vignesh_warar

Thanks! Yes, it’s on our roadmap. will try to ship it soon.

Hey, we use Bing + Google + the tiny index I built from my previous projects.

https://news.ycombinator.com/item?id=35826540

https://news.ycombinator.com/item?id=35510949

Currently, it uses both Google and Bing, but I am planning to add a small index from curated pages from forums like HN and Reddit, similar to what I did before

https://news.ycombinator.com/item?id=35826540

https://news.ycombinator.com/item?id=35510949

Sorry, I temporarily paused it. https://news.ycombinator.com/item?id=42194170#42202922 reply

Sorry, I temporarily paused it. https://news.ycombinator.com/item?id=42194170#42202922

Sorry, I temporarily paused it. https://news.ycombinator.com/item?id=42194170#42202922

Thank you for the feedback, everyone.

I am temporarily pausing and making sure I am within legal limits. If not, I will completely remove face recognition and try other routes to solve the problem.

Vignesh

It's just the Bing search API under the hood. The process is: Query -> Crawl -> Categorize profiles.

Thank you for trying. We are being rate-limited. We are currently fixing it.

Edit:

Fixed!

Thank you!

Exactly - this won't be useful for stalkers. We don't crawl or index social media pictures.

Will add FAQ section.

Thank you for the feedback :)

No, we don't. I think my copy might have been confusing. We are simply Perplexity for people - we just summarize a person's internet presence.

I just did research on this.

Clearview vs Introthem:

- Clearview does photo-to-photo matching. We don't do that, and I don't think I will ever build that.

- You have to provide the name, then we build the faces collection for analyzing at search time and delete it.

- We don't retain any face collection once the search is done.

I still don't know if I am breaking any laws, but here is how Introthem works.

Thanks for the info. We don't use any private data, only publicly available images. So it won't be a problem, in my opinion.

I will contact my lawyer and double check this.

Hey, developer here.

Thank you for the feedback.

I don't think I'd ever want the "chat with your screenshots" feature though

Here is the reason why this might add some value for the users: Personally, I take screenshots of web designs from time to time. So, when I design, I should be able to ask questions like, "What was the website that had a fancy gradient button?" This will return screenshots + website address(OCR) for reference.

I am not sure how useful this will be for everyone, though. I am still talking to users.

No, OpenAI implemented some safety measures: https://platform.openai.com/docs/guides/vision#:~:text=captc....

True, I will change it.

Hey, it is fully open-source; there are no restrictions. We are only charging for the hosted version. I mean, as the founder, I somehow need to pay my bills.

Hey there, Author here,

I agree — I should have done a better job writing the copy.

Here are some tasks you can accomplish with an AI employe:

- While I haven't optimized for testing, you can perform end-to-end tests simply by describing it. If the UI is complex, the AI employe will hallucinate; just click 'start recording' and demonstrate the tasks. It only records HTML changes, not the screen, microphone, or camera. and you don't have to worry about IDs or selectors—the AI employe will take care of it.

- Tasks related to OCR, like extracting data from PDFs, emails, and transferring it to CRMs/ERPs.

- Summarization tasks; for ex, there's a workflow example in AI employe where it automatically summarizes Hacker News comments for the website you're browsing. It automatically do the HN search, clicking the relevant link, and summarize the comments.

Since there are a couple of things an AI employe can do, I am simply confused about how to position it. Any feedback on how to position it is most welcome.

Hey, Repo author here, My intention is not to replace humans but to create more of a helping hand for them.

Hey, no worries. 'Employe' is not a typo; it is a word that carries the same meaning [1].

AIEmployee.com is not available, so I purchased something similar.

1: https://english.stackexchange.com/questions/264834/employee-...

It is not a typo.

aiemployee.com domain has already been taken, so I was looking for a word that is something close. It turns out 'Employe' is a real word, so I registered it.

"Employe is a rare dated alternative spelling of the more common employee (AHD)" [1]

1: https://english.stackexchange.com/questions/264834/employee-...

Fixed!

How does this work? Is data sent from browser to homebase (some cloud server?) and then openAI - so two or more third parties

Yes

Privacy - no policy?

We do have privacy page: https://aiemploye.com/privacy

Hello everyone,

I am happy to open-source AI Empoye: GPT-4 Vision Powered First-ever reliable browser automation that outperforms Adept.ai

Product: https://aiemploye.com

Code: https://github.com/vignshwarar/AI-Employe

Demo1: Automate logging your budget from email to your expense tracker

https://www.loom.com/share/f8dbe36b7e824e8c9b5e96772826de03

Demo2: Automate log details from the PDF receipt into your expense tracker

https://www.loom.com/share/2caf488bbb76411993f9a7cdfeb80cd7

Comparison with Adept.ai

https://www.loom.com/share/27d1f8983572429a8a08efdb2c336fe8

Love to know your feedback.

Hello everyone,

I am happy to open-source AI Empoye: GPT-4 Vision Powered First-ever reliable browser automation that outperforms Adept.ai

Product: https://aiemploye.com

Code: https://github.com/vignshwarar/AI-Employe

Demo1: Automate logging your budget from email to your expense tracker

https://www.loom.com/share/f8dbe36b7e824e8c9b5e96772826de03

Demo2: Automate log details from the PDF receipt into your expense tracker

https://www.loom.com/share/2caf488bbb76411993f9a7cdfeb80cd7

Comparison with Adept.ai

https://www.loom.com/share/27d1f8983572429a8a08efdb2c336fe8

I am happy to open-source AI Empoye: GPT-4 Vision Powered First-ever reliable browser automation that outperforms Adept.ai

My bad. Thanks! I fixed it!

I might be wrong here. I just know some product quantization techniques, but you can reduce the index by a lot! However, from my research, the more size you reduce, the more retrieval quality is also reduced.

Quoting from https://github.com/criteo/autofaiss

Using faiss efficient indices, binary search, and heuristics, Autofaiss makes it possible to automatically build in 3 hours a large (200 million vectors, 1TB) KNN index in a low amount of memory (15 GB) with latency in milliseconds (10ms).

I would highly recommend taking a look at AutoFaiss. You just have to set the maximum memory to build the index, and it will come up with the configuration.

Interesting!

If anyone is seeking a math formula for primes, here is one: https://en.wikipedia.org/wiki/Formula_for_primes

There is also a good YouTube video that explains this: https://www.youtube.com/watch?v=j5s0h42GfvM

Happy to see my favorite extension on HackerNews! I have been using this extension to learn French, and it works quite well for me.