Well that and the fact that they have a company to run and people to pay. Do you seriously believe inference is not profitable, despite evidence to the contrary being everywhere you look?

They are profitable on inference (at least anthropic). Inference is cheap, it’s not susbsidized. It’s training that’s expensive, but if the bubble really does pop then current open source models (which are pretty much sold at inference rate + small margin) will actually go down in price since the cost of hardware will have gone down (assuming less demand for GPUs). Considering that the current best open model is roughly one month behind, we won’t revert to pre-ai anywhere. It’s just too cheap for that.

Still, it’s lovely seeing people spreading misinformation based on something they saw on tiktok years ago. Keep up with the times or don’t spread outdated info.

Why does it not? According to the english definition of the word it literally does. You people don’t care about facts anymore, it’s all about feelings and what you’d like to believe.

I’m not reacting to a meme, I’m reacting to someone spreading lies and everyone swallowing it whole, despite proof to the contrary being publically available. When it’s the other side doing it it’s propaganda, but when it’s yours it’s just a meme?

Just proves my point. You don’t care about the truth behind the tweet, only about how it was twisted to your apparent benefit.

By definition, a “rogue” agent is an agent that behaves in unexpected ways, often damaging. Tell me again how there are no rogue agents?

If you’d used modern agents you’d know they do unexpected, sometimes damaging things on the regular, including the cheap models.

“I don’t understand it, so it must be a lie.”

Go and read the article. Nothing they explain is impossible, you can even reproduce it yourself.

Is this what lemmy has become? Statements like this getting upvoted by hundreds of people is crazy. Nobody here read that article but will happily follow the opinion that’s most cozy to them.

I wasn’t talking about the paper, when I talked about the paper you ignored everything I said and moved in a different direction, which was what I responded to. If you’re wondering about the relevance, perhaps you shouldn’t have brought it up.

Does it matter whether it’s part of it or done immediately after? For all intents and purposes it’s the same thing. Like I said, if the input and outputs are the same, what does it matter how the process works?

From the user’s perspective, where one question results in many inference calls, it would look like the LLM learns while it works, assuming such training would be enabled, which they obviously wouldn’t but could do.

Then you know that there is nothing stopping you from training during inference, or to modify weights in reponse to inference. It just so happens to be more efficient to train the next model instead, but again that goes back to what I said, there is no need to copy humans in this because humans aren’t efficient in everything, and certainly not in training.

Have you done any ML? Because if you had, you’d know that statement is complete bullshit.

You can interrogate the model too, and it’ll answer according to the context. Why are you assuming that the brain doesn’t do tricks but everything AI does is? We too have memories and since we can’t store them all, compression is performed (supposedly while sleeping), based on a few algorithms, some we understand more than others.

We, in creating LLMs, have copied the human brain as much as we can, hence why the core function is so similar. True, LLMs use a lot of “tricks” like you said, but if the same input produces the same output, who cares if the algorithm is different?

For example, take episodic memory. In humans, it’s a type of memory that can be conjured at will. Typically not acquired during training (or, in humans, generic evolution) but rather through experience of the individual. LLMs simulate this by taking notes, then later on running a specific command to recall the specific part of those notes that matter for their current task. From an outsider’s perspective, it’s even more accurate than episodic memory in humans since it can grow almost infinitely and suffers no loss of details. Yet it is a trick, therefore it must be bad.

Let me remind you, evolution took place over 4 billions years, modern LLMs has pretty much been around for two. Evolution never cared to make us good, it just wanted us to reproduce. We know that it’s possible for humans to produce fresh cells, we do it when we have kids. Why then can’t we use those fresh cells for ourselves to be effectively immortal? Evolution doesn’t care. Evolution as a process is flawed, and made humans flawed too.

Why then is it that when we change anything in the way flawed humans think, it’s seen as bad, even if by all metrics it’s not?

But we too are trained to say we’re in pain, trained by millennia of evolution. When you touch a hot stove and cry out, that’s just built in your training data.

How do I know anyone is ever in pain? It’s also just words, perhaps some moans too.

Just like this “bubble” has supposedly been about to pop for years.

They absolutely are, but most people on this platform would prefer to hear their comfortable lie instead.

How is it not better if it can do things humans can’t? OpenAI just solved another 100 problems no human could solve. Mythos finds decades-old vulnerabilities in common software on the regular. What can humans do better? You don’t have to be an expert to see where this is going. The entire field of math is basically in a panic right now, signing petitions left and right to try and stop AI from killing their field. Cybersecurity only fares better because not every company is allowed to use mythos and a big part of the job is security training for employees.

Because right now LLMs (and potentially other forms of AI) are the only technology capable of doing it. Humans can too, but are much slower and more expensive.

Well the idea is you don’t need to code it yourself, you can have it do it for you. Sure it’ll have bugs the first few times, but humans also make mistakes the until they get the hang of it.

Yes, and that’s called a harness, which everyone uses these days. The harness increases perceived intelligence (or accuracy) by absurd amounts. You can “teach” (or the equivalent of) LLMs anything with a custom harness.