Hacker News Reader

I watched this video a few months ago.

Virtual HLF 2020 – Scientific Dialogue: Sir C. Antony R. Hoare/Leslie Lamport

https://www.youtube.com/watch?v=wQbFkAkThGk

ontouchstarton We should revisit literate programming in the agent era

Literate programming in the sense of Donald Knuth is more about the chain of thoughts of the programmer than documenting code with comments or doc strings.

ontouchstarton We Stopped Using the Mathematics That Works

My point is that new mathematics will eventually come out of this mess. [0]

[0] https://youtu.be/2Op3QLzMgSY?t=75

ontouchstarton We Stopped Using the Mathematics That Works

Never underestimate the power of mathematics. On the other hand, a lot of mathematical breakthroughs in the history are not found by people doing mathematics but in physics and engineering. Calculus, statistics, information theory, etc.

Even for computer science, take a look at Turing Award from 1966 [0], we will see how short sighted we are if we only follow the trend. Time will tell and smart people will find new path.

[0]: https://amturing.acm.org/byyear.cfm

ontouchstarton We Stopped Using the Mathematics That Works

We are at the age of alchemy, wait for the age of chemistry and physics. New mathematical foundations are yet to be found.

ontouchstarton Ask HN: What Are You Working On? (March 2026)

When I discovered that some local llama.cpp can OCR PDF images generated by TeX, I started to revisit literate programming defined by Donald Knuth and explore using PDF as the source of truth artifact (instead of Markdown or program source code itself) for LLM to consume.

I only got to the point of having code and data as \verbatim in \LaTeX. Next step is CWEB.

Here is an example (with C and Rust code in \verbatim)

https://ontouchstart.github.io/rabbit-holes/llm_rabbit_hole_...

The ultimate goal is machine and human readable proofs on algorithms.

ontouchstarton LLMs work best when the user defines their acceptance criteria first

I made a comment in another thread about my acceptance criteria

https://news.ycombinator.com/item?id=47280645

It is more about LLMs helping me understand the problem than giving me over engineered cookie cutter solutions.

ontouchstarton We might all be AI engineers now

And you don't know what might trigger AI into overthinking. ;-)

https://gist.github.com/ontouchstart/bc301a60067f687b65dad64...

(This is an ongoing experiment, it doesn't matter what model I use.)

ontouchstarton We might all be AI engineers now

I am running local offline small models in the old fashioned REPL style, without any agentic features. One prompt at a time.

Instead of asking for answers, I ask for specific files to read or specific command line tools with specific options. I pipe the results to a file and then load it into the CLI session. Then I turn these commands into my own scripts and documentation (in Makefile).

I forbid the model wandering around to give me tons of irrelevant markdown text or generated scripts.

I ask straight questions and look for straight answers. One line at a time, one file at a time.

This gives me plenty of room to think what I want and how I get what I want.

Learning what we want and what we need to do to achieve it is the precious learning experience that we don’t want to offload to the machine.

ontouchstarton TeX Live 2026 is available for download now

I found out the other day that llama.cpp can work with PDF images offline without Internet connection. Combine with local Rustdoc and TeXLive, playing with plan TeX becomes fun again.

https://ontouchstart.github.io/rabbit-holes/tex_rabbit_hole_...

ontouchstarton Rust: The Unlikely Engine of the Vibe Coding Era

It doesn’t matter, you can replace “Rust” with any trendy language at the moment.

FORTRAN, COBOL, PASCAL, C, Java, C++, Perl, PHP, Python, Ruby, JS, TS, and now Rust…

The only thing matters is people pay attention to them and then have to use them because their job requires them.

AI would not change it.

We are slaves of the trend.

ontouchstarton Agentic Engineering Patterns

Hoarding is becoming an epidemic mental disease in the society of abundance. I don’t know what the solution would be.

https://simonwillison.net/guides/agentic-engineering-pattern...

ontouchstarton Ask HN: What prompt do you use to get Claude to consistently render LaTeX?

It seems that GitHub gist can renders some of \LaTeX, but not perfect.

https://gist.github.com/ontouchstart/bcffb186a753c5b75522fc8...

ontouchstarton Ask HN: What prompt do you use to get Claude to consistently render LaTeX?

LLMs only produce markdown [1], usually Math is wrapped with KaTeX [2]. The rendering happens in the Web UI. If some math failed to render, you can copy the code and paste to latex-sandbox [3] and fix yourself.

[1] https://docs.github.com/en/get-started/writing-on-github/wor...

[2] https://katex.org

[3] https://latex-sandbox.vercel.app

ontouchstarton Ask HN: Who is hiring? (March 2026)

What is NUS?

ontouchstarton Claude's Cycles [pdf]

I don't have time to do that myself yet so I just dug a quick TL;DR rabbit hole for fun:

https://ontouchstart.github.io/rabbit-holes/llm_rabbit_hole_...

ontouchstarton Claude's Cycles [pdf]

Fascinating report by DEK himself.

Time to sit down, read, digest and understand it without the help of LLM.

ontouchstarton Implementing a Z80 / ZX Spectrum emulator with Claude Code

Is it possible to build a full OS emulator on top of MMIX?

> The above tools could theoretically be used to compile, build, and bootstrap an entire FreeBSD, Linux, or other similar operating system kernel onto MMIX hardware, were such hardware to exist.

https://en.wikipedia.org/wiki/MMIX

ontouchstarton What does " 2>&1 " mean?

I did a google search on “dup2(2, 1)” in a fresh private tab on my iPhone Safari and this thread came up the second, between

https://man7.org/linux/man-pages/man2/dup.2.html

and

https://man.archlinux.org/man/dup2.2.en

A lot of bots are reading this. Amazing.

ontouchstarton What does " 2>&1 " mean?

That is correct. And also simple `man bash`:

REDIRECTION Before a command is executed, its input and output may be redirected using a special notation interpreted by the shell. Redirection may also be used to open and close files for the current shell execution environment. The following redirection operators may precede or appear anywhere within a simple command or may follow a command. Redirections are processed in the order they appear, from left to right.

       In the following descriptions, if the file descriptor number is
       omitted, and the first character of the redirection operator is <, the
       redirection refers to the standard input (file descriptor 0).  If the
       first character of the redirection operator is >, the redirection
       refers to the standard output (file descriptor 1).

       The word following the redirection operator in the following
       descriptions, unless otherwise noted, is subjected to brace expansion,
       tilde expansion, parameter expansion, command substitution, arithmetic
       expansion, quote removal, pathname expansion, and word splitting.  If
       it expands to more than one word, bash reports an error.

       Note that the order of redirections is significant.  For example, the
       command

              ls > dirlist 2>&1

       directs both standard output and standard error to the file dirlist,
       while the command

              ls 2>&1 > dirlist

       directs only the standard output to file dirlist, because the standard
       error was duplicated as standard output before the standard output was
       redirected to dirlist.

...

ontouchstarton What does " 2>&1 " mean?

https://www.gnu.org was there for decades, much older than SO and google.

I know GNU -> Google -> SO -> LLM is a culture shift, this is how human attention goes. The search engine and LLM only capture the latest group attention and we have short memory. That is why reading is a critical skill for us as human beings, we can't afford to outsource that to machine. (Same with writing.)

ontouchstarton What does " 2>&1 " mean?

Sometime all you need is to RTMF from the source instead of Nth hand information (N > 1)

https://www.gnu.org/software/bash/manual/html_node/Redirecti...

ontouchstarton First Website (1992)

Fascinating reading if you care about system design. Who would imagine this charged the world.

https://info.cern.ch/hypertext/WWW/DesignIssues/Overview.htm...

ontouchstarton What are the best coping mechanisms for AI Fatalism?

For example:

Allen Newell, 1975 ACM A. M. Turing Award Recipient: “Desires and Diversions”

https://www.youtube.com/watch?v=hCe0ZPGap_k

Wisdom from the previous AI era.

ontouchstarton What are the best coping mechanisms for AI Fatalism?

I am watching ACM A.M. Turing Award Laureate Interviews to set the perspective.

https://www.youtube.com/playlist?list=PLn0nrSd4xjjaSLBSzmno-...

ontouchstarton The Missing Semester of Your CS Education – Revised for 2026

I watched this YouTube video

Two decades of Git: A conversation with creator Linus Torvalds

https://www.youtube.com/watch?v=sCr_gb8rdEI

I was surprised that he only spent four months on it as a maintainer. What a great piece of work by a great software developer.

ontouchstarton The Missing Semester of Your CS Education – Revised for 2026

I am not at lecture 9 yet. I would love to follow their journey at human pace just like I did 30 years ago.

I do think your point is valid. The trend in the industry is putting emphasis on cosmetic qualities (format, workflow, testing), producing huge amount of metadata that consumes huge amount of human and machine energy for the peace of mind.

Complexity, maintainability, modularity have more to do with thinking about the problems at proper abstraction levels.

It seems we ended up spending more time on tools like writers spending more time on sharpening pencils or playing with fonts than writing something meaningful.

A low quality software can have beautiful code just like a low quality book has beautiful fonts.

ontouchstarton The Missing Semester of Your CS Education – Revised for 2026

When I was a Physics Ph.D. student in NYU in the late nineties, I took a course called UNIX tools in the CS department. It was a hands on course where the instructor did live REPL in the terminal and we watched him showing us all the tricks. I got hooked with UNIX since then. Got myself a dialup terminal in my tiny apartment in east village and dial in to the workstation on campus. The latency is so bad that I can’t only see the feedback after a few keystrokes. That was when I trained my vi muscle memory. (EMacs was out of the question.)

Later I got my own IBM 386 and installed Linux on it and started to program in Perl …

I am a big fan of Jon’s YouTube videos on Rust and I started to use Rust in non conventional ways.

I am going to follow this lecture series and “port” them to rustdoc and see how it goes.

Another rabbit hole to fall down, it is going to be fun.

https://ontouchstart.github.io/rabbit-holes/missing/

ontouchstarton Ggml.ai joins Hugging Face to ensure the long-term progress of Local AI

I have played with both mlx-lm and llama.cpp after I bought a 24GB M5 MacBook Pro last year.

Then I fell down the rabbit holes of uv, rust and C++ and forgot about LLMs. Today after I saw this announcement and answered someone’s question about how to set it up, when I got home, I decided play with llama.cpp again.

I was surprised and impressed:

https://ontouchstart.github.io/rabbit-holes/llama.cpp/

I am not going to use mlx-lm or lmstudio anymore. llama.cpp is so much fun.