Skip Navigation

InitialsDiceBearhttps://github.com/dicebear/dicebearhttps://creativecommons.org/publicdomain/zero/1.0/„Initials” (https://github.com/dicebear/dicebear) by „DiceBear”, licensed under „CC0 1.0” (https://creativecommons.org/publicdomain/zero/1.0/)S
Posts
4
Comments
229
Joined
6 mo. ago

  • On the wrong end of 40, I'm afraid. :P

    I hope you listened to the song, wore sunscreen and actually were kind to your knees.

  • Correct. The title is clickbait. The actual content doesn’t support the “all AI is bad” fever being stirred up.

    In any case, governments banning access to cloud-based AI is nothing new - Italy and the Czech Republic both come to mind (though I think the later was a knee jerk reaction).

    At this stage, it's probably a good preventative measure to not send sensitive information to the cloud (aka someone else's computer), especially a cloud in a foreign territory.

    Common sense, that applies from individual to government level: don't yeet your sensitive data helter skelter. Host your own secure infrastructure.

  • Could you elaborate a bit more on the issues around shortages of funding and professionals?

    My understanding was that Cuba has had one of the highest doctor-to-population ratios in the world, and that its medical schools were often treated almost like a public works project to train large numbers of physicians.

    Is it an issue of brain drain (graduate qualifying and then leaving for greener pastures?)

    (I'm not doubting what you say btw; just trying to address my own ignorance).

    PS: El Paquete is exactly the example of "fine, I'll do it myself" response I had in mind. At 1TB per week, that was/is one of the largest sneakernets in the world.

  • Doesn't Cuba punch well above its weight class when it comes to medicine, including universal free medical and dental? IIRC, hasn't the US embargo basically fostered the development of Cuba's own burgeoning biotech industry?

    https://en.wikipedia.org/wiki/Healthcare_in_Cuba

    I am sure that the US blockade has made life exponentially harder. Self reliance is one thing but access to raw materials is another.

    OTOH they seem like a resilient people. If there's anywhere on Earth that exudes "Fine; I'll do it myself" energy when dealing with America, Cuba is it. I hope this makes them even stronger. I hope they find a way to completely circumvent America. I actually hope that for the whole world.

  • I would have like to have read more about the examination and in what ways the medical students were under performing thereon.

    One presumes that the Brazilian cohorts are passing their national medical entry and exitexams (per their ministry of education?) and fulfilling their practicum requirements for licencing (per WHO/WFME recommended standards).

    If so, this raises several interesting questions.

  • You mean capitalism I think.

  • For sure. And in that spirit -

    Another option for people: pay $10 to get an OpenRouter account and use providers that 1) don't train on your data 2) have ZDR (zero data retention)

    https://openrouter.ai/

    You can use them directly there (a bit clunky) or via API with something like OpenWebUI.

    https://github.com/open-webui/open-webui

    Alternatively, if you have the hardware, self host. Qwen 3-4B 2507 Instruct matches / exceeds ChatGPT 4.1 nano and mini on almost all benchmarks... and the abliterated version is even better IMHO

    https://huggingface.co/Qwen/Qwen3-4B-Instruct-2507

    https://huggingface.co/DavidAU/Qwen3-4B-Hivemind-Instruct-NEO-MAX-Imatrix-GGUF

    It should run acceptably on even 10 yr old hardware / no GPU.

  • I have ASD; I made several tools that explicitly convert web sources to .md and JSON.

    The shitty thing is, a lot of sites - even if they have stuff available in simple, beautiful JSON format, refuse to give public access to it. Notoriously, movie session times for local cinemas. That should be a simple look up...but no.

    Oh well, at least cool shit like this still exists

    https://github.com/chubin/wttr.in

    https://github.com/scrapy/scrapy

  • I mean...you could just get them to deliver what you need :) Or try your local hardware store (if one still exists near you)

  • Fallout for me...but curiously (or not) my media server recommended "Jericho" and "Falling Skies" as thematic matches / worth a rewatch.

  • Agree. Unless Sama has the mother of all rabbits in his hat, I dunno how they unfuck themselves.

    OAI is betting hard on AGI...but AFAIK they're trying for it by "do the same shit, harder and faster". Man, I dunno.

  • Yeah me too. Opus 4.5 is awesome but my god...om nom nom go my daily / weekly quotas. Probably I should not yeet the entire repo at it lol.

    4.6 is meant to be 2x worse for not much better output.

    Viewed against that, Codex 5.3 @ medium is actual daylight robbery of OAI.

    I was just looking at benchmarks and even smaller 8-10B models are now around 65-70% Sonnet level (Qwen 3-8, Nemotron 9B, Critique) and 110-140% Haiku.

    If I had the VRAM, I'd switch to local Qwen3 next (which almost 90% of Opus 4.5 on SWE Bench) and just git gud. Probably I'll just look at smaller models, API calls and the git gud part.

    RTX 3060 (probably what you need for decent Qwen 3 next) is $1500 here :(

    For that much $$$ I can probably get 5 years of surgical API calls via OR + actual skills.

    PS: how are you using batch processing? How did you set it up?

  • Woof - the axes on that chart LOL. Suffice it to say, they're all pretty dang close. Interesting. Maybe the easter bunny can bring me something with >8GB VRAM so I can actually run em locally. I'm guessing Kimi-2 eats about what...500GB+ for 128K context?