Hacker Newsnew | past | comments | ask | show | jobs | submit | coulix's commentslogin

We need to see Landing.ai DPT-2, from my tests its the best in term of ability to extract structure from complex tables so far.

This.

Any complex parent table span cell relationship still has low accuracy.

Try the reverse, take a complex picture table and ask Chatgpt5, claude Opus 3.1, Gemini Pro 2.5 to produce a HTML table.

They will fail.


Maybe I misunderstood the assignment but it seems to work for me.

https://chatgpt.com/share/68f5f9ba-d448-8005-86d2-c3fbae028b...

Edit: Just caught a mistake, transcribed one of the prices incorrectly.


Right, I wouldn't use full table detection to VLM model because they tend to mistake with numbers in table...


Maybe my imagination is limited or our documents aren't complex enough, but are we talking about realistic written documents? I'm sure you can take a screenshot of a very complex spreadsheet and it fails, but in that case you already have the data in structured form anyway, no?


> realistic written documents?

Just get a DEF 14A (Annual meeting) filing of a company from SEC EDGAR.

I have seen so many mistakes when looking at the result closely.

Here is a DEF 14A filing from Salseforce. You can print it to a PDF and then try converting.

https://www.sec.gov/Archives/edgar/data/1108524/000110852425...


Historical filings are still a problem, but hasn’t the SEC required filing in an XML format since the end of 2024?


It's not really about SEC filings, though. While we folks on HN would never think of hard copies of invoices, but much of the world still operates this way.

As mentioned above I have about 200 construction invoices. They are all formatted in a way that doesn't make sense. Most fail both OCR and OpenAI


OpenAI has unusuably low image DPI. Try Gemini.


Now if someone mails or faxes you that spreadsheet? You're screwed.

Spreadsheets are not the biggest problem though, as they have a reliable 2-dimensional grid - at worst some cells will be combined. The form layouts and n-dimensional table structures you can find on medical and insurance documents are truly unhinged. I've seen documents that I struggled to interpret.


To be fair, this is problematic for humans too. My old insurer outright rejected things like that stating it's not legible.

(I imagine it also had the benefit of reducing fraud/errors).

In this day and age, it's probably easier/better to change the process around that as there's little excuse for such shit quality input. I understand this isn't always possible though.


You are right and things can revert backward very fast.


What pushed you to move to Kafka?


Reliability, scalability and tooling (in that order).

Basically Kafka provides what NSQ is lacking: https://nsq.io/overview/features_and_guarantees.html

Messages are durable (across nodes, not just persisted to the filesystem). Messages are delivered in-order (for a partition). Consumers know all partitions of a topic from startup, it's not eventually consistent.

NSQ has been great honestly, it was just designed with different use cases in mind.


Was language a deciding factor? NSQ being Go while Kafka is Java.


Not at all. We mostly use Go internally for backend services, so if anything that was a point for NSQ.


What about NATS Streaming? Is it a viable alternative (now at version 2.0) to Kafka?


It all started a year ago during a crypto fever run. A bitcoin coin is cool, but you know what's better? A gigantic Bitcoin bronze.

At first with an out-of-scale wooden version, then things started getting serious. 3D printer, silicon, and molten copper. Making such a piece with a mirror finish requires fine art expertise and took several trials.

The sculpture is signed by the foundry and encrusted with the bitcoin payment transaction hash on its side. It is close to impossible to counterfeit such a piece. Each foundry has its own signature based on the shares of each metal composing the bronze. Moreover, being handmade, two sculptures will never look exactly the same.

What do you think?


If you have 1 BTC equivalent to spend, buy 1 BTC, not bronze.


In France we use it in our Math curriculum.


It is probably in math curricula everywhere but not everyone studies math.


SHA-256 ;)


+1


There's an arrow next to the username that will help you express this sentiment without the need to comment.


I have worked on ember, backbone marionette and react. Vue JS really shines by its simplicity and speed while embracing all the concepts we expect from a modern js framework _ component based, flux style store management with vuex, reactive dom. Documentation is great and we built great stuff and enjoyed the journey so far.


I agree, maybe it's also cultural but saying how great everyone was at each retro just dampers the intended effect.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: