etra0

Just a few months ago I set-up a serverless gpu service just using whisper.cpp and it detects very well my very-local spanish modisms. I have an active program running on a vps that hits the serverless gpu with the audio when needed because I went with a telegram bot instead of building an app (and paying for dev fees, yikes!).

I wish at some point apple had something like that for in-device STT but so far I'm very happy with the solution.

Same, I think there's too much friction yet. Maybe at some point it could become a crosspoint feature I hope!

EDIT: Just checking on the crosspoint site they say such thing is not planned: https://crosspointreader.com/roadmap (section Out of roadmap: Active connectivity (RSS, news, browsers))

More than a special status as a host country, as far as I understand it is explicitly not a member state, hence it's separated from the rest of the flags (Which is curious because despite being the host country, it doesn't guarantee you being a member state of ESO).

Apple never released a driver that supports Vulkan out of the box, therefore MoltenVK was born, which translates Vulkan IR to Metal source then recompiles to Metal shaders.

This was the pipeline in Proton for macOS (I'm not sure if it's still is the case, been quite a while since I checked).

I saw this a while ago so it might not be totally related, but Sebastian Lague did a video on atmospheres for his planet generation experiment which was also very entertaining to watch [1].

There's something particularly entertaining on developing visuals and watching them come a reality — I hope at some point be able to experiment in this field.

[1] https://www.youtube.com/watch?v=DxfEbulyFcY

haha did something similar. Ended up vibe-coding something with Hugo and using backblaze + cloudflare proxy to host the images. So far everything has been free and I have a snappy 'portfolio' :-)'

I've tried Irys as well but the mobile only is kind of a deal breaker for me — I like seeing images in the big monitor to appreciate them more.

Of course I also have my webpage to showcase my favourite pictures but I feel I'm more picky in that site than in, say, Glass and instagram, since I want to show 'the best' there :-)

Lately I've been enjoying photography a lot but Flickr never clicked for me. Instagram nowadays is almost unusable for this as it prioritizes reels too much and 500px... I liked that one more than Flickr.

Right now, I'm using glass.photo and I actually quite like it. You have to pay, though, which is a high entrance barrier, but I feel the quality of what I see in the site is great, the platform works nicely and the community has been welcoming so far.

I yearn for a good site to share and comment photos which is a bit more open, though.

I did run with this setup for a few months (I believe like, 5 months already?) and when it works, it's nice, but 90% of the time it has been extremely painful.

Something breaks, one automatically updates and then it breaks the entire database, SCRAM mode, recovering is painful, and all the time I get warnings, spam and logs, it's anything but seamless.

Which is a real pity, because when it works it feels magical to use within my laptop, my phone and my tablet, all self hosted, but the pain won and so I'm searching for new alternatives.

Came here to appreciate the same. Not only it truly captures the scale, but does it in a great way.

Same for me, sadly.

One of the reasons why I learned vim was because I enjoy staying in the keyboard; I'm a fast typer and part of the fun is typing out the code I'm thinking.

I can see how some folks only really like seeing the final product rather than the process of building it but I'm just not cut for that — I hate entrepreneurship for the same reason, I enjoy the building part more than the end.

And it's the part that's killing me with all this hype.

non-native english mistakes and my stubborness to check with any LLMs :)

You have to realize, that some folks were born in the age of super-information and immediacy.

When I take a picture, I get the luxury of immediately see what I got. When I wanted to hear some music, I can search it up, and hear an entire album in a question of seconds.

It's an incredible privilege to do that, but at the same time, we got so used to speed, that pausing can be new for us.

This year I had the opportunity to travel to Europe and just sit in a café, sipping coffee, just observing, and it also felt new and different for me.

I shot with an analogue camera because I enjoy the feeling of waiting for the results, not being able to see the results at the split of a second.

This blog resonates with me because I've been feeling I want to pause more, to create more memories, to be in the moment. I should go to a café without phones and a notepad.

Well, last time the US intervened in my country it was to actually install a dictator that would benefit them.

It's too early to say it's a win-win situation.

The tip that I've been using quite a lot lately by debugging long log files is using `&` to filter what I want to read and `&!` to filter-out what's not useful (and they support regexes).

Admittedly, they are a bit slow sometime and sure, you could use `grep -v` then pipe which is way faster, but they've saved me on removing noise from logfiles from time to time when you don't always know what to filter beforehand :).

EDIT: It was in TFA.

That is exactly my point, though.

I didn't mean they do it on the first time, or that it is correct, I mean that you can 'run' and 'test it' to see if it does what you want in the way you want.

The same cannot be said to any other topics like medical advice, life advice, etc.

The point is, how verifiable is the output the LLM gives and so how useful it is.

Completely off-topic but I just love how the pettiness of Musk was abused by twitter community.

I had a chuckle reading all of these.

> 2) Ask an LLM, which gets you 80-90% of the way there.

Hallucinations and sycophancy are still an issue, 80-90% is being generous I think.

I know this is not issues of the LLM itself, but rather the implementation & companies behind them (since there are open models as well), but, what limits to LLMs to be enshittified by corp needs?

I've seen this very recently with Grok, people were asking trolley-like problems comparing Elon Musk to anything, and Grok very frequently chose Elon Musk most of the time because it is probably embedded in the system prompt or training [1].

[1] https://www.theguardian.com/technology/2025/nov/21/elon-musk...

> I find it hard to trust anyone with any kind of profit motive.

As much as this is true, and i.e. doctors for sure can profit (here in my country they don't get any type of sponsor money AFAIK, other than having very high rates), there is still accountability.

We have built a society based on rules and laws, if someone does something that can harm you, you can follow the path to at least hold someone accountable (or, try).

The same cannot be said about LLMs.

I haven't tried with this specific topic, but being the pleasers llms are, I doubt someone so focused on being anti-vaxxer will be convinced by an LLM, if anything, the LLM will give them reason at some point.

LLMs have certainly become extremely useful for Software Engineers, they're very convincing (and pleasers, too) and I'm still unsure about the future of our day-to-day job.

But one thing that has scared me the most, is the trust of LLMs output to the general society. I believe that for software engineers it's really easy to see if it's being useful or not -- We can just run the code and see if the output is what we expected, if not, iterate it, and continue. There's still a professional looking to what it produces.

On the contrary, for more day-to-day usage of the general pubic, is getting really scary. I've had multiple members of my family using AI to ask for medical advice, life advice, and stuff were I still see hallucinations daily, but at the same time they're so convincing that it's hard for them not to trust them.

I still have seen fake quotes, fake investigations, fake news being spreaded by LLMs that have affected decisions (maybe, not as crucials yet but time will tell) and that's a danger that most software engineers just gross over.

Accountability is a big asterisk that everyone seems to ignore

It works wonders! I build free-cameras and some other tools (all for offline games, of course) fully in Rust, and you'd be surprised how much you could do.

In one of them I hook into C++'s inheritance with no issue, just by understanding how everything works within the compiler you can do a lot.

I guess if you take it literally, yeah.

But I've seen plenty of boring pics of lakes and none were on top posts, contrary to these cases.

It is of course subjective what makes a good photo or not, but sometimes it is pretty clear why a picture reached top posts.

I can totally attest this.

I was an avid viewer of r/analog. I don't know if this was 'recent' or not, but every time someone post a naked picture, either good or not, it goes rapidly to Top posts.

Even though it used to had many comments like "This photo is not interesting other than the naked woman", the upvotes arrived anyway.

I think nowadays they mostly block the comments in those posts, but what used to be an inspiring subreddit that would pop from time to time in my feed, is not longer that interesting to me.

Kinda funny that it requires both "please" and "backport" for it to be considered haha.

I've used it for fun to analyze some data of some community in real time without having to host anything on my side.

Doing data analysis in javascript feels a bit weird at first, but it allows you to write a bit more functional code than python which I ended up liking.

I dislike the plotting api but the pro of being automatically updated without me needing to do/host anything is cool.

Big shout-out to WinCompose, it's the only way I found my keyboard usable while being bilingual :)

> As long as they don't start with Apple.

I guess it's just as good as any other of the vendors you mentioned. I don't see why we shouldn't start with Apple but at the same time I don't think anyone opposes to the other companies being forced too.

At least I know I would like to run personalized software on my Switch without having to rooting it by other 'ways'.

It is not whitespace sensitive afaik. This same version works as well

    loop do
            case message = gets
                            when Nil
            break
        when ""
    puts "Please enter a message"
        else
            puts message.upcase
        end
      end

(Scrambled on purpose).

LinkedIn posts really read like an alternative reality (which I would not like to be a part of, lol).

I cannot take seriously most of what I read over there. The comments are also often toxic, the whole business is... just weird.

What's funny as a personal anecdote, I've found more jobs through Twitter (pre-X) than through LinkedIn.

Seriously. And I've tried using LinkedIn for job hunt.