skerit

I just had a look, and they don't. It's really just a temporary fork of the conversation, so it's fully cached. But it does run every 30 seconds, so that is very bad especially if it's a subagent that's at 800k tokens.

And I don't really get the point of this feature, I've always found it super annoying that the main agent creates a subagent with the clear description of "Fix tile cache" for example, and then after a while it turns into something generic like "Running browser tests"

> the module ships as precompiled Bun bytecode, so editing its JS text inside the binary changes nothing

It's actually not super difficult to "decompile" it, tweak some stuff, and recompile. I use this method to change some ridiculous hardcoded instructions in the system prompt.

It's Google. They can't help themselves. This is what they do.

> while dismissing the black hole that is training

True, but the point the "subsidized!" people are convinced of is that for each generated token Anthropic is losing money, when it's exactly the other way around.

Yes, it's quite easy to disable.

Just go to /config, then search for "Feedback" and then you can disable the "Claude-drafted feedback"

I would suggest you do this, because it does use your tokens to generate this feedback.

I really hate how this feedback feature is enabled automatically AND that it is using YOUR tokens to generate it.

I wonder how the "It's all heavily subsidized!" people are going to spin this.

Maybe at a certain point their capacity is sufficiently being used up by the "full cost" API usage, and offering cheaper package deals would just cut into their profit because they can't get more inference?

I've shared a sub with my husband for a long time, and we never ran into trouble.

Then I shared a sub with 3 people in total and we were detected within a few days.

I don't know what metrics they use, because I use my sub on a handful of servers and different IPs. But I'm guessing there's some kind of fingerprinting going on, where they can actually determine that prompts are being sent by different people

> The major providers are unprofitable after you add in the training and research costs and other overheads

The very short lives of all these models also factor in.

I think the funny thing is the more inference, the more tokens a specific version of a model spits out, actually does make it more profitable.

But somehow a lot of people truly think that each token generated increases the cost of OpenAI/Anthropic/... and shout bloody murder about "subsidies"

And I'm certain there are people out there that actually think the API prices of OpenAI and Anthropic somehow equal their electricity costs.

Technically you can still run Claude-Code through proxies though, even with the subscription.

Right now I run my Claude-Code sessions through a very simple mitmproxy, logging everything that goes in & out. It's very low level, sure, but it works.

> So the Belgians enshittified speculaas to create speculoos

Yes and no. This "Speculoos is actually cheap Speculaas" isn't completely true. That's what people like to say today, but the "Speculoos" name has been around much longer than Lotus bakeries. It's just a dialect thing and there is no reason. Some people called it "Speculaas", some people called it "Speculoos", and _then_ a cheaper variant got made and they choose the "Speculoos" term for it.

It all just sounds really fitting because the "-loos" suffix in Dutch means "without".

I've been decompiling a bunch of games for the past year. It's insane what it can do, and honestly even Opus 4.5 was already really good at this.

I do wonder what this means for open-source versions of games, as in: could companies threaten with legal actions more easily now?

Really? Reminds me of Gemini 3. It took so long before being able to get a subscription for it, and then I would run out of their equivalent of the 5-hour usage window in a heartbeat.

Lol

Oh look, more people finding out they're not that special after all.

> one of the hardest things every fall, and that makes me depressed for weeks, is when one weekend, when the days are already getting shorter fast. One hour of evening day light gets janked overnight

Same! Though you have it way worse.

I live in Belgium, I mostly get up around 08:00/08:30. I already feel quite guilty for not making the best of use of daylight in spring and summer. Around the summer solstice, the sun comes up around 05:30 over here. I can't imagine it coming up at 04:30 in the middle of summer. (Which is probably already what happens in Norway?)

> The health evidence points to standard time, largely because morning light matters for our body clocks, sleep, and mood

Depending on where you live, depending on how long your commute is and by what time you have to be at your job, ... a lot of people don't get that anyway in the winter. There is no one solution, and you're going to piss of a bunch of people anyway.

Everything in the top 25 is above a 9.4 out of 10, so it seems a bit saturated? Also seeing that Claude Opus 5.5 is at #23 makes me think this is not the benchmark for me

> I haven't been this struck by a step change since 4.5/4.6. It's a bigger jump even than when Fable first arrived.

I had the exact same experience. And unlike Fable, it doesn't gobble up your entire usage limit in a few hours.

I always wonder what the "good enough" people are actually using it for.

Clearly you people have never been to Ghent https://11ways.be/ympmfp

(Ok the middle tower is the belfry, not a church. But there's actually another church just a few hundred meters to the right of the image, though it never got a peak)

I used to like coding, and I did look at it as a puzzle. But at the end of the day I always thought it didn't go fast enough, and I had to do so much more work to be able to finish my own projects. That actually made me sad. Now I actually can do all of that. And while Claude is doing that, I just play Minecraft or something.

> It’s beneficial for them to say that because it sets the ground for these models to be restrained legally and regulatorily later on.

I'm glad people are saying this out loud, because that is what they want. Not for the good of the world, but for the good of their pockets.

I used it for a while, and I thought it was quite hard to follow what was going on. And in the end it just went off the rails anyway, though that might have been a GPT-6 kind of thing.

So their original plan was to axe Terra, but then introduce an "Astra Light" model a week later? They had a nice lineup named for a whole 3 months, and they're already messing with it.

Opus 5.5 is just too eager in everything it does. The only good thing about this is that it's most often doing something correct.

I've switching from only using markdown in my prompts to using XML tags this year too. It's not only easy for the model to see when something ends, it's quite useful for me too.

> In Anthropic's testing, at its default "medium" effort the model matched or beat Claude Opus 5 at "high" effort on such tasks, in fewer steps and with fewer tokens

Opus 5.5 has been amazing, but I'm confused by how this is worded. It "matched or beat" Opus 5? There is no matching. There is only surpassing. By miles. Like Opus 5 was the biggest disappointment of the year. Opus 5.5 is even better than Fable. I do not understand why they're not acknowledging it for the leap that it is?

Retiring the Terra tier? Their space-inspired lineup has only been out for 2 months, they're already messing with it?

I like the Pelican test. And I agree this pelican looks very boring.