Hacker News new | ask | show | jobs
by d_watt 681 days ago
I really want 365 or gemini to do the obvious thing: "Using this template, summarize this document into a slide deck."

Maybe it won't be good, but it's surprising to me it can't do it barebones. Considering LLMs do an okay at summarizing into markdown, maybe the issue is a lack of proper intermediate DSL to represent slides it can train on?

But autogenerating at least a scaffolding for a deck seems so obvious, it's pretty damning it's so bad 2 years later.

5 comments

All my documents - Word, PowerPoint, OneNote, Excel - are all on OneDrive. Give me an NLP interface to query information from these sources. When providing the information, note the source where it came from. That's what I want!
Give me good sharepoint search!!

I'd love the R part of RAG, but Microsoft can't even do that it appears?

Wasn't that a specifically advertised copilot feature? Does copilot not actually support it? They are still claiming to support this it seems https://support.microsoft.com/en-us/office/create-a-new-pres...
When I tried it with a copilot trial at one business, it basically just dumped huge text blobs into big text fields.

They push the "generate image based on prompt" as the main add for slides to make them not just bullets, but no real foundational slide creation is baked in.

Things may have changed since I demo'd it ~6 months ago.

Google Slides is just generated images and transform text as far as I've seen.

We have information and we need to organize it so it can be presented so leadership can make a decision.

We now have so much information that the effort needed to take this step is so obnoxious that need an abstract organization system to exist on top of it.

I wonder if the scaling properties here just serve to highlight the deeper organizational problem that we still haven't solved properly yet.

Many years ago, someone remarked in a meeting it was too bad we (Microsoft) couldn't ship these PowerPoints.

Yet, here we are.

The problem is that’s requires them to do actual product discovery work.

Right now, they can chuck out a terribly implemented LLM.