cmgriffing

I can't be the only person who saw "NUMA" and then read the word "memory" in the title as "Memora-hee, Memora-haa".

We should thank companies for warning us during the interview process that they are so separated from reality (especially in the AI era)

If AI can’t make them recognize a work life balance has value then it’s easy to see they don’t believe the “force multiplier” BS they are peddling

I remember having to do this with icomoon back in the day when icon fonts were all the rage because svg support in ie11 was lacking and downloading the entirety of font awesome was a huge payload hit.

Ollama cloud is also a fantastic deal. They also don’t have a monthly limit, just session and weekly.

Only slightly related, but I have been using Oban as a Postgres native message queue in the elixir ecosystem and loving it. For my use case, it’s so much simpler than spinning up another piece of infrastructure like Kafka or rabbitmq

I shudder to think how small the macOS ui text will be on this but I’m willing to find out.

This might sound crazy but I started live streaming partially because I wanted to work on side projects.

You will be less productive than getting into a focus state off line. But 0.5x is still infinitely higher than zero.

On stream, you don’t get to randomly hop into a game or doom scroll social media or hackernews.

I was digging into the cursor effect just to see why it's so laggy for some people and noticed that this is actually a Next.js site.

That is news to me that Google is using Nextjs for anything.

I think this is a good case for applying Hanlon's Razor. The person that did the forking and removal of copyright text may simply not know that it needed to stay there.

I would love to know what processes MS is considering to prevent this in the future as well as what kind of auditing might be done to look at other projects that started as forks.

I would love to know of some of the companies you know of that have gone this route. Maybe nothing we've heard of, but still sounds interesting.

I recently discovered the way Tanstack does this. Basically, if the heading of the section is in the viewport, then the list item is highlighted.

So you could have multiple items highlighted, but it still "works" somewhat intuitively for the end user.

The drawback is that it requires JS via intersection observer. But maybe the CSS standards committee could see value in this kind of thing eventually.

Not weird to do all that. Just weird to not disclose it.

Promoting it? No problem. But promoting something you profited from without disclosing it violates FCC rules for broadcasting. I would say influencers aren't technically broadcasting but they are in principle.

While his formal affiliation may have been short-lived, do you think he got a cut of the sale to Shopify?

If so, not disclosing that when he promotes Remix is a bit shady.

Nice dude and all, but that is one thing I take issue with still.

Use AI to spot blind spots

Tools like ChatGPT can highlight UX issues you might miss. It’s a quick sanity check—not perfect, but better than guessing.

Tools like ChatGPT can highlight UX issues you might miss. It’s not perfect, but it’s better than guessing. Some prompts to try:

Was this an intentional joke?

Their first key insight is interesting. It says that coding ability predicts model performance in the game. I wonder if it also predicts performance of human players in some way?

FNV-1a might be worth looking into. It's pretty fast

We are destroying software by directly contradicting ourselves when we describe how we are destroying software.

LOL, you and I posted very similar replies at the same time.

Quick note: when trying to do full page screenshots, Chrome does a screenshot of the current view, then scrolls and does another screenshot. This can cause some interesting artifacts when rendering pages with scroll behaviors.

Firefox does a proper full page screenshot and even allows you to set a higher DPS value. I use this a lot when making video content.

Check out some of the args in FF using: `:screenshot --help`

This brings up an interesting thought too. A photo is just a lossy representation of the real world.

So it's lossy all the way down with LLMs, too.

Reality > Data created by a human > LLM > Distilled LLM

If these tools had a runtime check, then the cache key creation would have failed out.

72 is the max length of id, username, and password combined. If that combination is over 72, then failure and the cache key would not have been created. So, no, the attacker would not need to guess only one character of a password.

Oh interesting. I hadn't heard much about it recently. Thanks.

Consider the Volkswagen scandal where code was written that fudged the results when in an emissions testing environment.

The only person to see major punishment for that was the software dev that wrote the code, but that decision to write that code involved far more people up the chain. THEY should be held accountable in some way or else nothing prevents them from using some other poor dev as a scapegoat.

Take my upvote :)

In the case of Restream (or cases like their "stage"), thats not always doable. What about superfluous hero animations that marketing sites love to do these days? Pausing some generic canvas rendering logic is usually more annoying than you would want.

Be careful using:

    dialog::backdrop {
        backdrop-filter: blur(2px);
    }

It seems you can use the transform3d trick to kick it to the GPU to help fix it.

I feel like the value of changing from `git checkout -b` to `git switch -c` is not worth the cognitive overhead of making the change.

I'm using them on a new mac mini. Haven't tried on my older intel mac mini yet though.

Awesome! Thanks!

For anyone else curious about how to configure these, check out the docs: https://ghostty.org/docs/config/reference#custom-shader