@usr_bin_roygbiv Not only this. These plans math is based on the average user, not the extremes. If the average user only spends half of the weekly quota, the math is completely different
@LLMJunky It's not just about Fable. 5.5 is not super clearly better than Opus 4.8. So OpenAI either launches something near parity with Fable or their model needs to sit VERY CLEARLY between Opus and Fable but with a much more competitive price, otherwise it won't be well received
@she_llac@SemiAnalysis_ Yeah, it's easy to think that because each model has its own pricing and each task is different. So even if the person is doing the same task generally (eg asking the agent to code review a PR), the fact the PR is different and the model has different price, will have huge impact
@she_llac@SemiAnalysis_ Your analysis is now outdated as they changed limits since then. Btw, you mentioned before you already did the math for Codex, are you still planning to share it?
@LLMJunky Same thing that would happen if these companies just stopped offering plans. Most people would drastically reduce/completely stop using AI. Most people aren't even on the 200$ plans because it's too expensive even tho it gives thousands of dollars worth of usage
@synthwavedd Intelligence is just one part of the equation. Cost matters a lot too. Most companies won't let their employees use Mythos for regular stuff as that's not sustainable. Opus and GPT-5.5 cost was already becoming a problem
¡Sorteo de 5 Bundles de Good Gamers: Go Goated x Fortnite!
🗓️ Del 8 al 10 de junio
📌 Requisitos:
- Seguir a: @GoodGamers y @NotPaloIntel
- Dar Me Gusta y RT
El set estará disponible en la Tienda el 11 de junio. ¡Suerte!
@419twizzz@jxsh75fn_ it's not any items, only specific ones. and the text says "due to special circumstances" so maybe they changed something for the affected cosmetics and therefore are granting free returns
@KinchAnalytics Does it only count tournaments and ranked? Also, do game modes (solo vs squad) influence this? And is this PR the ELO used for skill based matchmaking (I assume not)?
@TSieranski@bridgemindai AI pages love to pretend they have inside information and claim a new model will be released on day X or be hugeee and when it doesnt happen, they claim the company postponed the release or nerfed the model. Majority of this pages know nothing and just make up stuff