No they hallucinate because they lose context but are tasked to answer anyways.
Nothing to do with context. Context windows are an LLM specific thing, but all neural networks no matter the kind or size “hallucinate” because they are a piece of linear algebra where you put numbers in on one side and get numbers out the other.
If you can’t see the difference between a compiler and a large language model
If you can’t see the difference between a given example and the underlying logic…
And if the entirety of your argument is a mystic “and all that stuff”
What, you want me to list the entirety of sciences downstream from physics that are involved in generating and predicting movement in mammals? Because that could, like, take a while…
That’s… the joke. The original proposition was that AI code bans can’t be enforced, a thought which I was leading ad absurdum.
I hate this attitude that if a rule or standard can’t be enforced perfectly, we shouldn’t strive for it.
“Murder bans may be impossible to enforce in the first place”
What do you mean? That’s exactly how LLMs work. That’s why they “hallucinate”.
I’m just a movement generator.
Which is way more complex than just predicting text, because you have to predict physics and all that stuff. There is a reason most robots controlled by neural nets have failed hilariously so far.
then you’ll have to acknowledge that some form of reasoning is going on, no?
By that logic you could say a compiler is reasoning. But the reasoning displayed was happening when the compiler (or LLM training material respectively) was written.
Can a lemmy post be directly mastodon
Kind off no and yes. Server side there is no federation AFAIK, but some frontends/apps can show and post to both networks. Kbin and forks for example.
Europe didn’t really “ditch Windows” though or did it. You see news of some places switching to Linux or planning to switch
Yes, that is correct. The nature of the EU itself and many of its members individually is rather decentralised, so you will see it happen bit by bit. But there is an increase in national governments interest in doing this over the last decade or so, and them doing it will make it easier for businesses and state/regional governments to follow. A keyword to look up is “Digital sovereignty”, that’s what they usually call it.
Nah the original clearly says “Leserkommentare”, so it’s user comments.
Well the script tag has a deprecated attribute called language, so in older code and media you will see <script language="javascript"> a bunch. The weird thing is they instead used the lang attribute which just doesn’t exist.
Oh I see. You mean like v3.0.1 will be stable, and v3.1.0 will be stable. Yeah that’s true. But v3 (as in v3.0.0) is breaking. At least that’s how I have always seen the terms used.
No, v3 is a release with breaking changes:
This release includes several breaking changes; read the full migration guide here. It’s worth mentioning that many of the breaking changes are updates to API endpoints and only affect third-party tools that integrate with Immich’s API. For the vast majority of users, updating works exactly as it always has.
https://immich.app/blog/v3.0.0-release
That’s why the major version was increased, it implies breaking changes in semantic versioning, which Immich is doing since v2 apparently:
Given a version number MAJOR.MINOR.PATCH, increment the:
- MAJOR version when you make incompatible API changes
- MINOR version when you add functionality in a backward compatible manner
- PATCH version when you make backward compatible bug fixes
Well it for sure killed my entire “Suspension of disbelief”.
Well if you want the real basics of self-hosting then depending on your level of knowledge you may want to follow a Linux tutorial like this one: https://labex.io/linuxjourney
This will teach a lot of basic knowledge and terminology you may need when setting up or maintaining containers and VMs. Admittedly, this is somewhat tangential knowledge to your immediate goals, but if you want to actually use your self-hosted stuff remotely later this will become critical to doing so safely.
As for the map, for me it was something like:
Local LXC/QEMU → Remote LXD over SSH on a RPi → Proxmox on a dedicated box through the webinterface
In hindsight I would have started with Proxmox directly though I guess. One big downside is it doesn’t integrate well with Docker containers, you have to set up a full VM as your Docker host. On the upside though you can install LXC containers from https://www.turnkeylinux.org/ in a few clicks, so it’s very good for just testing stuff out and playing around.
Well I don’t know. Their response overall is quite incoherent in places, but they mentioned somebody, presumably German, interfering twice now:
we didn’t port the feedback from our German QA to other languages
https://x.com/GOGcom/status/2062981104362242551
This was noticed before distribution, and out of respect for local sensitivities, the material was not sent to the German community.
https://old.reddit.com/r/gog/comments/1txmcyd/ts_gotta_be_racist_gog_explain/opxtour/
And yeah I can absolutely see a German translator/QA person just going, “No, I’m not sending out a fucking SS rune. That’s a fucking crime.” Because apparently it was only censored in the German language version of the newsletter, Germans with language set to English still received it with the runes. So that part, the “hey the German censor was the only one who caught this”, that I can believe. And some PR person rephrasing this embarrassing state of affairs as “German QA” and “respect for local sensitivities” that I can believe too.
Beyond that, there is pretty deep lore on this already. The developer used an Indian sun wheel before being made aware of possible “connotations”, and then apparently misinterpreted the term connotations, see here (same reddit thread as above, but different comment). Then the GoG rep saying in the reddit thread above the S-like runes were supposed to be Sowilō runes, which are indeed sun symbols, although not Slavic ones, which may comes in pairs, but not in the same orientation, and seem to render different depending on fonts. But the email seems to have used Greek Kappa, which may or may not render differently, looks like it doesn’t. It’s confusing, but plausible enough to have a difficult time disbelieving it outright.
But on the whole, those three/four symbols just don’t come together by accident. Some person put those there on purpose and is trying to play it off as a mistake. I suspect the author of that first damage control message on reddit.
is it cool if I skip ahead to the next exit?
No, you will be fined upwards of 240€, lose your license for a month, and get 2 points on your license (8 points means you lose it permanently, you will have to redo the license and take a so-called “idiot-test” beforehand).
Man fuck Axel Voss! Damn copyright shill. Guess we can take solace in the fact that he seemed to be the only one clearly taking the publishers side here.
And if I’m not mistaken, the European Commission representative argued in his reply to Voss (around 12:20) that “collective management organisations” or “cultural heritage institutions” might well be allowed to preserve games that are not commercially available anymore already under the current framework.
Well yeah kind of, except for the caveat that there are some things in those tests a website can’t really check. The tests are standardised and publicly available. There isn’t really any reason to not implement a (half-way) proper test, even if you are just harvesting data.
Above average that has to count for something right?
Well they aren’t though, 345/1000 people having a worse score than you means there are 654/1000 people who have a better score.
“IQ 100” is a moving target that gets regularly updated to the new average (which relatively speaking has been mostly climbing for decades).
Might well be behind here, but to my knowledge we don’t have a single robot outpacing a single human in most tasks. They don’t do one-shot learning from their mistakes, they don’t learn new movements randomly. Because most basically just use static weights when operating, because you don’t really want an industrial robot to get ideas. But even test systems that can learn usually just have an equivalent to the movement model of a brain, maybe a vision model, just like an LLM is just a language model. You would need a system that has all of those, plus the other brain areas, especially a prefrontal cortex like model for integration.
I’m sure people are working on it, but I haven’t heard of anything successful yet. I imagine there might be a being like that in secret which is currently tortured in some billionaires tech dungeon. Poor thing.
Yeah but that’s coincidental. It’s the model weights, prompt, and the RNG aligning. They can mock reasoning, because they do it by what they always do, predict more text, but it’s not like this has any effect on themselves. They aren’t really understanding a mistake when you point it out and growing neurons and synapses, i.e. they won’t have different model weights the next time you ask the same question. They can only really change when the powers that be release an update, which includes new training data, and hence model weights.