fermuch

The screenshot from the page itself (the split original/translated) does not translate the image labels. Is that on purpose? It only translates text and does not do OCR?

I guess it is a joke site but the questions are very real, maybe a bit too focused on databases and javascript. Is there a purpose for all of this?

I remember people using Yahoo and AltaVista a lot. And before gmail everyone had a yahoo email (if they didn't already have a hotmail). Times change, kings fall from grace.

xhigh tells it to overthink and re check everything. Low tells it to only do the minimum thinking necessary. I would suggest to give qwen medium which doesn't inject any thinking directives into it and also to give as much context as you can, ideally around 500k tokens or even 1M if you can. Big complex tasks like these make the model hit the compaction trigger a lot and they end up re thinking the same thing several times in my experience.

I think /handoff on pi (or at least oh my pi) is what you are looking for

devenv also has tasks/services. For example you need to start redis, then your db, then seed it, and only then start the server. All of that could be aliases, yeah, but if you define them as aliases you can have them all up with `devenv up`. It even supports dependencies between tasks ("only run the db after migrations ran")

Kagi Translate is now available as an app for Android and iOS!

The app supports over 248 languages and offers context-aware image translation, live voice-to-voice conversations, and a rich dictionary with audio, to name just a few of its features.

I use my Meta Quest 3 with Termux & Termux X11 to run full desktop apps, and it works like a charm. No emulation, too! Native ARM64!

I wouldn't put it in past tense...

The only thing you'd achieve doing that is to change the "main function" of a device to somethings silly, like a thermostat being sold as an art decor with the optional additional of functioning as a thermostat too.

A lot of times they are converted to BigInt and then stored as strings so different languages don't mess with the real value

Materialized views are a great tool for aggregating data in CH since they are automatically updated on insert from the original table. I recommend you to take a look and try it out, maybe it'll go down to single digit milliseconds!

3. Normal people accept surveillance and techies are tracked down / silenced.

Given our current surveillance state with social networks, I don't see (1) or (2) as real options.

Drowning in my own shower is a new kind of fear you've just unlocked on me

The link isn't working for me. For those who were able to see it: does it improve anything by using that instead of what ls does now??

It seems to be global. Google cloud, anthropic, everything. Even AWS it seems.

Under the "tools" part of the README it shows the following observation tools: - browser_snapshot_dom - browser_query - browser_accessible_tree - browser_read_text - browser_screenshot

So most likely the LLM can chose how to "see" the page?

Yes! There is a benchmark on ClickBench: https://benchmark.clickhouse.com/#eyJzeXN0ZW0iOnsiQWxsb3lEQi...

Going from memory, the GP comment is from a Markiplier video where he created a "game" in which you choose what path to take.

On the release stream he was asking people to "vote" which path to take, and IIRC it was by writting blue ball emoji or red ball emoji.

People were banned and he took it personally. Contacted his internal youtube contacts and all. There should be videos about him talking of it.

How is this different to winlator? It seems to depend on termux

> The only true random sources in the universe are quantum. But it's not practical or performant to hook your workload up to a raw quantum measurement like the spin of a fundamental particle, or radioactive decay.

I was under the impression that's how entropy generators for servers work?

Isn't that how china works? One timezone for everyone

Official support, and more eyes seeing it. Maybe better integration since they'll need to make it integrate fully to reach mainstream kernel.

I've tried to use `fetch()` and it works. How did you handle that from WASM? Or is it cached?

This seems like a cool idea which might see some uses in the future. For now it seems like docs are lacking details, so I could not figure out how it works.

Does it only support risc? I saw the example written in C++. Which versions are supported? Is it common to use in kernel development? (AFAIK only NT uses cpp, right?)

Anyways, cool project. I could see it as an alternative for embedded development in the future.

She is also great at communicating her advancements, which is a greatly useful skill to cultivate.

I think the idea is something like "let things run normally until they don't need that extra core. If it doesn't need the core, we shut it down. And it never wakes up"

Which... Makes it hard for the user to notice the change? Cores will naturally go offline, but performance stays the same while you still use it.

I think this is just a demo of what you can do with user space scheduling and shouldn't be taken seriously...

Just like adding code to textual models helps the model develop its reasoning capabilities, it seems like adding more languages helps in other areas too. What is needed is more good quality data to train on...

Would something like this apply to MAMBA/JAMBA too?

No? I haven't tried it yet, but in the docs it says "SQLite3 file format, and import and export facility"