SkoomaDentist
The Greater Finnish Empire
No bio...
User ID: 84
Even I, with a 7 years old second hand laptop can run surprisingly decent models. Any gamer with midrange GPU can run Qwen 35B A3B models at a decent speed (~20t/s or better) without having to have any particular technical skill beyond knowing the magic word "heretic" and doing a huggingface search for it. Anyone with a 24 GB GPU can run uncensored Qwen 3.8 even without any "dangerous company".
Anyone "dangerous" already has full access to those models right now if they only so desire.
soon the worst scum of the world will have very powerful and uncensored, non-refusing models running on cheap hardware
What do you mean soon? You can already run completely uncensored, non-refusing models on regular consumer hardware. Huggingface is chock full of heretic uncensored model weights.
The left has already won.
There is an important point that is being left out: The left that "won" is the US influenced left. Not the traditional "workers rights"-left but the "We must always consider the feelings of BIPOC LBGTQ people"-left.
Yes. Building stuff without human hand and finger dexterity is absolutely magic.
Pray tell how they are going to magically get manufacturing capabilities when even the best lab robots are not even remotely a match for run of the mill humans in most manufacturing tasks.
I assume there's a reason for the very common assumption that the vast majority of OpenAI's and Anthropic's homegrown software is vibe coded.
They and what manufacturing base?
Are you sure any places using per-house / apartment heat pumps don't already use them for cooling during summer?
Certainly here in Finland "cool your house during modern hot summers" is why 99% of customers buy a heat pump in the first place.
What is the AfD position on German Irredentism?
With AfD the problem is less their policy on German Irredentism than what their policy on Russian Irredentism is.
NVidia's moat is Cuda and having a large number of models in use around helps support that moat since any competitors will have to implement enough of Cuda to support most of them instead of higher level layer that only works for major closed AI providers. It's not so much that HF has much moat (which they only do by being the default in most software) but that ensuring HF's existence NVidia helps support their own moat.
Same here in Finland. And I'm not talking about some multimillion euro inheritances but just the casual "your parents were middle class and left a house / apartment to you and your sister"-level inheritance.
Stress hurts. I mechanically know the rules, but in high-stakes or in front of especially skeptical it's easy to ramble or try to add too many humanizing details that weren't in the original plan.
This is fairly easy to notice when it happens and totally understandable but not really what I mean.
I'm talking specifically about the things anyone can do that almost inherently will make their presentation / talk better simply by applying an almost mechanistic set of rules: Do X. Don't do Y. If you do those and "fail" due to stress, it'll come off as a one time failure instead of "man, this guy sucks as a presenter".
We train them wrong. From 2000-2005, it was common for introductory communications technology courses to emphasize animations, and forget to talk about why you should never fucking use them. I've seen college courses that encouraged people to make sure their slides had enough detail that they wouldn't need notes.
This might be it. Probably 80% - 90% of my teachers failed those basic things in slide design. Ie. never put too much on one slide and keep visual distractions to absolute minimum. Also whoever suggested adding animations to slides should be shot and their family cursed to the seventh generation.
A lot of people just seem to struggle fully modeling the extent other people might not share information they have, even if the know it at an intellectual level.
Aka "people are goddamned idiots". That would be the only thing that explained why people in the university kept repeating the same mistakes they were explicitly taught to avoid in a very easy to understand way on that course that every single student had to take.
Why are most people so godawful at presenting and demonstrating things when even the tiniest bit of knowledge and effort would make them 10x better?
This was inspired by a recent demonstration video I saw that took ages to get to the point while clearly not being about engagement or view time farming, but I see the same issue over and over again whenever people present things on stage or in videos. Even many professionals seem to be really bad at it.
A long time ago in a university far away we had a mandatory "Introduction to speech communication"-course that concentrated on the basics, including how to give actually good presentations. In retrospect, that course was by far the most bang for buck of any university course I ever took. After I took it, people started somehow thinking I'm a natural presenter by just remembering a few key points and applying them to my presentations without having any particular natural talent at it (and my presentations weren't much good either before the course).
It really isn't rocket science. Most of it could be summarized by "Get to the point. Less is more. Hook the listener by starting with the key idea briefly and only then cover any background and details. Repeat the key points at the end. Keep slides to max 3-4 points per slide and reserve anything else to auxiliary materials. Always show the entire slide at once so the viewer pays attention to what you say instead of wondering when the next point will be revealed. Show brief examples if applicable."
Really basic things, super easy to learn, super easy to apply with the tiniest bit of practise (if I can do it with my "Ackshually that's a very interesting detail, let's totally get sidetracked by it"-ADHD diagnosed brain, there's no reason most other people can't). Enough to easily get me into the top 10-20% of presenters and that includes people whose entire job it is to present and market things (like a marketing director at an old job saying "Skooma, why don't you present this since you're so good at it?" - in a group where half the people were from the marketing dept). Following those rules also makes presenting much easier as you have to do so much less (by leaving out pointless details and fluff and realizing nobody is going to remember most of it anyway so it doesn't matter if you leave out some tiny detail).
Yud has been wrong about a lot. IIRC he originally thought there would be a fast takeoff Foom event from some solo researcher or small team working on AI with limited access to compute. There’s a page out there with all his wrong predictions of which there are many.
This keeps happening to all the people (including many here on The Motte) who are all about theoretical computation and don't have the slightest clue about just how intricate and deep reaching the required manufacturing and design ecosystem is for all the required hardware. Nvidia and other such AI specific IC makers are just the small visible top of the iceberg. An AI "reaching recursive self improvement" won't do jack shit when it can't control the millions of people who are handling everything else. Likewise absolutely massive part of AI improvement is the training on real world data instead of just some abstract compute and without that data (provided directly or indirectly by humans), there is no improvement, because the AI is just playing in the tiny sandbox.
Not to mention the fact that incorporating new data into the models is a pipeline that takes three to six months as soon as it goes beyond trivial database lookup.
Yet it takes until Book 4 before we get a major character death, and Book 6 before we get one that Harry and the readers really care about!
Yes, that's my point. It gets established for almost five books that no one the readers truly care about is harmed, then Sirius gets offed at the end of book 5 and in the final battle of the entire series in book 7 more people get offed with no preparation and with nothing left of the book for their deaths to even influence. Those deaths just feel like a poor writer going "the tropes say I should kill some characters, so I'll just off them right like that".
In contrast Dumbledore's death is long foreshadowed and provides good progression for the story and thus feels like natural instead of being an ass-pull just to piss of the reader.
So it's not that character deaths or loss for the main characters are bad as such but many times they're poorly done and just make the reader feel bad.
I love downer endings
I, along with probably the vast majority of readers, dislike / outright hate downer endings. I hate writers killing off characters I care about just "because the emotional conclusion needs main character to suffer loss" and other such bullshit (see eg. Harry Potter). That's just lazy and sloppy and feels like the writer pulling out the rug from under the reader in an intentional "hah, hah, fooled you, idiot!"-middle finger gesture instead of proper plotting and writing.
What I have no problem with and even like is the A Song of Ice and Fire type thing where it's made clear that this is the type of book where People Die (or even Anyone Can Die). The book sets proper expectations and then the "will they survive or not?" question becomes a part of the enjoyment when reading.
I've been trying the free tier Sonnet 5 for some LLM prompt template writing / storyboarding help and it's been surprisingly helpful (A bit like an improved pre-enshittification-era Google but nothing groundbreaking - I still need to hold its hand constantly and hate that I can't remove individual messages from the current chat). I found this anecdote from Reddit quite amusing given I haven't noticed anything whatsoever like that.
Acts very rude, assertive and feels like its dragging its feet when trying to do any task.
It once said I was just "rambling" when I was asking it about foreign legislations, and complained that I wasn't working on my main project, for example.
As well as the next comment in the same thread
sonnet 5 is way to much like chatgpt 5.2, it's the first Claude I've used that is unfriendly.
Perhaps my expectations for an LLM are just calibrated to be properly low or the fact that I told it to "avoid em-dashes" silently fixes that "unfriendliness". If anything, I have to be careful against its fairly regular glazing, lest I start to think my ideas are too good.
During the same week I also grabbed the Chatgpt offer of Plus subscription for $0 for the first month (cancel any time), but I find I just can't tolerate its "words words words more words a bit of actual information fucking more words words words"-tendency.
I was expecting a bit different kind of Mr Fox.
y = mx + b.
This is heresy! It's obviously y = ax + b.
What did the Mr Fox say?
That's a completely different claim. Is it the drag queens who are the initiators or is it the activists around them, the ones arranging the event and the library staff?
There's a big difference between
"Hey, you really should come to read in this event. I've reserved the space, done the arrangements and convinced the staff." "Oh, ok, I guess I'll come, it might be fun."
vs
"Hey, I really want to do this thing. Is there any way I can make it happen?"
If the drag queens would keep it there, there'd be no problem (and most of us wouldn't even know that it exists). But they want to move it into public spaces
Do they?
Are drag queens themselves pushing to be in public places or some other adjacent but still different set of people who are doing that?
TBF, it's unfortunate that crusade failed. Not because of the lyrical content but because of the unmusicality of 99.99% of such music.
Instead, they do a ridiculously exagerated caricature of women.
How is that any different from a significant amount of all theatre performances, other than concentrating on a specific large and unhomogenous group?
Agreed. I see data centers more as something between commercial and very light industrial sites, something like warehouses but with more electricity / telecoms / capital investment need.

It can be surprisingly hard to force a smile when 1) you can't see yourself and 2) you need to look towards a specific unnatural spot.
More options
Context Copy link