kbyatnal

7 ms·

1.
by kbyatnal 4mo ago
agreed! dogfooding is the best — we'll integrate this into our core product, which will force us to improve it very quickly
2.
by kbyatnal 4mo ago
Your issues are mostly with the file picker, which is a brand new component that was just added yesterday in response to a feature request after launch: https://x.com/andrewlu0/status/2064726998186971400?s=46 It h
3.
Show HN: Extend UI – open-source UI kit for modern document apps(extend.ai)
252 points by kbyatnal 4mo ago|81 comments
4.
Searching The Pentagon UFO Files(ufo-files-declassified.com)
1 points by kbyatnal 5mo ago|0 comments
5.
PoliTax Split: PDF splitting benchmark from presidential tax returns(extend.ai)
1 points by kbyatnal 6mo ago|0 comments
6.
How we built a prompt optimization agent(extend.ai)
5 points by kbyatnal 7mo ago|0 comments
7.
by kbyatnal 8mo ago
Deepseek OCR is no longer state of the art. There are much better open source OCR models available now. ocrarena.ai maintains a leaderboard, and a number of other open source options like dots [1] or olmOCR [2] rank higher. [1] https:&#x2F
8.
by kbyatnal 10mo ago
Ultimately, there’s some intersection of accuracy x cost x speed that’s ideal, which can be different per use case. We’ll surface all of those metrics shortly so that you can pick the best model for the job along those axes.
9.
by kbyatnal 10mo ago
Claude coming shortly (in the next ~1 hour)
10.
by kbyatnal 10mo ago
We wanted to keep the focus on (1) foundation VLMs and (2) open source OCR models. We had Mistral previously but had to remove it because their hosted API for OCR was super unstable and returned a lot of garbage results unfortunately. Paddl
11.
by kbyatnal 10mo ago
Sonnet/Opus is being added shortly!
12.
Show HN: OCR Arena – A playground for OCR models(ocrarena.ai)
216 points by kbyatnal 11mo ago|63 comments
13.
by kbyatnal 1y ago
Yeah that can occasionally work and something we've tested, but it introduces a lot of noise unfortunately and makes systematic evals difficult.
14.
by kbyatnal 1y ago
School transcripts are surprisingly one of the hardest documents to parse. The thing that makes them tricky is (1) the multi-column tabular layouts and (2) the data ambiguity. Transcript data is usually found in some sort of table, but they
15.
by kbyatnal 1y ago
thanks! Datalab is great, I've met Vik a few times and their team has done some impressive work. We can also support the conversion to markdown use case, and might be a better fit depending on your use case. Feel free to create an acco
16.
by kbyatnal 1y ago
It's very dependent on the use case. That's why we offer a native evals experience in the product, so you can directly measure the % accuracy diffs between the two modes for your exact docs. As a rule of thumb, light processing mo
17.
by kbyatnal 1y ago
Exactly correct! We've had users migrate over from other providers because our granular pricing enabled new use cases that weren't feasible to do before. One interesting thing we've learned is, most production pipelines often
18.
by kbyatnal 1y ago
Feedback heard. Pricing is hard, and we've iterated on this multiple times so far. Our goal is to provide customers with as much transparency & flexibility as possible. Our pricing has 2 axes: - the complexity of the task - perform
19.
by kbyatnal 1y ago
good question! Our goal is to provide customers with as much flexibility as possible. For certain use cases, you might be willing to take a slight hit to accuracy in exchange for better costs and latency. To support this, we offer a "l
20.
by kbyatnal 1y ago
thanks! A lot of customers choose us for our handwriting, checkbox, and table performance. To handle complex handwriting, we've built an agentic OCR correction layer which uses a VLM to review and make edits to low confidence OCR error
21.
by kbyatnal 1y ago
There's certainly a lot of tools that focus on individual parts of the problem (e.g. the OCR layer, or workflows on top). But very few that solve the problem end-to-end with enough flexibility for AI teams that want a lot of control ov
22.
by kbyatnal 1y ago
thanks! Yup that's correct, we offer a set of APIs for handling documents: parsing, classification, splitting, and extraction. We've seen customers integrate these in a few interesting ways so far: 1. Agents (exposing these APIs a
23.
by kbyatnal 1y ago
There's definitely no shortage of options. OCR has been around for decades at this point, and legacy IDP solutions really proliferated in the last ~10 years. The world today is quite different though. In the last 24 months, the "T
24.
by kbyatnal 1y ago
thank you Fabio!
25.
Launch HN: Extend (YC W23) – Turn your messiest documents into data(extend.ai)
61 points by kbyatnal 1y ago|33 comments
26.
by kbyatnal 1y ago
Extend | Senior Software Engineer, ML Engineer, AI Engineer | NYC | Full-time | $250k-$350k + equity Extend is building a LLM-native document processing platform (massive market w/ low NPS existing solutions) ( https://extend
27.
Extend (YC W23) is hiring engineers to build SOTA document processing(jobs.ashbyhq.com)
1 points by kbyatnal 1y ago
28.
by kbyatnal 1y ago
Thanks for the reply. Not sure what you're referring to, but I don't believe we've ever copied or taken inspo from you guys on anything — but please do let me know if you feel otherwise. It's not a big deal at the end of
29.
by kbyatnal 1y ago
Founder of Extend ( https://www.extend.ai/ ) here, it's a great question and thanks for the tag. There definitely are a lot of document processing companies, but it's a large market and more competition is always be
30.
Extend raises $17M to build a document processing cloud(extend.ai)
1 points by kbyatnal 1y ago|0 comments

More ›