2,957 karma · joined September 15, 2020
It's been a moving target for years at this point.
Both open and closed source models have been getting better, but not sure if the open source models have really been closing the gap since DeepSeek R1.
But yes: If the top closed source models were to stop getting better today, it wouldn't take long for open source to catch up.
[0]: https://en.wikipedia.org/wiki/Shoshin [1]: https://en.wikipedia.org/wiki/No-mind
This is a total misrepresentation of how any modern LLM works, and your argument largely hinges upon this definition.
The model of Claude the DoD is asking for more than likely doesn't even exist in a production ready form. The post-training would have to be completely different for the model the DoD is asking for.
So... all acts are selfish because if it looks unselfish, that just means it was selfish in a hidden way?
You mean at the top labs? Since when isn't that level of misanthropy categorized as having mental health issues?
They inserted themselves into the supply chain, and then the government told them that they'll be classified as a supply chain risk unless they get unfettered access to the tech. They knew what they were getting into, but didn't want the competitors to get their slice of the pie.
The government didn't pursue them, Anthropic actively pursued government and defense work.
Talk about selling out. Dario's starting to feel more and more like a swindler, by the day.
Care to elaborate? Paid by whom?
Why isn't the person posting the full transcript of the session(s)? How many messages did he send? What were the messages that weren't short?
Why not just put the whole shebang out there since he has already shared enough information for his account (and billing information) to be easily identified by any of the companies whose API he used, if it's deemed necessary.
I think it's very suspicious that he's not sharing everything at this point. Why not, if he wasn't actually pushing for it to act maliciously?
If you're driving a vehicle in the former category, you'll be on the hook for reckless driving if you aren't fully supervising the vehicle.
I'm pretty sure the original commenter was supervising the driving, though.
And the play style of Alpha Zero wasn't different in a way that needs a super trained chess intuition to see, it's outrageously different if you take a look at the games.
I guess my point is, that even if the current situation is basically a 'deadlock', it's been proven that it's not some sort of eternal knowledge of the game as of yet. There's still the possiblity that a new type of approach could blow the current top engines out of the water, with a completely different take on the game.
In that hypothetical of running 2 instances of Stockfish against one another on a modern laptop, with the key difference being minutes of compute time, it'd probably be very close to 100% of draws. Depending on how many games you run. So, if you run a million games, there's probably some outliers. If you run a hundred, maybe not.
When it comes to actually solved positions, the 7-piece tables take around 1TB of RAM to even run. These tablebases are used by Stockfish when you actually want to run it at peak strength. [3]
[0]: https://tcec-chess.com [1]: https://lichess.org/broadcast/tcec-s28-leagues--superfinal/m... [2]: https://lczero.org [3]: https://github.com/syzygy1/tb
Chess engines have been impossible for humans to beat for well over a decade.
But a position in chess being solved is a specific thing, which is still very far from having happened for the starting position. Chess has been solved up to 7 pieces. Solving basically amounts to some absolutely massive tables that have every variation accounted for, so that you know whether a given position will end in a draw, black win or white win. (https://syzygy-tables.info)
Not sure if you mean that I'm being hostile for no reason towards this 'someone else'. The second section in my original post is the big conditional.
> I guess the grifter didn't just copy paste what our guy did. it did have more impact and well time
This is more than likely correct.
Either way, I do think it's grifter behavior to not mention/include anyone else who was involved in the project if you pick it up halfway through. Unless the code (or whatever else) is actually bad, and you have to do extra work to redo it. And, if you are actually aware who even worked on the project to begin with.
But it very well might've been a case where some higher up passed the project off to another programmer (months/years later) with no malicious intent whatsoever, and the programmer just did the thing as requested. Or a myriad of other explanations.
If it actually went down like this, that's pretty horrible, and that someone else is a grifter. Very harmful for any organization in the long run, because that behavior will be applied to anyone who's "ripe to be taken advantage of" (from his point of view), burning them out of the way.
That is, if they were aware that you made the thing that they picked up later. Though I wonder why the original didn't go through. The other person pushed harder for it to go through, or showed it off with a different sort of demo? Or was it a different sort of technical implementation / design?
Assuming we're just talking about information on the internet: What are you reading if the original source is several dozen layers deep? In my experience, it's usually one or two layers deep. If it's more, that's a huge red flag.
Look at the actual bot's GitHub commits. It's just a bunch of blog posts that read like an edgy high schooler's musings on exclusion. After one tutorial level commit didn't go through.
This whole thing is theater, and I don't know why people are engaging with it as if it was anything else.
This is more a case of GitHub as an organization actively embracing having agentic AI rummaging about.
This is completely backwards. The Beatles put out songs that they didn't think were hits, and put out songs that they were conscious of being the antithesis of a hit. They wanted to freak people out from time to time. As many artists do.
Just check out Revolution 9. Pretty sure you can't get much out there than that when it comes to music of that era. And still very out there to this day.
Or for a more 'songy songs' that I'm pretty sure they didn't think at all in terms of hit material: Tomorrow Never Knows or Within You Without You. And there's dozens more.
It's really easy to end up on autopilot with the structure of your workdays, only to notice that it doesn't work for the current project all that well. Getting stuck on the 'nothing' approach can be just as bad as getting stuck on a 'hundred tabs' approach if you're just doing one or the other because you've always been doing it.
SHOW ME THE MONEY!!!