11.05.2026

Author: cb341

pages/threads.md

@@ -9,7 +9,101 @@ math: true
 
 Conversations, thoughts, half-ideas, things I am starting to explore.
 
-## 11.05.2026
+## 11.05.2026 (afternoon)
+
+(yes this entry contains filler, are you yet frustrated by it?)
+
+recently discovered the RFC writing style for the internet. \
+[RFC 2119](https://datatracker.ietf.org/doc/html/rfc2119) defines a handful of keywords meant to remove ambiguity from specifications:
+
+> 1. MUST   This word, or the terms "REQUIRED" or "SHALL", mean that the
+>    definition is an absolute requirement of the specification.
+> 
+> 2. MUST NOT   This phrase, or the phrase "SHALL NOT", mean that the
+>    definition is an absolute prohibition of the specification.
+> 
+> 3. SHOULD   This word, or the adjective "RECOMMENDED", mean that there
+>    may exist valid reasons in particular circumstances to ignore a
+>    particular item, but the full implications must be understood and
+>    carefully weighed before choosing a different course.
+> 4. ...
+
+we SHOULD apply this to everyday language.
+
+daily language is mostly filler. greetings burn tokens. \
+when both sides agree the goal is to exchange information, not feelings, email is high friction for no reason.
+
+on a different note: \
+can we make it socially acceptable to write emails in a colder, denser way?
+
+**BEFORE**
+>
+> Guten Abend zusammen!
+>
+> Wir haben eine Reservation für die Drachenhöhle am Freitag, 08.05.2026.
+> Ursprünglich hatten wir mit etwa 18 Personen geplant, aber es sieht nun so aus, dass wir am Ende etwa 8 Personen sein werden. Da fragen wir uns, ob ein Wechsel zur Hobbithöhle Sinn machen würde. Für eine kleinere Gruppe wäre das wahrscheinlich die gemütlichere Atmosphäre.
+>
+> Falls das noch möglich ist, würden wir uns sehr freuen. Ansonsten kommen wir natürlich auch gerne in die Drachenhöhle :)
+>
+> Vielen Dank und freundliche Grüsse
+>
+> Dani Bengl (it/its)
+
+intention is buried. greeting, setup, filler, hedge, thanks. politeness at the cost of signal.
+
+what i actually wanted to say:
+
+**AFTER**
+>
+> Reservation Fr. 08.05.2026, Drachenhöhle. \
+> Teilnehmer: 18→8. \
+> Wechsel zu Hobbithöhle möglich?
+>
+> Dani Bengl (it/its)
+
+the organizers will extract that core from my original anyway. \
+then they formulate their own core ("room available, will move", or "possible but maybe keep the bigger room"), pad it with filler, send. \
+i receive, scan, decode, repeat.
+
+why can't human <-> human communication skip the encode/decode step?
+
+maybe it can. open the channel with a header:
+
+> the following message is optimized for density. no impoliteness implied.
+> <message>
+
+and inside, reduce vocabulary for load-bearing words. MUST, MUST NOT, SHOULD, MAY. capitalized as in the RFC.
+
+builds on top of [a-optimized-language](https://cb341.dev/blog/a-optimized-language/), section "Why I prefer talking to an LLM over humans".
+
+i'll try it. either non-engineers read me as cold and heartless, or they appreciate the efficiency.
+
+## 11.05.2026 (morning)
+
+most modern languages signal sentence type at the end. \
+declarative, imperative, interrogative, exclamatory. \
+you only learn the tone after you've already read the words. \
+hard to read aloud. you have to guess, then correct.
+
+why not put the marker first?
+
+> Why do we even bother with languages?
+
+vs.
+
+> ? why do we even bother with languages
+
+reads more structured. same shape as:
+
+> Q: Why do we even bother with languages?
+
+and while we're rewriting orthography. capitalization currently marks sentence starts. \
+the leading marker would already do that. \
+free up capitalization for *meaning* instead. emphasize words that MUST NOT be skimmed, like bold or italics but in plain text.
+
+> if a leading `?` opens a question, does the trailing `?` still earn its keep? probably drop it. one marker per sentence...
+
+## 11.05.2026 (midnight)
 
 caught myself asking Claude to move a file for me today. \
 not debug something. not explain something. move a file.
@@ -40,7 +134,7 @@ Radiation brute-forces mutations. It increases the error rate in DNA copying. Mo
 
 So theoretically: set an area on fire repeatedly while exposing it to controlled radiation and you might sculpt organisms that are extremely hardy to both. Like lab-directed evolution but on a landscape scale. Chernobyl’s Red Forest + later wildfires already did a messy version of this experiment.
 
-It feels like the ultimate “move fast and break things” approach to biology. Flood the system with noise and variation, kill almost everything, and see what’s left standing. Evolution by catastrophe.
+It feels like the ultimate "move fast and break things" approach to biology. Flood the system with noise and variation, kill almost everything, and see what’s left standing. Evolution by catastrophe.
 
 Reminds me of the AI conversation. We can now brute-force code into existence by describing it. No need for the slow, careful craft. Is that better? Or are we just trading understanding for speed and volume?
 
@@ -49,10 +143,10 @@ Same question here: does flooding the genome with mutations and extreme selectio
 Nature already does this slowly with fire, volcanoes, solar radiation. We could do it faster. The question is whether we should, and what we lose in the process.
 Most mutations are bad. Most attempts fail. The “forward” leaps are rare and expensive.
 
-Still... the black Chernobyl frogs are real. They happened in ~30 years.
+Still...\
+the black Chernobyl frogs are real. They happened in ~30 years.
 
 Curious what other traits we could force if we tried.
-No conclusions. Just another void staring back.
 
 ## 04.05.2026