Your nav menu on mobile doesn't collapse once you select an item. Couodnt tell if the link worked or not until I closed the menu. Maybe have it collapse on click?
There are certain demographics you can follow where there is a constant wave of hype of simple products, copy/pasting/forking, SaaS, yada yada. This has become fairly prominent since ChatGPT.
Specifically Github provides a feed, but also the ability to view other's stars. Stars are particularly useful for trending discovery in whatever domain/community you're interested in.
Any service that does not prominently show who is responsible, don't even bother using it. If the owners are not willing to put themselves on their own company website, you should not be using that service/site.
Gotta give them credit for at least having a public pricing page though. That is something I will NEVER understand, some companies somehow think that making this information a PITA to access somehow helps their business?
Your FAQ is clueless about corporate data concerns relative to the price points you're offering.
How does we prioritize privacy protection?
(grammar)
At Hansei, safeguarding your privacy is paramount. Your data is end-to-end encrypted and can only be accessed by you. You can delete your data anytime from our servers. Query and response data are stored to improve our services, adhering to our comprehensive privacy policy(https://hansei.app/privacy-policy). Queries and responses also pass through OpenAI and are subject to their privacy policy as well.
Most companies don't have a privacy officer, they have a data officer, and they don't worry about privacy, they worry about data loss, IP loss.
Even then, you both dissemble and flat lie in this answer, with the contradictions evident in the response itself.
End to end encrypted? What end to what end, and in what way that's not just SSL/TLS over the https request?
Accessed only by you? Not by, say, an embeddings tool, or not by, say, OpenAI where they "also pass through"? What about DBAs, SREs, Dev/QA, and the Founder/CEO?
More specifically, how do you (in logic, not policy) prevent yourself from being able to access the company's data even in case of break glass need, such as a law enforcement warrant? If someone uploads a document store of CSAM or bomb building instructions, you're claiming you cannot be responsive to a search warrant or national security letter? Think carefully before you answer. Being able to say this while performing a service using the data can take years of work and code/config/infra/system automation.
Understood you say nobody can access but then query and response (are you saying response doesn't contain data?) are stored "to improve services". In what way, other than -- ultimately -- humans learning this data incidentally, while "improving"?
This answer is transparently end-to-end nonsense to a CDO or CISO. The only way it could have been worse is if you'd had that "we take your security very seriously" stock phrase with an appeal to military/bank grade security or bits of encryption there.
Hey there, first off, I want to say a big thank you for taking the time to delve into our privacy practices and for your valuable feedback. We appreciate the scrutiny – it gives us an opportunity for reflection and growth.
What we mean by "end-to-end encryption" here is that the data being encrypted at rest and in transit. It's not just about SSL/TLS over HTTPS, but also about the data being encrypted in the storage.
When we say "accessed only by you", we mean the user has control over their data. Yes, currently the data does pass through OpenAI for embedding and query response as we're using OpenAI models and subject to their privacy policy.
Regarding the access of data, it's important to note that it is strictly controlled and monitored. In terms of law enforcement warrants, we comply with legal obligations. If required, we can provide data, but it's not a regular practice and is done under strict legal procedures.
Yes, the "query and response" data is used to improve services. But it doesn't mean humans are learning from this data. It's more about machine learning algorithms using this data to enhance the user experience.
I hope this clears up most of your concerns. Our priority is to offer value-aligned services to our clients, and their data safety is a part of this commitment. We might not be perfect, but we strive for continuous improvement in our data security and privacy deployment.
This the first thing I look for to establish the credibility of any site, app, or service. It's marginally OK to link to a LinkedIn page, but you have to populate it with credible information. Seeing 0 employees is a huge warning that you should not use the service. This causes me to look harder to understand why this information is obscured.
Very nice. Although I wonder how this is sustainable as a business when it seems like every day someone is releasing a frontend over LangChain/GPT for free. This week alone I think one collected at least 5 open source projects that can do the same thing. This shouldn't discourage anyone from attempting it though. It's a super crowded space and inevitably, this process will become part of the OS soon enough. Anything in your home folder will be transparently sliced, vectorized and stored locally by the OS, with priority API access to their cloud LLM backends.
I have no way connection with this product or their makers.
But do every product has to be different from every other product? Is it a criteria for success ? or, you are making a statement that if someone else builds it, another person should not build the same ?
I think the intent of the question was not to discourage people from doing something that has been done. That's how we all learn. The question is what makes this different than the 5 other OpenAI API frontends released in the past 3 days alone, for free, that you feel like it is worth charging for? Since the LLM is doing the heavy lifting, the stack used to build this can run on a lemon. Python and Node are just the glue needed to do it with little effort in less than a week.