>Aging and death is not an accident or a deterioration from damage or a drift away from a correct instruction set. It is an absolute hard necessity, without which we would all be long gone.
Not buying this at all. This view is completely at odds against multicellular organisms that can literally clone themselves.
Everything we know about aging so far points towards the deterioration of renewal mechanisms that keeps our system functional. Cells die due to a host of reasons besides programming. For example, they may become non-functional and eliminated by the immune system.
But that deterioration seems to be programmed, or at least not selected against, which is surprising. We know animals with drastically different rates of aging, and some that are effectively immortal.
The flaw with this argument is the assumption people who control the LLMs is the sole problem. Plenty of people use LLMs for personal gain at the cost of others.
Power is a leverage. The people who control the LLMs have way more power and capacity for harm than Average Joe Misusing Tech. Your comment is pure ‘whataboutism’.
Not sure if something like this is already on the table, but I would like to see Claude responses more in line with Simplified Technical English [1] by default. I find those writing styles a lot easier to read. This has been standardised as ASD-STE100 [2]. I've seen few people making SKILL.md files with that in mind, which works great, but having this by default without invoking the skill command would be better.
That's the opening line to Pride and Prejudice, where Jane Austen (I guess the "bot" part of the name is intentional) states something that many people of the time would superficially agree on, but which is intended to be highly ironic.
I guess the point of GP (and of Jane Austen) is that people never actually universally agree on anything. And when they superficially do, there is actually a large undercurrent of disagreement.
Another fun quote apropos here would be "I love standards, there are so many to choose from"
There's a great book from a different line that delves into this as a form of what the author calls 'manifold objectivism'. In short, there are some big concepts like Christianity or Islam that people refer to as if it's the same thing but almost nobody has a shared meaning when they refer to such big things. Even so there can be concrete dialogues about these topics from different frames and with different underlying meanings.
There's trouble though when either there is conflict that isn't reconciled between the interlocutors or worse, when there is a satisfactory conclusion between the two interlocutors who never accounted for the divergent definitions ... there are many 'objective' understandings of what a thing is.
"A Fundamental Fear: Eurocentrism and the Emergence of Islamism"
> short, there are some big concepts like Christianity or Islam that people refer to as if it's the same thing but almost nobody has a shared meaning when they refer to such big things. Even so there can be concrete dialogues about these topics from different frames and with different underlying meanings.
See Wittgenstein’s Philosophical Investigations for a deeper treatment of that topic.
You can prompt it, and it almost immediately forgets. Or context compaxtion purges ut. Or it somewhat communicates with you in it, but writes all comments and docs in its original bloated style.
Training supersedes random markdown files and prompts
Let's say Claude Code's system prompt is updated to recommend responding in Simplified Technical English. It'll work for the current generation of models. But Anthropic will train the next generation via RL on Claude Code traces as they do today. As long as they don't change their reward design, the next generation is going to be pulled towards Claudish again, because speaking Claudish gives higher rewards, so in the end the prompt doesn't really matter.
Ideally these things shouldn't work like this but this seems to be the sorry state of RL right now.
You miss that it is not just your prompt but also the various system prompts, plus how the model was trained. But you can reduce the verbosity (also with a setting in /config).
And memory and code comment styles - i think that’s a big one people forget about. You can prompt it all you want, when it sees elaborate comments in memory and code it will follow the style
IMO that would be a mistake because Simplified Technical English cannot properly represent business domains specifically when talking about using specific concepts from those domains. It can explain those concepts but I think it will fail short or naming them.
So I think making that default as it is will create bugs. I ran an experiment here https://allaboutcoding.ghinda.com/explain-to-me-in-simple-te... (of course it is fit to my usage) see section "What about understanding and facts" and ASD-STE100 fails, in my experiment, to return facts as I have defined them in those cases compared with no instruction or just saying "use Simple Technical English".
Well that is why it is technical English. I wouldn't use it to explain business concepts, but it is perfect for explaining logical flows and how something works.
I was replying to someone saying to make this the default.
While this is good for _technical english_ it is not a good default as a good default should work for most of the people in most of the cases.
IMHO the case for explaining logical flows and how someting works is just one case even when using Claude Code in programming so a default will not make sense.
If someone has a way to tame GLM into doing this then please…
As a workhorse, GLM is so good, but goodness, its prose, wherever needed, makes me feel like going to a park and kicking all the benches there endlessly. And it doesn't change!
I have instructions to combine style of Brooks, Steven King and Economist Style Guide, while avoiding anything that may put wrong emphasis („genuinely“, „load-bearing“) does not contribute anything meaningful („X rather than Y“). The output is quite good* so far even on Opus 4.7-5, when it concerns product requirements. The key is to identify the pattern and focus on the meaning of the undesirable language.
I've tried controlling it via fine-tuning Claude.md and in-session messages/instructions. It always results in failure and then "Yes, guardrails are already there. I still failed" and it feels like "I am like this. Deal with it". I've even tried languages that avoids negatives e.g. "Don't.." "never.." etc. Nope. Just doesn't work.
Few more tips to make failures more rare:
1. interactive orchestrator just ensures that process is followed and the progress towards North Star objective is steady and with predictable quality.
2. Agents use specific skills which have output gates, one of which is always quality of writing and reasoning. Orchestrator accepts work only if gate criteria are met.
This is going to make work much slower and token consumption higher. ROI from fixed rate subscriptions is still quite good. ROI from volume-based subscriptions needs to be watched carefully.
By default? I doubt it would be useful or desirable for 95% of users. I'm guessing most people find natural language easiest to read, STE seems to be like a project almost akin to Esperanto or Lojban.
I don't know, in my experience the chat output has been fairly natural.
The quoted sentence seems to be from some "model internals". Of course that can be incomprehensible, what an AI assistant is really doing is just statistically generating text with some bells and whistles to make it do useful stuff. Poking around in the internals doesn't seem like a good use of time? Comparable to cat:ing a compressed file, of course you get garbage.
I'm talking about the value of making the output in STE style, which doesn't seem useful. The value of having internal stuff in a given writing style would have to be evaluated as to whether it improves model performance.
> The quoted sentence seems to be from some "model internals".
I get these kinds of responses all the time.
> I'm talking about the value of making the output in STE style, which doesn't seem useful.
It is extremely useful. Because model output is literally what you wrote: "statistically generating text with some bells and whistles to make it do useful stuff". And in the past few months all Anthropic models have been outputting insanely bloated jargon-laden bullshit. If I see another "this leg of the decision tree holds the simple fact", I will scream.
I don't want this "internal" monologue when it talks to me, writes documents, or outputs comments in code.
Eh? STE is designed for procedural writing, e.g. "if you have a problem, press this button". I generally don't ask advanced AI to give me step-by-step instructions.
The giant ones in data centers are models that broadly encompass a variety of data for general purpose applications. You can certainly train models specifically for processing DOM content, making them smaller and more efficient.
There's a video on YouTube by a paleontologist giving his own theory for why sabre-tooth cats had these teeth which would be more hindrance then help in the kind of use most people imagine for them. Fossils show they couldn't open their jaws wide enough to use these teeth in stabbing fashion, except on something very thin.
The theory is that the sabre-tooths evolved to specialize in feeding on long-necked camelids (camel family) and the teeth evolved to let them deliver a fatal slicing bite to the neck. The sabre-tooths teeth fit neatly between the neck vertebra of a camelid, and have serated edges to let them slice through the tissue - the teeth are made for slicing, not stabbing (edit: as much as specialized stabbing).
Being so specialized they likely went extinct when their prey did (for whatever reason).
Yes, but if you imagine the dynamics, the cat chomps down on the prey's neck (which unlike anywhere else is narrow enough that it can open it's jaws wide enough to stab through) and just the natural struggle would let the serrations do their work. The cat wouldn't need to cut all the way through the neck - as soon as the arteries are cut then the prey will weaken and bleed out, which seems to be a fairly common MO among some predators.
Cats don't really chew much anyway. If sabre-tooths have a rough tongue like other cats, they can probably scour flesh off bones with that with little need to work over the meat with their teeth.
Is that actually true? If you observe domestic cats feeding, they chew a lot. Small fragments at a time and chew the crap out of it. Whether that's also true for big cats, i don't know.
My cats rarely chew, either dry or wet food, even treats they inhale and swallow without chewing. If I want them to chew, I have to give them meat trimmings or something large enough that they can't swallow it without chewing.
They seem smart enough not to choke on larger pieces of meat but it's still weird to me that they don't chew dry food.
They chew kibble, but they don't really encounter food in the wild that's anything like kibble. If you give them wet food, you'll notice they practically inhale it.
My cat almost doesn't chew at all, she's only fed wet food which she swallows in chunks and the rare prey she catches outdoors is teared and swallowed in chunks as well.
If you let your cat outside, please put a bell or something on them. The average cat hunts far more than they eat, and generally destroy local bird populations.
I was recently told this by my vet when my cat needed some teeth removed -- there was no real impact on the cat's life because they don't really chew their food.
>They don't stop the mouth from closing. Presumably they don't stop it from opening either.
That much is obvious. The question is how they get around sabre tooth without being a major obstruction to bite pieces of meat off.
As for the other animals, Elephants mostly use their trunks to put food in their mouths. Pigs, their teeth are not that huge comparatively speaking and less of an obstruction, plus the have a very long snouts. Walruses, I have no idea how they manage to feed either.
Canines are their hunting knives, where they need to penetrate deeply and then hold the piece in place until it is disabled. Canines are unable to chew properly, as they don't have a counterpart that works as the anvil, they just perforate. Also, the structure analysis show that canines would shatter if used to chew bones, so they where reserved to bite the neck or the belly while hunting.
It is speculated that the Sabre Tooth tigers (500 kg/1000 pounds) could take down a 3000 kg/6000 pound) megahervibore, that probably had a skin and fat as thick as a modern elephant. They needed those canines to kill those preys.
> Walruses, I have no idea how they manage to feed either.
they blow air on the sea floor to dislodge clams, use their whiskers to locate them and and then suck them out of their shell. I did not know walruses survive mostly on clams before today.
>What I find fascinating is the application of information that is not present in the data but that can be confirmed by the data. I’m thinking specifically of deblurring algorithms people use to recover text.
I would argue that high frequency information is not missing in a blurred image, rather it's spread out spatially. Every sample in the blurred image is a convolution of multiple samples (i.e. high frequency data) from the originating source. Theoretically, it should be invertible, provided the exact function is known that does the convolution, but there are practical limitations, such as knowing how to apply the inverse operations at the edges and dealing with the extreme sensitivity to errors.
Not buying this at all. This view is completely at odds against multicellular organisms that can literally clone themselves.
Everything we know about aging so far points towards the deterioration of renewal mechanisms that keeps our system functional. Cells die due to a host of reasons besides programming. For example, they may become non-functional and eliminated by the immune system.
reply