I prefer to have a Clippy agent that identifies code that should be copied and pasted into random locations for no good reason. You may think it's not necessary given how LLMs operate, but I have found it's critical for 10x gains.
Well I would hope so. That's the most basic "translation" a computer can do. Converting one set of characters into another set of characters from a lookup table. How many r's in strawberry is a more difficult question.
Is this some unique baseXX encoding? Please educate me. You do know that just because bitcoin uses it doesn't make it special, and doesn't stop it from being the second most basic translation task? The most basic being something like rot13 (direct character substitution). Are you saying this is somehow special because extra characters are added at the end to make the "grammar" work?
Sigh. Read your own link? Yes it’s not the same baseXX encoding with padding as any other.
It makes it not a straight 1:1 map unlike b64, and you need to be doing actual compute in order to get to the result. You are operating on blocks and calculating checks on each, etc. The algorithm isn’t complicated by any means but it’s certainly not the “second most basic translation task”.
And the CryptoNote variant is further different as it operates on 8 byte blocks. So you linked the wrong documentation.
Sigh. Apparently you didn't understand what I meant by second most basic. Basic + an extra step is still pretty damn basic algorithmically.
You are correct in that I didn't catch the "CryptoNote" variant. But a different block size? Come on. That's not even a material difference. This shit still isn't magic.
But you're right that it is more on the order of "how many r's in strawberry" difficult than it being less difficult.
There's still not some big magic difference between modification per block and modification at the end. It's still right up the alley of a translation machine.
I stand by it being the second most difficult translation task.
And I stand by the original post being goofy hype.
Whatever dude. If you can't see the difference between doing in-thinking freetext arithmetics and counting characters in a word, I have no hope that talking to you will be useful.
GP's phrasing was not precise, but I think most people understood it was liability ownership and not copyright. Ownership with respect to copyright and ownership with respect to legal liability are two completely different concepts.
That's correct. What I meant was that if my bot writes defective code for whatever reason, I am responsible for that (though conversely, if I get more done because of the bot, anthropic gets the praise).
Parent of your response: > ...may be hard to identify which company or who’s responsible when you fired all of the people from the department that it’s supposed to police those type of things.
You responded to a post quite literally about identifying sources, which is exactly what a response team does. Maybe you respondrd to the wrong post?
It's good that you clearly see the "Great Idea" spam for what it is. That comes from the reinforcement-learning human-feedback (RLHF) that is used to keep you coming back for one more pull.
I thought this would be a push for a nationwide open weights effort / initiative, but I was surprised to see this:
> Distillation ... reflects a long tradition of learning from, building upon, and improving existing technologies, a tradition that has helped drive innovation since the rise of the open-source software movement. By contrast, unlawful efforts to extract value from closed models raise legitimate concerns. Those concerns should be addressed through targeted legal and commercial frameworks rather than sweeping restrictions on techniques that play an important role in AI innovation.
Sounds like they are saying "please protect our IP theft" that created closed weight frontier models in case we arbitrarily decide to close our models. But don't get rid of distillations in general so that we can all also keep benefiting from open models. We don't want to lose the ability to benefit from the work of others as we launder IP into closed models.
Surprised Linux Foundation kept their name on it with that.
"..a tradition that has helped drive innovation since the rise of the open-source software"
i've said this in other comments but the fact that these companies are trying to align with open-source when there's no "source" included with their models is pretty damning. I think it lays bare the absence of any kind of noble or righteous motive with respect to distillation.
The entire reason we are in this war is the current administration's ineptitude. Anyone who voted for this president should be ashamed they fell for a con.
Here's my hot take. Bush, Clinton, Obama, were all liars. Trump is actually the most honest president I experienced. Partly cause he is so bad at lying, but also cause he wants people to like him. People knew what they were voting for, he did not hide his idiocy.
We won't progress if we don't fix the mistakes that led to where we are. It's not just "Trump bad".
Dude straight up lied to everyone. Most blatantly corrupt incompetent fraud in US history. Just because he's narcissist enough to not care doesn't make it better. And yeah, lots of mistakes to fix that led to McConnell / Dumpty. Mainly erosions of separation of powers. At least the other presidents you mentioned tried to stick to what they campaigned on and didn't try to light fire to the government for personal gain.
I'm not sure "honest" is the right word. Trump loves to lie and make stuff up. It's just usually blatantly obvious, while Trump is also seemingly unable to keep anything secret. Maybe "transparent" is the right word. Just not on purpose
100% this. OpenAI created a for-profit subsidy (that is now an SBC and only controlled 26% by the original non-profit). Guess when they did that? After GPT-2 (Feb 2019) and before GPT-3 (June 2020). Not remotely surprising they started the move to for-profit as soon as they had something they thought could be profitable. Then in 2025 they switched to an SBC, dropped the non-profit control well below 50% and completely got rid of the profit cap. Just in time for the IPO talk. It was a mask the whole time.
reply