xnorswap

I've been rinsing Opus 5.5 and I've been unable to get over 20% of my plan usage, it seems to use a tiny fraction of the usage that Opus4.8 or 5.0 used.

I don't know if it's an accounting trick or genuinely less usage, but since 5.5 I've not had to think about limits at all.

I don't want to adjust my system volume to change the volume of an over-loud video. My system volume is set for my comfort so my notifications and other applications are the right volume.

A volume slider on a video player is a really basic feature that every one should have.

Whom I used to confuse with Bombay Bicycle Club.

Meal? Washing machine wash clothes, dishwashers wash dishes, so something may have been lost in translation.

And even so, the energy efficiency of modern dishwashers with turbidity sensors is very good, and actually can be more efficient than hand-washing, especially if that requires multiple bowls full of water.

Obviously there are limits, but if you stack during the day and stick it on overnight, then it's pretty efficient, even for a single person.

Right now there's a huge friction with IDEs, that if I have it open while claude is working on something, then any caches it has get stale very quickly, and it churns processing trying to keep up with edits.

Any leverage that claude could benefit from LSPs or language/AST MCPs is undone by that tooling not being able to keep up or stay fresh, so the IDE just "runs hot" for no purpose.

Which means I'm incentivised to close my IDE while claude is working, and now I've got used to not looking at the code.

Like you, I really want a system that works better. I want claude to search and edit through better tooling, it would be faster, more accurate, and should in theory be able to leverage compilers to apply edits better and more accurately.

I don't want claude wasting an awful lot of time grepping for things that a compiler cache can answer, but if that compiler cache can't keep up then it's actively harmful.

I'm sure there must be a better way, there must be a way to have smarter more change-aware cache invalidation, and smarter more cache-aware claude tool use, but so far none has happened.

Long term I'm sure the pairing will emerge, but unfortunately I don't think that's going to come from the IDE vendor side, it'll have to come out of some harness research lab, where Anthropic or a competitor finds a way to reliably leverage this stuff for a speed and accuracy boost.

It's completely bonkers to me that an LLM can almost "natively" read profiler output, but doing so involves vast amounts of text and context ingestion, when there are real tools sitting there that it ignores in preference for either reading vast amounts or iffy python one-shot scripts to parse stuff.

That's so obviously LLM slop. It doesn't seem like that was noticed so much at the time though.

They handled that comically badly, the proportions meant the gaps went from being perceived as small to being a feature:

https://www.bbc.co.uk/news/uk-44910195

Visually it looks like they took out half the peaks.

Your mind tells you that for every gap there used to be a peak there, regardless of the truth.

The task bar gradient felt straight out of windows 2000 too, immediately felt anachronistic.

The "display settings" also claims it is set to 256 colour while the display is clearly far more colours than that too.

This is "rub your gameboy the right way to catch more pokemon" levels of insanity.

Random performance is random, your brain will jump through hurdles to fit patterns where there aren't any.

Sometimes a table is more clear than a chart.

And as long as we keep changing the meaning of coding!

Out of interest, what was the ciphertext?

Edit: Found it from here: https://mvueh-enigma-solved.carterl.chatgpt.site/

    ICRVSORMCCWQTATYEVFXDBZGGSNXWLPSYWZYTCBSWULRTBZCVGODVJUSLSOOMJQJZSXSEBZPEYMDNXJYTC

Yes, it could be a mass-delusion drive ad-spend, but I think you're vastly underestimating how much these logos actually go in to "brand recognition".

Next time you walk into your local electronics retailer and look at TVs, you might see a Samsung, LG, HiSense.

Chances are, if not for massive amounts spent on advertising sports, you'd overlook some of these brands in favour of the ones you recognise.

It won't be a conscious, "I saw this during the north London derby so I'll buy it", but subconscious name recognition will have a huge impact on willingness to buy.

They clearly don't, otherwise it wouldn't generate so much advertising revenue.

I haven't tried it since it was first released but it didn't seem to work at all for me back then.

It was so slow that the roslyn results would be lagged well behind any edits it was making, which would just leave it confused.

I'm feeling the same, personal projects have never felt better, but work has slowly felt more miserable.

Code reviews come back with a ream of issues that I just want to scream "I don't give a fuck".

Pre-LLM no-one would have cared that if another class was configured differently ( it isn't ) and called the new class in a different way ( it doesn't ), that you'd in some race-condition occasionally leak a resource, but only on shutdown when the program is about to terminate anyway.

That's the kind of issue that claude loves to find and write up in a code review with the exact same verbiage and write up style as if it had just found a naive SQL injection from string concatenation.

Prior to LLMs I doubt anyone would even have spotted the first issue, let alone bothered to write it up.

But instead it's another round of review, another round of trying to decipher the LLM's comments to discover whether it's a meaningful problem or a nothingburger.

And it's all so tiring.

I wouldn't say that's flat at all, if you switch to the "architecture" graph it gives a clearer picture without the between-version artefacts of dips that don't exist: https://stats.documentfoundation.org/downloads#month,arch

What matters more than the peaks are the floors, which are clearly increasing.

There isn't any need to "look things up related to the conversation", and if that is seen as acceptable then that itself is a sign of how things have changed, and how addicted we all are to information.

Many wikis do this, Wikipedia used to publicly log IP of all editors.

I didn't downvote, but I can understand it as a reaction to, "Ask claude yourself".

What do you mean by "acknowledge", and is that same level of acknowledgement the level used in benchmarks?

Trackbacks! Yes thank you, I'd forgotten the name.

Your musing about 2-way links reminded me of the short-lived "webhooks for backlinks" trend, which didn't last long before spammers found and abused it.

The idea was that you had a well-known webhook location, and when people linked to your blog post, they'd ping your webhook with their URL, and you'd display it under your blog post as a "backlink".

If that's obviously a way to inject marketing spam you'd be right, but I swear this stuff got quite popular before being overran.

Unfortunately google is now so bad, it's impossible to find genuine history of the phenomena I'm describing, so I just have to trust my memory that I'm not inventing a past that didn't exist.

They didn't measure which programming language because we all know it's python, which it tries to use, every session, despite repeated memory files to not use python.

Recently it's even taken to installing python to get jobs done.

Excellent article, I went into it thinking the headline is obvious, but the examples are well crafted and it isn't just the obvious stuff.

Because open protocols were abandoned, because they generally lacked security, discoverability, and perhaps most significantly of all, moderation.

Even FINGER had .plan files, where you could get status updates from people.

But how are you going to query whose plans you want to read, and who is going to moderate without centralisation?

It's the moderation in particular that dooms most federalised efforts. The web itself is only really moderated through the relative lack of discoverability, with Google and hosting providers having become a form of gatekeeper to remove the very worst of the web.

Twitter in the early days took off because it did solve a technological problem of the time. You only had your SMS enabled phone, no WAP or smartphone, and wanted to provide updates. They gave you a number you could text, which would post updates. Hence the character limit ( less than the SMS limit, to leave space for your credentials ).

Plenty of alternatives exist for getting information out, people just don't know where to look for it.

That technical problem disappeared, but it had by then grown to provide a service to rapidly get updates from people, one that had a sense of authority and the moderating effect and discoverability of centralisation is impossible to ignore.

> Wtf, UK?

We can't even blame this one on the relative lack of infrastructure investment, because busses have never been any better as far as I can tell.

In fact, the only reason the bus fare is as "cheap" as £3 is because of a price cap, which is optional for the bus companies to follow, and has the effect of meaning that they no longer offer return tickets, so you get to choose between the price capped £3 single ticket or a "day saver" £7.

Outside of London, Manchester, Nottingham and a few selected other places, busses in England tend to be fairly dire, especially as you get out to smaller towns and villages.

Even if the routes are there, it's not uncommon for the actual busses to bare little resemblance to the published timetable.

In some places you'll get fancy live information screens which happily announce the arrival of a bus that doesn't exist, or more frequently, to simply state that the information isn't available, please check the paper boards.

Bus routes tend to be especially bad in rural towns and villages, but even in my fairly reasonably large town, it's quicker and more convenient for me to walk the 1.3 miles to the train station than try to get a bus.

If I plug in "bus" on google maps planner, it has me walk 20 minutes to a bus stop to sit on a bus for 7 minutes, versus a ~25 minute walk.

Alternatively, I could walk 5 minutes to a nearer stop, wait 30 minutes for the next bus, which itself then takes 26 minutes to get to the bus station, which is itself a 3-4 minute walk from the bus station to the train station itself.

And that's assuming the bus is on time running to schedule, and doesn't then hit traffic.

I'd also be paying £3 one-way for this privilege.

Once you're inside the TFL zone everything changes and busses are frequent, generally reliable, and cheaper.

I would disagree with that, I've seen people anchor their own identity to that of brands like Apple or Disney.

They may not be aesthetics you find cool, but clearly they appeal to a large number of others.

Just this morning I had a worrying result when I googled "UV index <location>" and it confidently told me the current UV index was zero.

It was not, and at least because it was obviously wrong I clicked through to a page with the real UV index.

But had it said "The UV index is 1.7" I might have believed it, but it'd still have been wrong, potentially dangerously so.

Now to be clear, live data is precisely the sort of thing I expect it to get wrong, and I never intended to trigger let alone read an "AI summary", but it's still worrying for those less inclined to dismiss it.