qingdao99

This is/was fairly common, I've experienced it on the Chinese subway a few times and I've seen a few clips of it happening online. No idea if it's official policy or not, though.

I think this account is a bot.

Anki supports HTML and JavaScript so technically anything is possible.

When it's expected to be handling reminders, calendar events, and other device functions for millions of users, it will be considered critical.

Some Gen Z hate audio messages and will not listen to them out of protest, others will send them every day. For most, they're just sprinkled in occasionally in primarily text chats (with their friends - likely not with anyone they're not close with) when there is some reason to send a voice message, e.g. conveying emotion or when typing is difficult.

That's my experience! n≈30 Gen Z.

Blocked for me! Virgin Media is my ISP. Maybe your ISP is less restrictive/compliant (not sure if the block is actually mandated).

I think you should implement client-side prediction, it feels very laggy.

https://developer.valvesoftware.com/wiki/Prediction

This was my guess. I believe I have experienced this as an adult too: I played on a special Minecraft server (an anarchy server) for a year or so, and despite spending so much time on it, I can hardly remember my first few days on there. I have no memory at all of being at the spawn location. I believe it's because of what you described - as I continued to play on the server, my understanding of how it worked changed completely so my initial memories no longer made sense under my new framework.

You could probably find a company or a person in Taiwan to ship them to you from Taiwan.

What did they mean? Can't figure it out.

I've always thought that you could probably just walk off somewhere else once you're on the tarmac.

as long as you'll pay for it

And it's far better in my opinion - ChatGPT actually creates its own prompts from what you give it and feeds those to DALL-E (you can see this by clicking on the images it returns and reading what the actual prompt was), although you can disable this by just firmly telling ChatGPT not to modify your prompt.

Additionally, the image generation itself is quite different. ChatGPT's DALL-E seems to create much more stylized images - much harder to get plain shots that don't heavily embellish your description.

You will casually breach the GFW if the VPN you pay for is not blocked yet.

Wow, which game?

It says the same for the comment you replied to. I don't think it's accurate for this task.

> don't concatenate user input and logic unless the bounds between the two are guaranteed to be explicit

Which is achieved how for an LLM?

Not an expert but:

The lag compensation is, more intuitively, just allowing the client to move on its own without needing server verification yet. The server maintains the authoritative game state, but (assuming no cheating is going on) the clients are able to move as they think they will. In the background, the client is actually sending its inputs to the server, and they technically haven't happened until the server verifies them (however many milliseconds later), but we allow the client to move on its own regardless.

The yanking back only happens if the client's new state (position etc.) don't match the server's state.

I would recommend signing up to GPT-4 API access and, upon hopefully getting it, using a third-party frontend like https://bettergpt.chat/ rather than the official ChatGPT page.

You'll never hit a ratelimit as far as I can tell, and it's usage-based so it will probably come out cheaper than $20/mo for regular usage.

I got GPT-4 access within just a couple hours of release and I'm sure I didn't write anything interesting, I probably just ticked the box for "student" or "hobbyist".

Edit: Nevermind, it took ~2 days. Point still stands, I just checked out the form and I definitely ticked "General exploration of capabilities", which is probably the least interesting option if OpenAI is being selective.

I think "high quality, well-prepared sources" would include blogs and articles, which are likely to become heavily influenced by AI (blogs and articles are high quality compared to Reddit posts for example, which were included in the past).

In fact, there's no reason to think that academic papers won't start using language models to write better.

Tainting your text with AI can be as simple as pasting a paragraph in and asking if there's anything to improve.

> But once one exists you can use it to make more relatively cheaply by generating training data.

Is that a real technique? Why not just shrink down the model itself directly somehow, is that not possible?

It's not defending bigotry to state that Antisemitism is not exclusive to Nazism.

You must not have spent a lot of time online in spaces with edgy teenagers. It truly is just about being edgy a lot of the time.

Indeed - edgy teenage online humor.

Antisemitism is not exclusive to Nazism at all. This is a news article. They don't just write things for no reason. They wrote "Antisemitic slurs" because they saw Antisemitic slurs.

Saying slurs doesn't mean you believe in a particular ideology. He may have just been saying them to be edgy (unless there is further evidence I have missed).

I think it just depends on how critical the software is. If it's just a webpage running in your browser then it's not a big deal if it crashes on some weird input. If it's a database, each crash can cause huge amounts of damage.

I like the optimism.

You must not be familiar with the biases that ChatGPT shows.

https://twitter.com/aaronsibarium/status/1622425697812627457 (refusing to support saying a slur to prevent millions of deaths)