strangescript

Shouldn't they be loaded right now with all the BTC they bought when it was low?

Why do articles testing ancient LLMs keep getting posted?

* Yes I am aware I am not running R1, and I am running a distilled version of it.

If you have experience with tiny ~1B param models, its still head and shoulders above anything that has come before. IMO there have not been any other quantized/distilled/etc models as good at this size. It would not exist without the original R1 model work.

Everyone is trying to say its better than the biggest closed models. It feels like it has parity, but its not the clear winner.

But, its free and open and the quant models are insane. My anecdotal test is running models on a 2012 mac book pro using CPU inference and a tiny amount of RAM.

The 1.5B model is still snappy, and answered the strawberry question on the first try with some minor prompt engineering (telling it to count out each letter).

This would have been unthinkable last year. Truly a watershed moment.

These presentations often serve as a comfort blanket rather than a plan of action. Oh man something incredibly disruptive is happening to us. Lets talk about it. Whew, okay, we understand it, lets go back to being complacent.

Years later, "man we tried, we had that meeting and everything, we just couldn't compete"

There are tests they are passing that they can't be hardcoded for by design. They still have all kinds of flaws and inconsistency but getting upset they answer "2+2=4" because someone trained them on what the answer to 2+2 is supposed to be is silly.

And two years ago this result would have been thought impossible.

Picks up goalpost, looks for stadium exit

"We have created artificial super intelligence, it has solved physics!"

"Well, yeah, but its kind of expensive" -- this guy

The dems started this when they black balled Bernie in 2016. They were too focused on their own self interest and they are going to reap what they sowed long into the future now.

As soon as something can be automated, people start acting irrationally upset, especially if the thing is seen as even remotely "creative". Those people are going to have a bad time moving forward.

I find it a little confusing as well. I wonder if its because so many of these companies have went all in on the "traditional" approach that deviating now seems like a big shift?

This is all semantic bullshit. I could raise a child devoid of any outside contact and teach them that klingons enslave our people and we must hide in the woods. Things you see in the skies are their hunting machines. They could completely "know that" and its a total fabrication.

What if that person firmly believes what they say is true?

Guessing you used 4o and not 4o-mini. For stuff like this you are better off letting it use mini which is practically free, and then have it double and triple check everything.

This. Its what people want to hear. If you heard that voice in a vacuum with no context and asked someone what famous person is it, I doubt many people would say Johansson. Some, sure, but not the majority.

The only thing Sam did wrong was play too fast and loose with the implication of "Her" given that he had been talking to ScarJo. Lawyers should have told him to just STFU about it.

"Is this thing we defined to make us feel special achievable?" What we define as "consciousness" is just an emergent property of large, dense neural networks w/memory.

The human brain is a marvel of mechanical and electrical engineering, not a mystical, otherworldly device. Maybe there is something going on at a quantum level to explain the incredible performance our brains achieve, but that is still nothing that we can't build.

This would be interesting but also feels a little restrictive. Maybe something like LoRa could bridge the capability gap but if a competitor then drops a much more capable model then you either have to ignore it or bring it into your app.

(assuming companies won't easily share all their models for this kind of effort)

Or all these properties are emergent from ever increasing compute. The underlying architecture doesn't need to fundamentally change. The roadmap is increasing compute. Agency is simply never ceasing input and resulting output, both internally and externally. Do you have Agency when you sleep? I still don't know what "extrapolate beyond their experience" means. Does Einstein discover relativity if he is never taught math? More to the point, why was he so brilliant in the first place. Why could he "extrapolate beyond their experience" better than others in his field?

We want AI to hoop jump something we don't even understand ourselves. The only empirical evidence we have is as we increase compute, the results get better.

The fundamental problem with this strategy is model size. I want all my apps to be privacy first with local models, but there is no way they can share models in any kind of coherent way. Especially when good apps are going to fine tune their models. Every app is going to be 3GB+

"extrapolate from our experience" "synthesize something coherent"

These are non-scientific concepts. You are basically saying "humans are doing something more, but we can't really explain it".

That assumption is getting weaker by the day. Our entire existence is a single, linear, time sequence data set. Am I "extrapolating from my experience" when I decide to scratch my head? No, I got a sequential data point of an "itch" and my reward programming has learned to output "scratch".

Or, you prices are outrageous. I have plenty of money to blow on Mceedees, but its absolutely stupid how expensive it has gotten.

This a pointless nearly unenforceable rule to make people feel better. Sure, if you generate something that seems like a real event that is provably false you can be caught, but anything mundane is not enforceable. Once models reach something like Sora 1.5 level of ability, we are kind of doomed on knowing whats real in video.

I am sure they got the heads up about NVIDIA's announcement and cut this contractor loose. As per the contract agreement, the project would become open source.

This is definitely a problem, but you could also ask this question to random adults on the street who are high functioning, job holding, and contributing to society and they would get it wrong as well.

That is not to say this is fine, but more that we tend to get hung up on what these models do wrong rather than all the amazing stuff they do correctly.

This feels a little click baity. I am sure there are tons of exploity software repos on github as well. Need to use some common sense and protect yourself and report things that seem off.

I guess it might be surprising to some that code from HF can have exploits just like GH.

I might be mistaken, but can't you still get models from HF that don't use safetensors?

Meanwhile, I am going to take my horse and buggy down to the local blacksmith to get some work done...

I like what you built. I think calling something a palette generator that only does one color at a time isn't "correct". Or at least not what I am looking for. I want something like yours that builds out a suite of colors for an app.

There is a difference between being buggy and being poorly designed. This is something Bethesda used to understand.