Ok, I’ll bite. Assuming that the article means Large Language Models exclusively when mentioning “AI”.

Good specialized models aren’t that easy to build for a lot of use cases.
Coding is one of the more problematic cases. People tend to think that coding is about code, so if we just create a model that is specialized in writing code, we’re golden. However it turns out this isn’t true. Coding is not only about code itself, but about creating an application that has a specific purpose and features in a specific domain. The LLM needs an understanding of not only the plain code, but needs knowledge about the subjects the application is about, and these are mostly in the domain of general knowledge. Another issue is that the model itself needs to communicate with the user in an appropriate way, and this needs general knowledge as well.

“Summarize documents, draft memos, analyze spreadsheets”

All these examples require an LLM to have a good general knowledge to grasp what the original text is about. You won’t get a decent summary when the general knowledge of the subject at hand is subpar. My point is, bigger LLMs with more general knowledge handle these tasks better than smaller ones. I’m not saying smaller models can’t do the job, but there will be differences in quality.

It’s not that hard to let existing LLMs use specialized knowledge.
The actual problem is to prevent this specialized knowledge from being stolen by US/Chinese AI companies. So instead of focusing on building own “specialised” models, let’s work on preventing the ongoing theft of knowledge, and using the right tools for the job.

Olaf Scholz? Is he still relevant? I think you might be confusing the former German Chancellor with the current one, Friedrich Merz. Who actually said “that Google is also investing in Germany and Europe, and that we can ensure European sovereignty here even with an American provider.” - which is utterly ridiculous.

The article mentions that the state is currently home to 257 data centers. Just one thing to note in case you aren’t aware: That’s not exclusively AI datacenters and there is a massive difference in power draw and pollution between regular data centers and the AI counterpart.

I never wanted to be anything, and I was exceptionally successful at that so far. :/

I guess it would be good if a developer would disclose what AI was used for. Can’t make everyone happy here, because we have all nuances from “any AI use is bad” to “IDGAF” on Lemmy, so the dev should give more details on actual AI use. I personally would not object to using AI helpers for boilerplate stuff like this, but I can guarantee that a lot of people here think differently about it

Should have used IMAP, amirite?

Guess a real bank employee would know that withdraw used as a noun is called a withdrawal, but fun story.

I still can’t wrap my head around how this strategy (“buy our models, they are so powerful they might kill all humans”) would work in their favor, apart from one thing: restrictions would hurt Chinese models probably more than themselves. But I find it very unlikely that characters like Altman or Amodei would be capable of playing that kind of 4d chess. It’s a weird flex

What do you think where this is going to end up? My personal assessment of the situation is that only the mega-corporations can afford enshittification like that, and even those are now getting a taste of their own medicine (Sloperators recreating their flagship products… it’s a race to the bottom with enshittification on one side and sloppy replacements on the other). Smaller sized companies will simply get buried when the quality of their products falls to garbage levels. This is what those smaller sized companies don’t realise yet: When everyone can churn out the same low quality slop, their USP and their customers are gone soon.

Let them, but insist they should also use Claude to QA their code and thus take responsibility for the whole thing. If they trust AI to be good enough, why not go full circle? That shouldn’t be a problem, right?

Joking aside, I know there is a real danger of those people shifting the responsibility for their created code mess downwards to the developers and QA, and when shit inevitably hits the fan they know who to blame.

While that is partly true, it oversimplifies a few things. And these are the things that make LLMs a dead end for intelligence:

• Evolution means constant learning and adaption. LLMs can’t be adapted on the go, they are stuck with their initial training data and retraining is a slow and externally controlled process. Apart from training, every additional information you give it while running must be loaded into the context, the space for context is limited and must be flushed before supplying different context. That’s a hard limit for evolution, at least with current technology.
• The physical world is more important than you might think because everything hinges on this realm, including the electricity, cabling and hardware involved in running a model. The idea of an intelligent entity that exists purely as an ethereal cloud of logic unbound by physics is hard to justify, because the shape of logic itself is defined by the physical ground we are standing on. Therefore, an AI would need a first-hand access and understanding of the physical world to be successful. LLMs can’t have any of that, because they are purely language-bound.

I personally think Yann LeCun (a pioneer of machine learning and AI) is on the right track with his research and his opinion that LLMs are a dead end for artificial intelligence. The reasons are more on a technical than a philosophical side.

Ghost pairing ist ein irreführender Begriff. “Digital incompetence” trifft es eher. Das potenzielle Opfer muss gleich mehrere dumme Dinge tun:

• die eigene Telefonnummer in eine nicht vertrauenswürdige Website eingeben
• einem Warnhinweis ignorieren und einfach auf “Ok” drücken ohne zu lesen oder zu verstehen was dort angezeigt wird

Aber tatsächlich wäre es vermutlich sicherer wenn der Nutzer im Messenger initial selbst seine Paarungswilligkeit anzeigen müsste bevor eine solche Paarungsanfrage überhaupt möglich wird. So wie bei Bluetooth, wo die Auffindbarkeit explizit eingeschaltet werden muss und nur ein paar Minuten aktiv bleibt.

It’s kind of scary to imagine living in that time and using all these products (or even worse, being a worker in an asbestos factory) cause you have no way to know this stuff causes cancer. I wonder what kind of extremely unhealthy things we are exposed to right now just because we don’t know.

Not the OP, but any software developer who has spent some time with the inner workings of LLMs can tell you that. Open weight models are a good opportunity to explore how LLMs work on the inside and to experiment with them a lot more than you can do with the so called Frontier models. The technology however is the same. These are just very fancy statistical prediction machines, and the reason why you are so impressed with them is that the statistical average of what sounds reasonable in a conversation is exactly what people expect to hear on average, given a specific context. It’s a neat little trick that works only only as long as the context fits (contentwise and spacewise), and once you realize that you’ll see LLMs as what they truly are. There is no magic, and the “intelligence” is the intelligence of the people who built those systems, the systems themselves are not intelligent in any commonly used sense of the word.

It just proves that people disagree with you, for whatever reason. Could be that they don’t like to be called stupid. Who knows.

Sorry bout the sarcasm. My point is, not only do you have to specifically instruct the LLM to go looking for voter data, but you’ll also have to let it roam freely and unsupervised on the internet for quite a while to get to such a result (which is a bad idea to begin with). In the end, it’s the sloperators fault to let it go off the rails and a scenario like this is entirely preventable

Thanks, I hate it. Not the browser, but the ad. I don’t want stale nostalgia. I want a modern, secure, privacy-oriented ecosystem and tools which help taking the internet back from the claws of megacorporations. Tools which help building new communities and resilience to corporate enshittification. The ad just rubs me entirely the wrong way.

So, there I was, happily vibecoding another Minecraft clone when suddenly it went looking for voter data in Georgia. Yeah, happens every time. Can’t do anything about it. These Superintelligences just have it their way.

4 of the current “Top 10 movies” in german Netflix are “Ice Age”. Tells you everything you need to know about the quality. There is simply nothing left to watch.

I don’t know and in my opinion letting children use LLMs unsupervised is a bad idea regardless of which provider is used. But as long as teachers don’t understand the basics of how LLMs works and treat “AI” as a magical answer box there is not much hope left anyway. Sorry if I’m bitter here. Teachers are expected to know about the materials and technology they are using to teach children. The blatant lack of knowledge and ignorance of the teachers are what makes school AI usage dangerous for children.