Wishlisting is bookmarking and always has been.

Ah my bad. I thought this was the Codex app!

it might honestly just be a config toml issue. Have you tried asking ChatGPT to inspect its own config to see why it keeps coming up? That might help, i've used it to help resolve its own configuration problems before.

I don't build games. I'm one of those people when it comes to apps. I'm so tired of the use of AI becoming a personality or a lifestyle. I'm just not interested in dealing with it. Life is short. I'm just going to make software for me that does exactly what I want it to and makes me happy.

It's not even an issue if AI writes it. It's just that people don't know what good writing looks like. And because of what AI is trained on, AI thinks good writing is persuasive marketing copy, even in a freaking GitHub README. You could be utilizing an LLM so much better to condense down the README into exactly what it needs to be with no fluff and no BS, making sure that you're not actually missing any important details, but that's not how people use it. They love all the shiny emojis and all the other stupid em-dash usage and everything else that goes into it.

I would absolutely hate to be the support for something like this. Doing it myself is fine, making it a paid product? Imagine doing something that either wrecks somebody's expensive several hundred or several thousand dollar TV and now you're on the hook for fixing it. And that's before we get into any of the other annoyances that come with this..."[completely unrelated thing] was working just fine before you did this!!" Maybe it's a small issue, maybe it's a big one, but what if you can't fix it without physically connecting to it? Are you going to have a plan for doing remote ADB work with them? I get what you're saying, but if anybody's looking at this thinking about commercializing it, I would strongly think twice.

I think it's the height of disrespect that the author of the DLSS 5 swapper didn't pick a different name. You think there'd be some camaraderie and etiquette there.

Who knew!?

You should absolutely be centralizing all of your hooks and skills and then symlinking locations for the various harnesses that don't support the universal locations. I don't think any of them differ so much in their hook or skill syntax that you couldn't do that, right? Granted, I use a lot more skills than I do hooks.

Anyone who is literally copying and pasting code into a chatbot instead of using an agent probably didn't know what they were doing in the first place

"A Google AI overview once told me to put glue on pizza because it holds the cheese better. After mocking it on social media, I never tried it again. AI must still be that bad." <- an alarming number of people, based both on personal experience and a sneaking suspicion.

Unless you managed to root and/or reflash your TV's ROM, you absolutely still have ads, bloatware, spyware, etc. If all you've done is install a custom launcher none of that stuff is missing. Still cool thing to do, but I think it's worth noting since this trend is suddenly blowing up. Anyone who has an Android TV with access to the play store can install Projectivity for free and get this "speed boost" without any extra configuration: play.google.com/store/apps/d…

Breach *related* to a distillation attack. The distillation attack was not the breach itself. I can't understand why so many people are misreading this.

Damnit Jordy, every time I think about building something new I see Droppy now has it and I'd already doing it better than me. 😂

Would absolutely love to see OMP support!

Jev can do this if it's more aggressively usied. Josh Pigford made a pretty solid Chrome extension for it. I'm wondering if you can extend it further.

Verification can't stop idiots from copying and pasting stuff out of LLMs. "Read this Twitter thread and give me an engaging comment to make" It is not exactly difficult to macro or hotkey or turn into a skill.

Given how much X encourages and literally pays money for engagement, they probably don't need all that much to come out ahead. Besides, people like Ethan probably aren't the target. It's just like Reddit bot posting. They're looking to build up their reputation and game the internal ranking so when they want to do something that actually involves real money, they think they're going to be better positioned.

There's absolutely nothing substantive about the comments in his screenshot. You can instantly tell when someone is making a comment just for the sake of making a comment. A novel idea here: not everything needs your take. And I'm not referring to you specifically, I mean in general. There's a certain class of people that you can tell are using LLMs because LLMs are not designed to shut up. LLMs are designed to leave comments, to engage. Intelligence is soon going to be determined by your ability to know when to not say something, and that's how you're going to stand out in this sea of pretenders who use LLMs to say way too much.