BoorishBears
I think the presentation of the points is confusing people, the model could understand weights directly for extremely small networks in the 10s of thousands of weights.
Once they moved onto even the smallest LLMs, what it did was closer to abliteration: create examples where the model does something you don't want, create examples where the model it doesn't do the thing, then most importantly do forward passes.
Once you're allowed to actually do a forward pass vs just reading raw weights, tracking the changes in activations and tweak the model to modify them is a very well established process that takes a few hundred lines of Python.
How much easier my life would be if Chinese open models were even just 1 year behind outside of X posts and one-shots.
My easy litmus test is would the year old model (Opus 4.5) being open sourced today would eat the traffic of every single Chinese model... and the answer is pretty much yes.
Every Chinese model would throw away their work just to post-train or distill Opus 4.5 and get a more reliable model.
So diverting significant resources is just going to harm the main goal and kind of suck at the side goal too.
4.5 is a great example of this: it wasn't perfect but it was better than every 5 family model at things like tone and writing, and despite being priced extremely ruthlessly and already finished training, they couldn't afford to tie up the resources to host it.
If you say the quiet part out loud, the lemmings will take that as approval.
"Only thoughtful application of AI allowed" -> "Of course my usage of AI was thoughtful! My custom pi harness with 1M context project brain is very thoughtful!"
Not sure why you think this wouldn't be something they did on principle either though. Most open source projects are built on someone's personal principles: that's what motivates them to keep working.
(hence the recommendation someone's free to fork it and do the work themselves)
My read is any LLM use should be opaque to them. You should be able to explain your bug find in your own words, have formatted as the project expects, not have any LLM generated comments (which trend towards being useless anyways)
The wording of the post is very clever, if you're green you read it and you say "I'm banned because I use Claude". If you're experienced enough you can read the message between the lines: "if you use LLMs, it should still have gone through your brain before it reaches mine, so I'm not turned into a meat puppet by proxy"
It's sad to see Carbon releasing at the same pricing because that implies Google is planning to do the thing everyone is doing: release promising model for subjective preference, RL-maxx a V2, replace promising V1 with V2 because it benchmarks better
Muse did it, Kimi did it, Deepseek did it, Tencent did it. No one seems to care abojt subjective preference anymore.
If they couldn't train a reliable detector with that they wouldn't exist.
You wouldn't even need to be able to read the information the watermark encodes to recognize it exists in a piece of text
At this point most people getting a 3DS are probably doing so specifically to jailbreak it
My EVGA came with stock firmware that started at a 420W and allowed up 500W, of course 500W did nothing for gaming
Ironically, that technique usually pans out since a lot of useless fads get hyped. But every once in a while it causes people to completely and utterly miss the point.
It's so much more compute efficient for my goals, though I do deal with subjective preference which makes exploration during training much harder
> Agent observability is crowded, so instead of a bullet list of how Laminar differs from X, we’ll explain how we’re solving agent observability from first principles
my brain shut down and bounced minus the comment because it's a launch and maybe you'd want feedback.
I'd be using this for post-training in a custom harness: so what you focused on during post-training and what it excels at are the most interesting things
At a micro scale their code is right, but tons of invented complexity no sane person would introduce. I've found the two most common triggers are the models polluting their own context windows (poorly worded instructions in comments/issues/AGENTS.md) and a poor "taste" for breaking changes (they'll constantly try to make their additions non-breaking, and if prompted they can make breaking changes, it has to be very carefully scoped or they start assigning great reward to ripping out things)
Which is interesting... after gpt-image-2 came out with this bug replacing the piss yellow filter, it seems like every other release except Google's have the same issue.
Off the top of my head, this exact artifact appeared in Qwen, Grok, and now Krea.
It makes me wonder if gpt-image-2 is being used to bootstrap post-training, or if their images are just so ubiquitous it's turned into a sort of "virus".
There were multiple suicides linked to the app with multiple states trying to legislate their vertical... at that point this diversion about who's confirming they did the obvious thing of hiring lobbyists is pretty moot.
That is literally a subreddit of first hand accounts with this app having kids talk about relapse and withdrawl.
https://secureainow.org/clipping/ai-lobbying-shifts-congress...
I don't know if this is maybe a denial because of a personal connection to the company, but there are subs where kids as young as 12 openly talk about withdrawls and coach each other through relapses from C.ai and "the purple ghost app" (Polybuzz)
https://www.reddit.com/r/character_ai_recovery/
https://www.reddit.com/r/character_ai_recovery/comments/1wzz...
The top level comment blamed cheerleading but let's apply critical thinking for a second as product people and look at the funnel:
1. Potential users sees new release 2. Potential user takes time to install and set up Google harness (you already lost 90% of people here, the funnel better close quick!) 3. Potential user cannot use new model.
Then what? You hope they developed good will towards your model through this experience?
You stuck a two week wait in your funnel? Is that the definition of not having your act together or what?
We're on HN, how many of us have heard from Google staff this kind of stuff is a result of them growing so big that AI Cloud has to compete with Vertex Cloud, and AI cloud exists because Vertex Cloud needed 2 years to get OAI/Ant style API keys working and were losing dev mindshare in the meantime?
You worded it as if it was some pithy gotcha, and I'm sure "lOw dAtA rEgiMeSs doNt eXist aNymOre" would have done numbers on LinkedIn, so maybe you thought it was new.
But no, it's right up there with "the cost of software is zero" and "saas is dead"
Refrains of the clueless grifter who deserves all the condescension one can muster.
Good engineers don't treat approach as A == B or even A like B, when extremely integral parts of their applications differ.
Zero-shot isn't just "more convenient", in a low data regime: it's the only workable solution, and 100x so if your plan involves the acornym "BERT" (because even the largest of those models has the world knowledge of a fart to draw priors from)
Better ergonomics while being faster and cheaper as the existing things really is enough to justify callling what you've done a new thing, in a world of finite resources and time. It's actually making me scream how many people don't get that.
Much more commonly, the statement is used by writers who want to wax poetic without actually having to commit to any meaningful position on the topic at hand... which is exactly how the AI-written article in question uses it.