ChatGPT Launches 'Company Knowledge' • Hacker News Board

Re: ChatGPT Launches 'Company Knowledge'

#21

Post by ares623 » Fri, Oct 24, 2025, 9:44 AM UTC

> OpenAI never trains on your data by default.

Are embeddings used for RAG considered company data (presumably calculated by OpenAI), or OpenAI’s?

(I don’t know if that’s how RAG actually works)

Re: ChatGPT Launches 'Company Knowledge'

#22

Post by blitzar » Fri, Oct 24, 2025, 9:45 AM UTC

cantor_S_drug wrote: post #17

blitzar wrote: post #14

Earlier quoted context omitted.

My startup is working on a similar problem ... inbound in 5 ...

Already exists. https://www.glean.com/

You get all the upvotes I have to give if your photo is on that leadership page

Re: ChatGPT Launches 'Company Knowledge'

#23

Post by tehmillhouse » Fri, Oct 24, 2025, 9:45 AM UTC

incoming1211 wrote: post #11

V__ wrote: post #10

Quick, let's ship another feature before the market realizes it's an ai bubble..

ai isn't a bubble.

I wish my forehead had more headroom for the amount of travel my right eyebrow needs to accommodate this statement.

Re: ChatGPT Launches 'Company Knowledge'

#24

Post by arnvald » Fri, Oct 24, 2025, 9:45 AM UTC

This is actually an interesting space, and I think there's a room for such products. Large companies struggle a lot with their knowledge bases, and discoverability is part of the problem (especially when using multiple tools: Confluence, Docs, Chat app, etc.)

The problem here is that there are companies that focus on this area and keep improving their products, while for OpenAI it's one of dozens of tools they launch, so it's hard to believe they'll keep dedicating adequate resources to make this a mature tool that's worth the investment (in form of time and money) for the clients

Re: ChatGPT Launches 'Company Knowledge'

#25

Post by tirumaraiselvan » Fri, Oct 24, 2025, 9:45 AM UTC

Now we know why OpenAI told its investors to avoid Glean[1]

[1]: https://www.reuters.com/technology/openai-tells-investor-not...

Re: ChatGPT Launches 'Company Knowledge'

#26

Post by dist-epoch » Fri, Oct 24, 2025, 9:45 AM UTC

hansmayer wrote: post #16

And in the next episode of "Pointlessly Searching For The Killer App"....

They already have The Killer App - ChatGPT

Re: ChatGPT Launches 'Company Knowledge'

#27

Post by matusp » Fri, Oct 24, 2025, 9:48 AM UTC

shanehoban wrote: post #18

Claude memory was posted 16 hours before this, kinda crazy how fast they churn this stuff out

They raised billions of dollars, and they're obviously spending it to develop just about anything they can think of. What's rather concerning for them is that the core AI seems to have pretty much stagnated by now.

Re: ChatGPT Launches 'Company Knowledge'

#28

Post by rvz » Fri, Oct 24, 2025, 9:50 AM UTC

incoming1211 wrote: post #11

V__ wrote: post #10

Quick, let's ship another feature before the market realizes it's an ai bubble..

ai isn't a bubble.

The parent comment is an indicator that we are in a AI bubble.

Re: ChatGPT Launches 'Company Knowledge'

#29

Post by consp » Fri, Oct 24, 2025, 9:50 AM UTC

incoming1211 wrote: post #11

V__ wrote: post #10

Quick, let's ship another feature before the market realizes it's an ai bubble..

ai isn't a bubble.

Looks a lot like web2.0 though.

Re: ChatGPT Launches 'Company Knowledge'

#30

Post by Leynos » Fri, Oct 24, 2025, 9:51 AM UTC

lukax wrote: post #4

> It’s powered by a version of GPT‑5 that’s trained to look across multiple sources to give more comprehensive and accurate answers. So another GPT-5 fine-tune. Codex also uses a custom GPT-5 fine-tune. Does fine-tuning make sense now? Or do you have to be OpenAI to fine-tune the models with a mix of existing data and new behaviours?

I think that you have to be OpenAI (or X, Google or Anthropic) to be able to fine tune models of this scale through reinforcement learning at present.

Look at Tinker for an example of where things might be heading though (https://tinker-docs.thinkingmachines.ai/)

At present though, I get the sense that reinforcement learning at scale is the current battleground (and has been for most of 2025). But we also see over time, the general models adopt the skills taught to the specialized models. Look at how the learning that made codex-1 went into GPT5.