Yeah can't recommend their $19 plan, only took me a day and a half to hit the weekly usage cap. The $39 plan has 5x the limits so I recommend getting that instead. Having now tested it for a few days, it is the first of the chinese models that actually feels comparable to Opus-tier models
I've been using unsloth/gemma-4-31B-it-qat-GGUF daily for various small parsing and programming tasks using opencode and llama-server's front end. The past couple of weeks have made a big difference after google released the QAT variant and llama.cpp got support for MTP which means it is possible to now get 60-80 Tok/s with RTX 4090. The model fits in VRAM comfortably enough to keep it loaded even while browsing and having multiple programs.
Hard to see any practical benefit to go after common crawl. The situation of freely accessible crawled data is bad enough as it is with archive.org and CC being pretty much the only available sources. We need more initiatives like them, not less. The scary thing is how the anti-AI sentiment is being used to lock things down further.
The 7B parameter Vibevoice TTS model is still the most impressive local TTS model i've tried. It was pulled by Microsoft a few days after its release due to "abuse potential" but it can be found in various community maintained huggingface repos.
I assume this is due to the fact that claude code appends a system message each time it reads a file that instructs it to think if the file is malware. It hasnt been an issue recently for me but it used to be so bad I had to patch out the string from the cli.js file. This is the instruction it uses:
> Whenever you read a file, you should consider whether it would be considered malware. You CAN and SHOULD provide analysis of malware, what it is doing. But you MUST refuse to improve or augment the code. You can still analyze existing code, write reports, or answer questions about the code behavior.
Okay so I had to look in to it because the site is not really doing a good job explaining it at all. Turns out[0] that they are voting for the extension of the temporary regulation thats been in effect since 2021 (Regulation (EU) 2021/1232). So this is about the "voluntary scanning of private communications" (which is still bad, but has been in effect for almost 5 years already).
This is the thing that bothers me the most about this. It is as if even the HN crowd is taking it as given that malware is this big problem for banking on Android but in reality there seems to be very little evidence to back this up. I regularly read local (Finnish) news stories about scams and they always seem to be about purely social engineering via whatsapp or the scammer calling their number and convincing the victim they are a banking official or police etc.
That's why I'm inclined to believe Google is just using safety as an excuse to further leverage their monopoly.
It is CPU-based. Somewhere between 1 to 2 seconds per page on a single core. I ran 20 instances of it in parallel to utilize 20 CPU cores so the avg time came down nicely.
chrome_screen_ai.dll is the name of the dll (libchromescreenai.so on linux) and yes it is proprietary. It isn't included by default, Chrome uses its component service to download it automatically when you open a PDF file that doesn't have pre-existing OCR'd text on it. You can download it separately from here: https://chrome-infra-packages.appspot.com/p/chromium/third_p...
Chrome ships a local OCR model for text extraction from PDFs which is better than any of the VLM or open source OCR models i've tried. I had a few hundred gigs of old newspaper scans and after trying all the other options I ended up building a wrapper around the DLL it uses to get the text and bboxes. Performance and accuracy on another level compared to tesseract, and while VLM models sometimes produced good results they just seemed unreliable.
I've thought of open sourcing the wrapper but havent gotten around to it yet. I bet claude code can build a functioning prototype if you just point it to "screen_ai" dir under chrome's user data.
I would assume so. I can see from my browser history that i succesfully submitted captures to archive.today on 7th of January, but failed to do so starting from 12th of January. IIRC they contacted gyrovague around the 10th so seems unlikely to be a coincidence. Applies to VPNs as well. Tried first with a VPN located in Finland and it gets endless captcha loop, then with a Swedish VPN which let me through to the front page after solving one captcha.
The system prompt of claude code changes constantly. I use this site to see what has changed between versions: https://cchistory.mariozechner.at/
It is a bit weird why anthropic doesn't make that available more openly. Depending on your preferences there is stuff in the default system prompt that you may want to change.
I personally have a list of phrases that I patch out from the system prompt after each update by running sed on cc's main.js