cpeterso

Why was SVN preferable over git for finished assets?

SPOILER:

> According to the new proof, that ceiling is roughly 14 riffle shuffles for a 52-card deck, if you cut your deck in a random place with each shuffle. Beyond that point, the cards will be fully mixed.

Great article and worth reading even though you now know the answer. :)

The Claude (and ChatGPT) web UI supports incognito/temporary chats that are discarded when you close the tab. Click the ghost or dotted speech bubble icon in the upper right corner of the page.

I use this as my default mode so I don't clutter my chat history with random, one-off questions I ask. Unfortunately, there is no way to change your mind and save a productive incognito chat after you've started chatting.

[delayed]

And Apple already has a $20B search partnership with Google they can build on.

It should be, unless there are new issues. This code change landed in Firefox 153 Nightly and 153 will become the next ESR version (July 21).

https://whattrainisitnow.com/release/?version=esr

more like the "Obama Awards Obama a Medal" meme:

https://knowyourmeme.com/memes/obama-awards-obama-a-medal

You can also include other md files like AGENTS.md in CLAUDE.md:

  @AGENTS.md

The company could also test the bots in their employees’ homes for no cost. If employees aren’t comfortable with the bots in their own homes, then they shouldn’t let them loose in others’.

I name my branches for the overall task. The description of the branch’s head commit in jj status doesn’t tell the whole story.

Having to list all the bookmarks you would list to auto-advance in your jj config.toml is a hassle and easy to forget.

I see people asking LLMs to omit comments, but comments like PDL (Program Design Language) could be helpful for interactive development: ask the LLM to write pseudocode comments, review them, and then ask the LLM to expand them to actual code. People say comments should explain why, not how. That seems useful for both human reviewers and LLMs.

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

Restarting Windows is actually “cleaner” than shutting down on modern PCs because shutdown saves some kernel state for Windows Fast Startup.

I’ve read (but haven’t tested yet) that you can still enable Opus 4.6 with:

  /model claude-opus-4-6[1M]

Or describing something as “the unlock”.

Dennis E. Taylor’s “Bobiverse” series is goofy, yet hard, SF about a guy (Bob) whose uploaded mind gets sold when the cryogenic hosting company goes out of business. Given a job piloting a deep space probe, he starts replicating and exploring the galaxy.

That’s why they suggested one developer use Firefox and another one use Safari.

Ideally websites should also test with the beta versions of Firefox, Safari, and Chrome to find and report regressions (browser bugs or web spec changes) so they can get fixed before the release breaks your website.

Here's a recording of a 2020 presentation ("Securing the Future of Safety and Security of Embedded Software") from NVIDIA at the AdaCore conference:

https://www.youtube.com/watch?v=2YoPoNx3L5E

What coding model are you using for the rewrite? Opus for everything? A prerelease model like Mythos?

Once Spotify has this AI metadata, I'd like a setting to hide AI "artists" from my recommendations and playlists.

Thanks! I’ll check out his other books.

I have the first edition sitting on my shelf, waiting to be read. I didn’t realize there was a new edition!

Do you recommend I upgrade to the second edition instead reading the first edition I already have? Amazon says the first edition has 720 pages, while the second editions’ volumes 1 + 2 have 1186 pages!

I use the "Unhook" Firefox extension to customize the YouTube page. You can hide shorts, comments, recommendations, the home page feed, and more.

https://unhook.app/

A post about big-endian testing with QEMU was posted on HN just a few days ago:

https://news.ycombinator.com/item?id=47626462

https://www.hanshq.net/big-endian-qemu.html

Clang’s C++2c implementation status page simply says “no” for both reflection and contracts. GCC’s says “yes”.

https://clang.llvm.org/cxx_status.html

https://gcc.gnu.org/projects/cxx-status.html

> Wisdom I received from, IIRC, the Perl documentation decades ago

Perl is a language renowned for being difficult to read and maintain.

OpenCode's creator acknowledged that the ease of shipping has let them ship prototype features that probably weren't worth shipping and that they need to invest more time cleaning up and fixing things.

https://x.com/thdxr/status/2031377117007454421

I was curious why some expressions in the code used the character ù, such as “If Zù500”. It looks like a character encoding error, but the code presumably works correctly. ChatGPT says the byte value for ≤ in TI-BASIC is the same as ù in ANSI/Windows-1252 (0xF9).

Looks like LLMs also find Dafny easier to write than Lean. This study, “A benchmark for vericoding: formally verified program synthesis”, reports:

> We present and test the largest benchmark for vericoding, LLM-generation of formally verified code from formal specifications … We find vericoding success rates of 27% in Lean, 44% in Verus/Rust and 82% in Dafny using off-the-shelf LLMs.

https://arxiv.org/html/2509.22908v1

And they can get “too build to fail” bailouts from the US government after they are dependent on ChatGPT.