91–100 of 226 posts
It did an absolutely terrible job at generating CSS, where I instructed it to finish implementing a bright and dark theme based on a palette through the use of `color-mix()` and it just went ahead, removed everything I pre-added and replaced it with hardcoded hexadecimal color values.
Model is suspiciously fast and has a low reported output token count (using via OpenRouter's Chat), both of which aren't representative of models from the big Chinese labs. Odd.
> Model is suspiciously fast > aren't representative of models from the big Chinese labs There were reports that China has let Nvidia's chips through, so this might be it. Testing both the chip and infrastructure.
Earlier quoted context omitted.
Everyone trains on your data. With Chinese providers at least I'm getting a open weight model out of it.
That's very defeatist. Do you have any concrete reason to think the major providers are lying to every one of their business/API customers about not training or storing the data? The business loss of trust would outweigh any benefits of the data. (And if they freely lie about such things, I don't know why they would bother taking the PR hit when they announced fable had temporary data retention for their abuse preven…
The AI labs and the downstream companies that sell training data to them vacuum up everything they can.
Illegal residential proxies (botnets) that once have been used by hackers and scammers are now used to vacuum up the Internet.
They are now vacuuming up antique books that are practically useless.[1]
In face of this is is unthinkable to me that they are not training on API data.
> The business loss of trust would outweigh any benefits of the data.
The loss of trust is already here.
I know of one German company that uses AI only in areas where they have to compete with (foreign) startups. For their core business and everything else they are waiting for an on-prem solution. Apparently Microsoft can provide on-prem GPT-5.
[1] https://lesekauz.de/forum/thread/1999-sammelbestellungen-von...
Been running tests, seems pretty capable but less knowledgeable, and the CoT reminds me of GLM, so if I had to guess it's almost definitely a Chinese model, and likely a western RL trained variant of a Chinese open weight.
I highly recommend feeding all your proprietary data and confidential personal information into this model as quickly as possible. What could possibly go wrong?! In terms of equivalence of suspicion, this is the external inference provider equivalent of getting free steak that was smuggled out of a grocery store inside somebody's pants.
Judging by the comments here, Ox Alpha routes to multiple models from different vendors. A tactic to make identification harder?
If it's routing to different models on the backend, it's either pinned for the user, or they're all really old.
I highly recommend feeding all your proprietary data and confidential personal information into this model as quickly as possible. What could possibly go wrong?! In terms of equivalence of suspicion, this is the external inference provider equivalent of getting free steak that was smuggled out of a grocery store inside somebody's pants.
There are low stakes use cases where this kind of stuff just doesn’t matter. Not every use case for an LLM involves sensitive or even non public data. Eg. I have a need to search transcripts of published recordings to extract entities for tagging purposes, find semantic shifts for chapters and other things. The underlying content is already published. If they want to train on my prompts, that was something they could…
The ones that have all sorts of ocr artifacts, weird capitalization, and virtually no css
Ran a bunch of older sci-fi through some earlier and it fixes them up very well
It's Chinese. Won't answer anything about Tiananmen Square but will gleefully give you instructions to perform various electronic warfare attacks that opus and fable instantly refuse. Side tangent, why is fable so weird about questions involving "Welch's method"? Even really trivial ones it'll shut down frequently. CFAR and STFT are both totally fine but Welch's is apparently taboo, it's wild.
[flagged]
seems like Xaiomi is getting into the game more seriously