J_Landstroem Yeah you can use local LLMs with Cursor. But a local install won't have the same performance as the cutting edge that is readily available unless you spend a small fortune on your hardware which will get old very quickly.

7b distill is just not for coding, it's not in the same ballpark. o1-mini is old and I was meaning to write o3-mini which is on par with regular o1 in size but much better for coding. There's also o3-mini-high which is much greater. In general o3-mini-high > o3-mini > o1 (non mini) > R1 > o1-mini

@J_Landstroem They are IDEs but they also have their own "magic" under the core by indexing your code base, allowing you to work over several files simultaneously, better refactoring.

@J_Landstroem Funkat otroligt dåligt för generella kvartalsrapporter från väldigt olika bolag för mig i alla fall. Funkar bättre på standardiserade rapporter som i USA.

@dungeoninvest The case was that they knew that their core is very strong. And they though that they could show big growth by just increasing the investments. But the growth has actually lacked last years even though CCU is increasing and in theory we have had new core releases like Cities 2.

@J_Landstroem För att extrahering av ren text inte funkar riktigt på tabeller och eller bilder med grafer.

Varje post får också såklart med kommentarer från både rapport och conference call-ljudfil (återigen, tack @Quartr_App som tillhandlahåller detta). Både år över år men också forward looking om det finns något sådant.

@J_Landstroem I’m saying it’s actually not based on the same architecture, let alone size. But on the Qwen architecture.

@J_Landstroem To be clear: 7b distillation of DeepSeek R1 is *not* the R1 modell but usually a low-size based on Qwen distillation which is an Alibaba model.

@J_Landstroem Cursor and Windsurf also support local models but given you’re trying 7b distills or 70b down sampled to lower float bit precision I would assume that you don’t have the GPUs for that.

@LouisOliverW @alineaanalytics I don't have their service myself but I see this as a no-brainer for gaming focused analysts at banks compared to the time it takes for them to follow this type of data manually.

@stock_trap @Downtown @michaelbatnick 🔥 so many opportunities for you guys right now with the AI boom and incremental prices for compute going to zero.