Thank you the Netherlands and the European Union for making me retarded today because I couldn't sleep well due to your dumb ass anti-AC regulation
I'm now less productive so will contribute less to the GDP today
So congrats you made us all become poorer
noticing claude is asking me more questions.. instead of just doing the task
this means you've got to be in front of your keyboard more
interesting hmmmm
would be good if claude instances would remember the date of each message
perfectly reasonable to come back the next day
but it never understands it's a new day
and it keeps telling me enough for today
should I just set the memory to run new Date() ?
Usage limits are up, effective today we're:
1) Doubling Claude Code's 5-hour limits for Pro, Max, Team and seat-based Enterprise plans
2) Removing peak hours limit reduction on Claude Code for Pro and Max plans
3) Substantially raising our API rate limits for Opus models
Introducing SubQ - a major breakthrough in LLM intelligence.
It is the first model built on a fully sub-quadratic sparse-attention architecture (SSA),
And the first frontier model with a 12 million token context window which is:
- 52x faster than FlashAttention at 1MM tokens
- Less than 5% the cost of Opus
Transformer-based LLMs waste compute by processing every possible relationship between words (standard attention).
Only a small fraction actually matter.
@subquadratic finds and focuses only on the ones that do.
That's nearly 1,000x less compute and a new way for LLMs to scale.
Introducing SubQ - a major breakthrough in LLM intelligence.
It is the first model built on a fully sub-quadratic sparse-attention architecture (SSA),
And the first frontier model with a 12 million token context window which is:
- 52x faster than FlashAttention at 1MM tokens
- Less than 5% the cost of Opus
Transformer-based LLMs waste compute by processing every possible relationship between words (standard attention).
Only a small fraction actually matter.
@subquadratic finds and focuses only on the ones that do.
That's nearly 1,000x less compute and a new way for LLMs to scale.
@TideBusiness your bank integrations are broken
auth happens before you're able to select which account you want to connect.
and why do you still need to choose your account even if you've only got 1 account..
one of my most common prompts to claude now is "please search comprehensively, don't be lazy"
it tries to cut so many corners and don't know if that's part of anthropic's cost cutting
incredible value still but yea.. weird given the recent changes they made to usage
anyone else noticing this?