It's insane that Cursor charges 25 extra cents per million tokens on a model that charges 2 cents per million cached tokens. Basically everyone is forced to use Cursor's own models at this point as otherwise the costs are simply too high.
I tried using deepseek v4 flash with OpenRouter. It switches between providers too eagerly which resets the cache. Then, each provider begins to rate limit me for providing so many uncached tokens, so it just keeps on switching providers. I'm paying for every token... why rate limit me? It was unusable compared to just using the official Deepseek provider which has a much better cache rate.
Fun fact - bedrock only saves chunks modified by the player because consoles are much more restrictive with save file size than PC. If the world gets too big, your save is effectively lost.
Java does save chunks after generation, likely due to world gen updates not seamlessly transitioning with older versions of the game and PCs not caring about save size.
You can genuinely feel just a few milliseconds of input latency with a mouse, it disconnects looking around with your physical actions. It's the same way you can feel the difference between low and high refresh rate monitors, even though you can't really count the frames themselves
I can't use Fable at work because all prompts are saved for "safety". I can't use Fable at home as the API pricing is much too expensive while codex is heavily subsidized with generous limits.
Fable is pretty fantastic for standard code monkey stuff. But yeah, anything related to security or science that's six degrees related to death, drugs, and/or fun gets flagged.
have fable do the planning and explicitly tell it to farm out actual tasks to opus 4.8 and sonnet 5 depending on complexity. it works pretty well but yes the API switchover guessing game is no fun
I've played around GPT 5.6 sol high at both work and home.
At work, it was able to one shot a dashboard. Of course, my prompts are vague as I'm not exactly sure what I want yet, but it did a better job than I could do as a backend dev forced to work on frontend sometimes.
Usage is also great, it just feels so much more efficient than older models in terms of thinking and time. Cost is barely better though.
It can burn a million tokens in less than a minute, at least at launch where there's likely less load on the servers.
At home, it feels like I'm fighting the AI less while letting it refactor code. I'm glad that I left this 12,000 line vibe coded port of a hand written codebase to future models to refactor. It feels like the model has better judgement than old models that would destroy your codebase so long as it meant accomplishing your prompt.
abyssmc.org was my server and the website has season 1 and season 2 links online while the newer seasons died. I could also get you the earth map with about 100 years of cumulative playtime, the first two seasons maps had much less playtime
My name is Jeong Dongwoo, and my email is doogwoo@gmail.com. If you send me an email, I will explain in detail how to use it. I would greatly appreciate your help
that is quite funny, tbf. looks like the person who opened is friends with eva, a member of "server scanning inc" [1] so they thought it was worth a shot
I've found that opencode and codex are the two subscriptions that still seem to subsize usage. Deepseek V4 has been the most powerful model in opencode IMO, I trust it with problems where I can validate the solution such as debugging an issue - but I only trust the proprietary GPT-5.5 and Claude Opus 4.7 models for writing code that matters.
It's still not uncommon for it to escape it's thinking block accidentally and be unable to end it's response, or for it to call the same tool repeatedly. I've watched it burn 50 million tokens in a loop before killing the chat.
