I gave some LLM's my password manager and some credit cards to try to do this sort of thing lately.
They failed most of the time. Simple things like finding the right password for Gmail sometimes was out of reach. Anti bot techniques sometimes stopped it.
Impressively, sometimes they'd successfully write hugely complex bash or python scripts to do tasks on web pages they hadn't managed to do with the browser automation.
Getting together a group of owners with 3% ownership by value or 3% by voting rights both seems do-able, and I'm sure would be done of there was a serious case of mismanagement.
We all know this is the first political position. They'll walk back half of it, and what remains will appear to be a compromise, but was what was intended all along.
We all also know they know they will be absolutely ravaged in the next GE, at the scale of the Tories.
The question - why hand that to Farage and his far right? Is Keir Starmer a far right operative in Labour? His track revord would suggest so, but do we have any receipts?
I don't think so, I kind of wish but he was a skilled human rights lawyer, the head of CPS, an utter moron wouldn't be able to achieve this much, he's not a lettuce.
I'm afraid it's a malicious intent here, and I also wish I could see a competent government, but until one takes on FPTP and media dictating the policy, nothing changes.
The hope is in the angry youngsters, maybe they could vote Greens+LibDems in.
Alternative interpretation: the purpose of a system is what it does, and the PMs and governments are competent at something else than what you think they should be competent at.
Yes, it's like selling a share to a group of mates who hear from their mates that AI is hot so they want in. Still does nothing for the profit not being there to pay those investors back in any other way than via new investors (ie pension funds).
The providers must have a more efficient approach. Most cache every request for 12+ hours, and they certainly can't spare 100GB of ram per request for 12 hours.
have things changed around this recently? I know openai optionally allows 24 hours but thought it was ~1h without that, and anthropic used to quote 5-15 minutes or something.
Anthropic is 5 minutes, though you can pay more to get 60 minutes I believe.
The #1 was to make Claude code token quota go further is to never let the 5 minute cache TTL expire. Either send a new request within the window, or use /clear and copy/paste, or use /clear and a framework that automatically generates session state that gets replayed from files after /clear.
or maybe they don’t actually cache (fully) but lie and just don’t charge the user right now. at least half the users, who are probably also using the most similar tokens / prompts, wouldn’t really know the difference in latency (or care)
If it actually cost that much RAM, they would almost certainly add extra things to the API to manage cache lifetime. Ie. A 'please cache this for X minutes' flag, or a setting for a single re-use cache (the most common use case)
Except the providers also cache the parsing of the prompt (the KV cache), and that has substantial cost savings (easily an 80% saving on typical coding use cases).
That caching is done server side and not passed to the client. Which in turn means they still need state management on the server side, although it perhaps doesn't need the same level of global replication and availability.
from the march changes, it looked like they increased cache eviction rates on the VRAM at claude causing everyone to start burning tokens as they had to regen token state.
Screw those things up, and those taxes will bankrupt you because they can exceed all your other earnings.
reply