The default macOS file system is just weird.
It's case-insensitive for storage but case-sensitive on display.
I learned this hard way because some files refuse to download on one of my git repo.
9,555 karma · joined November 11, 2008
It's case-insensitive for storage but case-sensitive on display.
I learned this hard way because some files refuse to download on one of my git repo.
https://en.wikipedia.org/wiki/Murder_of_Poon_Hiu-wing
A guy commited homicide and walks free after 19 months in jail.
Until the day it no longer works. Sad.
modern Python went in a somewhat different direction with asyncio, but with tasklet and continuation it could be a much powerful combo.
But IRL it's too vague. The exploit hunting is a better use case.
It's gunna to happen anyway.
We are sailing into the age of vertically integrated intelligence.
AI shouldn't be "helping" with homework. AI should be the homework.
Start with any topic or problem, students should be encouraged interact with AI, and learn stuff using Socratic method.
Teachers should rate the chat session instead.
And iterate the process to high profeciency.
Type your current status in a bar.
And the thing that aggregates them? Twitter.
People like to peek into what others are doing.
... opens the SO link
yeah there's no answer to it (even if someone does, it would likely be closed by 5 others)
The repo gave you guide and benchmark of how to tweak ds4 pro for better performance, with a clickbait title (Fable 5 level), what kind of more proof do you require? More benchmarks? Reports for media outlets? STF model on HF?
What about the runtime size? I care this because I intend to run a stripped WASM version of DuckDB in browser.
https://andymatuschak.org/books/
It explains a lot and works really well.
I tried out in ChatGPT with a simple prompt:
> ...paste link... Give me series of quiz see if I really understood the article well. Ask & answer one by one in turns.
Really fun experience.
The cycle was a "process".
two hard things in computer science
https://blog.est.im/2026/stderr-04
I think LLMs are reall good at conjecturing based on existing knowledge, but inventing new tools/lemma and new paradigm? Not really.
Srsly bro? Websocket is simple? Talking about handling state persistency and connection partitions, such a headache.
The closing } ) ] token has to be generated to match exactly the beginning { ( [ many lines before. You can challenge yourself writing Lisp by hand without cursor moving backwards, and try close correctly by counting ))))))) you'd have a big headache.
A long, nested sub-routine with many () {} will cost LLM's context and makes it underperform, because the attention head have to track the state. On the other hand the indentation level can be infered as a single token[1] and saves reasoning effort.
Note these discussion is about "code generation", not parsing.
1: https://platform.openai.com/tokenizer Try input many spaces.
Source: https://github.com/est/gitweets
Demo: https://f.est.im/
The site and data is self-contained as a git repo.
A git commit is a post. Blobs are pictures.
Comments are stored in git notes.
Quant company usually squeezing every penny.
So, death sentence even to frontier models?
100% local and no leaks.
What? I never realized it was the case. I was calling my handcrafted MCP server without the header all the time!
Turns out it's just a webp.
LLMs are surprisingly bad at generating JSON.
> Requirements: JDK 17+ and the Android SDK (platform 35).
Bummer. I am stuck with Android 14 for the moment.