The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.
The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.
Access to all other Claude models is not affected.
We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.
Read our full statement: https://t.co/bwn0sximKZ
Noticing:
fable automatically records gif outputs and then analyzes each frame for quality concerns. this is a raw gif
1. i've never seen the anthropic/claude logo on an output at bottom right like that?
2. insanely low FPS
3. interesting to choose gifs instead of many photos
i ate dinner while it made v2
kinda silly but at the same time wow
most of the improvements came from taking a screenshot from a real league of legends game, asking it to determine specific improvements, log them in priority order, then /goal through each item
thanks ThreeJS!
some bugs i saw here:
- not all the dummies seem to actually move
- R - W behaviour is wrong (which makes sense because all I said was to copy leblanc's kit with no reference docs)
- at lots of players + particles; saw some fps slowdown
...guess what the next /goal is :D
people underestimate how much a few /compacts will save you
keep context under around 30% and your usage will remain intact
use ultrathink ultracode ultraplan etc sparingly. if you have to use this for every prompt you're going down the wrong path
use other LLMs to optimize your prompts (i use "big claude" (the web version) to write most of my claude code prompts)
use /goal with specific outcomes. this gets better everyday but it can eat tokens if you let it run for too long
dont be afraid to restart. sometimes a fresh context window is all you need
use images generously (and instruct your agents to do so as well). surprisingly cost effective
if you are being priced out of LLM usage that is a skill issue