simon83

So the rumor spreading around the Claude subreddits is that Anthropic is A/B testing plan quotas. A lot of people are complaining about reduced usage since about 2-3 weeks, while others don't see any significant reduction. But nobody bothers to provide data so their complaints are just feels.

Yes, I keep hearing that, but I never saw any concrete numbers. That's why I tasked Opus to write a script to keep track of this.

Here's the gist how it works:

The transcripts log the token count, so you can figure out what those tokens would cost in API prices. Each response returns the current 5h and weekly quota percentages, this is what is shown in the statusline. The returned quotas are not stored anywhere by default. So Opus wrote a small script that sits in the status line and records the percentages while I work. Dividing dollars spent by the percent used give what 1% of quota is worth.

On the 20x plan, 1% of weekly limit is worth about $20 in API prices. A full week is then about $2,000, or $8,000 a month. 1% of the 5h window is worth about $5.20.

All of this only works correctly if you use your sub on a single machine, within CC, as chat usage and other machines can not be tracked this way.

I'm on the 20x right now because I got offered 50% off as a returning customer (I cancelled a few months ago). So next billing period I'll switch to the 5x plan and measure its weekly and 5h limits the same way again.

Yes that's what I remember as well. I haven't used Claude Code in a few months and only started to use it again around 2 weeks ago, but it seems they reduced the the subscription value quite a bit. See my other comment above how I measure the API dollars to weekly rate.

Also the Claude subreddits are complaining since last week about nerfed usage. But since it is reddit, nobody bothers to actually measure the true value.

Do you have it decompile and recompile automatically after every update? Or have you turned off auto updates?

In my sessions it is ~16%, not using workflows due to the nature of my work (a lot of research subagents where fixed workflows don't work well)

Oh wow! I showed Opus the gist and asked it to analyze my own transcripts, and since those summary calls don't show up there, it calculated an estimate based on the 30 second timer.

I'm on the 20x Max plan and my quota is at 98% (resets today), so I had a quite heavy week of usage, including a bunch of orchestrated subagent runs (no workflows).

API equivalent spending for this week: $1,970

Hidden status-label calls: ~$380

Share of the true total: 16%

That's incredible...

Subagents have a default TTL of 5 minutes though, with an opt in to increase it to 1h, for a premium for writing to cache.

> I can't reason as well or as quickly. Things go over my head. My working memory is so much smaller. I get lost in code all the time. I forget what I'm working on.

Is that while being high on THC, or being sober but having consumed lets say the day before?

I also don‘t think a customer facing chat bot brings much value, but an internal, employee only chat bot could be really useful, depending on the organzization of course. The company in my last position was a rather big one with an insanely huge Confluence instance. I‘ve spent (wasted) so much time searching information there. Having a chat bot, trained on all that information, would‘ve been really useful, I think.

As a Linux noob I decided to try Arch Linux 4 years ago because I wanted to actually learn how the various parts work that make out a whole Linux distro (things like the bootloader, systemd, various desktop environments etc). Other distros give you everything ready out of the box, which made it hard for me to actually dig deeper under the hood, because I had no clue where to start. With Arch Linux you more or less build a whole Linux system from the bottom up.

Following the Arch guides to get a fully working system was easy enough for a complete Linux noob like me. But you still need some basic knowledge about computing, for things like networking, file systems, drive partitioning and such. If you learned this knowledge from other operating systems then Arch Linux is fairly easy to setup by just following their guide.

I understand that all of this is based on some fundamental mathematics, a couple of algorithms here, and some statistical analysis there. And I find it absolutely amazing that we can use all of that knowledge and encode it into something that resembles intelligence. This makes me think if our brains and the concept of intelligence are really as special and mysterious as we assume.

Google: "confetti has left the cannon"

> No results found for "confetti has left the cannon".

I'm amazed that a "stochastic parrot" can come up with such a beautiful idiom.