This is awesome, I was thinking of creating an agent to generate spec and sequence diagrams (using Mermaid), but this takes it to a new level being generic. I tried a question about how does LLM post training work and it made a nice workflow about it.
If there were a way to magically ensure that only the good guys had access to the information, I'd be way less concerned about these measures. (There are only a few things that I care about being secret for the sake of secrecy, and I can easily keep those off of computers.)
Every encryption backdoor is a huge vulnerability. Even if we somehow ensure that the powers-that-be remain entirely trustworthy (something that, historically, we can't even manage for a century), they're not the only people who'll have access to the backdoor. It's not possible to make an encryption backdoor that only authorised parties can use: as they say, the laws of mathematics do not respect the laws of Australia.
I assume thinking medium and hard would consume more tokens.
I feel the timing is bad for this release especially when deepseek R1 is still peaking. People will compare and might get disappointed with this model.
My vibe question checking suggests otherwise. Even o3-mini-high is not as good as r1, even though it's faster than r1. Considering o3-mini is more expensive per token. It's not clear o3-mini-high is cheaper than r1 either even r1 probably consumes more token per answer.
I agree with most points, however going viral has the benefit of strong backlinks for SEO.
I had a website I shared here on HN, which made it to the front page and resulted in being mentioned in multiple newsletters and websites. The website site was on Google's 1st page when you search for the term, all due to the strong backlinks I gained.
More of a system: AWS Textract or Azure Document Intelligence. This option requires some coding and the cost is higher than using a vision model.
reply