QWEN 3.6 27B can run fine on a 16GB video card and if you give it more time it’ll be as ‘smart’ as bigger models.

As much as I’d like this to be true (don’t believe all the benchmarks), in reality, using e.g. gpt 5.5 is still a lot less pain in the ass, mostly has to do with more reprompting (gpt is just smarter, oneshots stuff more often) + a lot slower (on an RTX 3090 for reference).

I’ve tried using it for some time, but I think I’m faster writing (better, although that’s also true for gpt-5.5) code by hand, than using this (+ I need the valuable VRAM for other stuff, as I’m a graphics/shader programmer most of the time).

That said, it’s already fairly impressive how much progress these smaller models have made the last year, it’s usable, you can “vibe-code” at least simple stuff.

Right now it just makes more sense to not use the API directly, but the subscriptions, they seem to be better priced. Also, Anthropic seems to be more expensive…

How that will look in the future, I don’t know, I’m fairly sure they’re progressively increasing the price, or rather reduce the amount of tokens you can use in the subscription (as they did with gpt 5.5 and claude opus 4.7). The chinese competitors are getting increasingly more interesting. It’s also quite impressive how well small models like qwen 3.6 27B run already on a (not so affordable) 24+GB GPU, unfortunately still far from the quality of say gpt 5.5, but probably comparable to o4 or something like that, certainly usable.

Right… But in the age of AI, data-harvesting the right data (i.e. the human/non-AI-slop) becomes very interesting to a lot of companies, “age-verification” is an easy argument (for policy-makers) of e.g. social-media companies to know whether the user is an actual human, thus the verified data is a lot more valuable.

Around 15 kCal/kg daily, maybe not very high indeed, but it’s significant, say I have 10-15kg more muscles than the average untrained person, that would be 150-225 more calories at rest, also you’re not always resting as you say already, the number will likely be higher, when considering daily activities, also since there’s more strength, it’s likely that the muscles are more used when doing sports (faster running/cycling, more weights, etc.), which is also relevant.

(For my case, I have a very inefficient metabolism likely (rather results in heat than fat-increase), I eat a lot even when I’m not trained, but right now, as I’m a little bit more packed I eat absurd amounts of food (due to a lot of drumming mostly btw. wouldn’t have thought it results in that much extra muscle/calories-burning))

Actually the more I have read into what is a healthy active routine (more on the athletic side of things, because I have overloading issues) Cycling to work, is like the perfect active rhythm for the body.

Especially when you do it fast (VO2 max).

The body needs time to recover and needs just little stimulus for growth/maintenance (so marathon is not good for your body, it’s just wear and tear), and after ~6-8 hours, needs stimulus for further growth again (so cycling back from work).

I have always wondered why I was comparatively muscular to others (for some part it’s likely genetics), I think it was/is because I cycle like a complete madman to work (well… VO2 max), because I’m an adrenaline junkie.

I just recently discovered that this is actually quite healthy (well not in the casualty sense obviously, but for cardio and general health).

If you’re interested I can recommend this talk: https://www.youtube.com/watch?v=anB-UMXIDQA

I think that just means, that you have changed your microbiota to like more healthy stuff (I too just don’t like unhealthy food anymore, but that took years…).

Carbs though are something that you really crave while being active, the muscles want to be fed, whether that happens via Oreos or say more healthy fruits or something like that is a different story.

Yeah and don’t forget that the newly built muscle needs more energy as well just to maintain.

I have to eat painfully much… (I’m quite active, and even when not I eat a looot more than others…)

There’s vaultwarden, I don’t think it’s difficult to migrate to that.

I have to admit I chuckled, because it’s probably true… (and it’s not difficult to dwell in cynicism these days…)

So you haven’t had “zero” problems. AFAIK wayland is already usable with AMD since almost a decade or so… (well not every program was supported yet a decade ago obviously, but, at least these kind of issues that Nvidia has/had are non-existing AFAIK).

It doesn’t fix the drivers but for many the installation and set up is where things go wrong. That’s how it was for me.

For me it was never the installation, just the risk after updating the drivers that yet another issue appears (sometimes old ones were fixed though, to be fair).

That doesn’t improve the quality of the drivers though… But you seem to not have had issues yet… Are you on wayland though?

There’s always a new issue. One time I can’t resume of suspending (I think this is still an issue…). Then shutting off a monitor leads to a crash of the driver-stack. I could go on. Just the fact that Nvidia took so long to support GBM properly is a tragedy.

In the US…

Different in other parts of the world, even in Europe this is a high salary (but it reminds me to still ask for a raise…)

The beginning of time…

I honestly think, it’s very cool for prototyping ideas at this point. It’s also parasitic. Although I think because of (maybe) different reasons: It gives people the power (which they unfortunately use way too much) to imitate an art, but in an non-arty imperfect way that doesn’t comprehend details (of the art), resulting in slop. For software that can go very wrong as we see here. This is also a reason why I mostly quit open-source, because now everyone can code a bad version of a library, it sucked the art out of good open source etc. and it’s increasingly difficult because of good wording/“look” etc. to differentiate on quality of code, previously you could often check a code-base review it somewhat and know how good the quality is, now it’s more like “is this slop or not?” (in which case I go a big circle around it, because reviewing is often not worth it)

At some point though, I think this automation of work is inevitable, we need to think about a society that can peacefully exist without having the requirement to work to exist. I actually think this could easily be utopian, everyone can focus on what they actually think is fulfilling life.

Though, it’s sad and concerning that technology is developing faster than society can adapt, which is why I’m mostly with you, because people (or representatives like politicians) just aren’t “programmed” for these fast-paced changes, to adapt the technology such that the future may be more utopian as it currently is heading towards a dystopian future…

I only trust math with my data (i.e. encryption…)

Yeah true, I actually also have thought about trying muscaria, but I’m fearing a little bit the anticholinergics group of drugs… (among others because of potential to increase risk for dementia etc.)

Not only that mushroom, the death cap as the name suggests has similar effects (same toxin), but is actually the most deadly mushroom. But yeah Amanita are the most deadly/toxic species of mushrooms.

Munich with their new major also had a win in this direction just now. I do think that local policies likely are more important anyway for at least traffic policies.

Don’t wanna be that guy, but: You misspelled Microslop.

US is liked

Brave assumption, there might be some people outside the US who like them, but I guess that’s true for all countries. I for my part started hating the US, it’s such a bad development for humanity, since capitalism/oligarchy takes over. Even China starts to look good in comparison…

Also:

world stability for a huge portion of the globe

Just LOL. Probably the force that causes the biggest global instability. Do I have to mention Iran as most recent event?

There might’ve been a period where that might’ve been true, but it certainly isn’t anymore.