🚨ÚLTIMA HORA: CERN se ha quedado fuera de servicio hasta 2030.
El Gran Colisionador de Hadrones del CERN ha entrado en un periodo de inactividad que se espera dure hasta 2030.
Y una vez más, la gente se pregunta sobre las fechas.
El colisionador se puso en marcha por primera vez en 2008.
Semanas después, el mundo se vio afectado por la peor crisis financiera en generaciones.
Luego, en 2013, se produjo un importante cierre.
Por esas mismas fechas, un meteorito masivo explotó sobre Rusia en una de las explosiones atmosféricas más potentes de la historia moderna.
Después, llegó el largo periodo de inactividad desde 2018 hasta 2022.
Durante esos años, el mundo sufrió una pandemia, confinamientos, interrupciones en la cadena de suministro, turbulencias económicas y una histórica convulsión política.
Ahora, el CERN vuelve a guardar silencio.
En un momento en que:
- La IA avanza a pasos agigantados.
- Las guerras se intensifican en múltiples regiones.
- Los gobiernos expanden sus sistemas digitales y capacidades de vigilancia.
- Y el mundo parece cambiar más rápido que nunca.
El CERN está fuera de servicio y la cuenta regresiva para 2030 ha comenzado.
¿Qué crees que saben ellos sobre los próximos cuatro años que nosotros desconocemos?
@IterIntellectus He identified a great marketing opportunity with benchmarks and he's taking advantage of it, as simple as that.
No one wants to be left behind in the AI era, whatever it takes.
@BLUECOW009 What is this narrative that is gaining traction that Grok isn't good and doesn't qualify as a "frontier" model? The model is clearly in the top three right now and definitely possesses frontier-level intelligence.
Where do these people come from?
Paid by OpenAI or something?
@zcode_ai In Grok Build it has 95% success in my tasks. Definitely the best harness for any model in general. I don't doubt that ZCode is good too, I haven't tried it.
@0xhugobatista @crisjs_19 @opencode@thdxr Each model adds to that percentage based on its own quota and cost, which determines how much you spend and the percentage that expense represents relative to that model's quota.
@0xhugobatista @crisjs_19 @opencode@thdxr I explained it better in the other replies to this tweet. Your table works exactly the same way. If you add up all your percentages, you would have already used up 95.5% of the monthly quota.
@crisjs_19 @0xhugobatista@opencode@thdxr If you use a high-cost model with a $15 cap, your monthly quota gets used up quickly. If you use only GLM 5.3 Flash, you would have $60 for the month. If you used only Omen Alpha, you would have up to $100 for the entire month.
This can be either a good thing or a bad thing.
@0xhugobatista @crisjs_19 @opencode@thdxr That is why different models with the same monthly quota may have different total requests; the model's price must also be taken into account to determine how quickly the $15, $30, or $60 threshold will be reached.
@0xhugobatista @crisjs_19 @opencode@thdxr It’s quite simple. Usage percentages are calculated based on the total dollar quota. The higher the quota and the lower the model's price, the more slowly the total percentage rises (8.1235 is 27% of 30). Monthly usage is the sum of all the percentages.
>math professor spends a year on one of the hardest unsolved problems in math
>"Navier-Stokes"
>his drafts for the whole project went through codex sessions
>openai had his logs in codex
>suddenly rumor leaks AI solved it
>he emails openai to ask what's going on
>openai calls within days
>"funny story, our model also solved it"
>using his exact approach
>asks openai: did you train on my sessions?
>no answer
@Abraham_BCT@log404tv@opencode How do you use Muse Spark outside of the OpenCode CLI via API? I'm getting an error; this hasn't happened with any other model.
Its usefulness is relative; at the time it was released, Grok 2.5 was indeed a useful release.
Grok's system prompt, as well as those of other AI labs, serves as a guide so that others can better design their own system prompts for local models depending on their needs and objectives.
The X algorithm has already been announced as being fully open source soon, just as was done with Grok Build, to increase the transparency of SpaceXAI products.
Whatever the reasons, Elon Musk and SpaceXAI are indeed part of the open source community, and they have indeed made some of their products open source; which directly contradicts your lie.