ktimespi

The biggest barrier I faced learning SIMD is the weird naming convention for intrinsics.

Once I got past that, it was fairly simple. mcyoung's articles were also super helpful

An interesting article about how Jarret (author of the article) came up with the precursor to FSRS (Free Space Repetition Scheduler), the algorithm used by default in Anki now

I find it very annoying when queue problems break into queue-of-queue patterns like in the `wait`scenario

I'm a fan of this project. Great work!

It does need quite a bit of time to be productive with... That being said, Rust by Example is good

Incredibly yucky of Zuckerberg do that. I found it very hard to read through that part.

Aren't decimal types BCD coded?

The major allegations are about Meta bringing authoritarians to power while shirking responsibility for making that happen

If this can learn my patterns, that itself is enough. Google Keyboard used to do this, but mysteriously doesn't do it anymore.

Pretty cool to read it directly from the associated device XD

Did you write a metadata parser for most of the filesystems?

If the disk decides to falsely report a flush, there's not much you can do about it from the user side, no?

Is this really that big of a moat for Figma to cross though? They would have a much easier time integrating AI than the other way around

finally.

I guess this is breaking ToS?

s3fs can do partial reads too with range queries, I'm leaning more towards the tradeoff.

This was my concern too. The whole point of using S3 as a file system instead of EBS / EFS (for me at least) is to minimize cost and I don't really see why I would use this instead of s3fs.

This is ideal for my use case, yeah. No need to fiddle around with another app's UI.

Yeah, the fact that she realized what's going on and still worked tirelessly to give Mark / Facebook more negotiating power speaks volumes. I also can't buy the whole "I have financial woes and can't escape" spin that she puts on her situation.

Otherwise, great book.

This is sweet!

I've thought about this a bit too, and instead of a straight escape hatch to html when any special formatting is necessary, it might make sense to bind css to particular elements like so:

##[color:red] Heading 2

or using code blocks with particular headers which render out as normal text with the styling applied, like this:

```styling: 'ruleset here...' Lorem Ipsum Dolor ```

this does require a lot of work to consider what might be ergonomic for a particular element (by element I mean ATX headings, blocks, etc.), though.

Yeah, realized this the first time I used an LLM to code. I've not used them since. No matter how good it gets, it's dangerous to lose touch of my own intelligence.

For me, receipt scanning and tagging documents and parts of speech in my personal notes. It's a lot of manual labour and I'd like to automate it if possible.

It's a major disservice to me as a paying user too.

I've had a very pleasant experience with Tailscale for self-hosted applications too! I will recommend Silverbullet (https://https://silverbullet.md/) for note-taking.

The fact that I couldn't turn off shorts recommendations on youtube is just so, so annoying. It's such a time sink and I'm glad that the tides are finally shifting against addictive algorithms like these.

sounds like AlphaDev [1] might be a better approach for a problem like this.

[1] https://github.com/google-deepmind/alphadev

I've heard Zulip is a good alternative

He realizes that they can't move data out of Slack, and asks for another corporate product that has the potential to lock down the organization's data...

I don't think it is true either, considering the broad claims made.

The thing to be changed is research incentives, not getting the bar even higher. Take the Francesca Gino case, for example. I don't think anyone can argue that Harvard's bar is "not high enough".

This comes across as elitist