Is it bad that I look at every library now, no matter how awesome, and think "Well this didn't exist on or before September 2021, so do I really wanna learn this?"
I mean surely an interface of the form blah blah blah, "Now, take the content and turn it into slides using a document creation API" is the best right?
Should every library include a documentation endpoint for Scraper Plugin of the right context length? Is prompt injection an attack, or an opportunity? This is new ground for adoption.
This is a real effect where newer (and maybe objectively better) tech is now at a disadvantage because LLM hasn't been trained on it (or there aren't enough examples to yield good results).
Of course being an "early adopter" has always been a challenge, but I suspect the lack of LLM capabilities will slow adoption even more.
Wow, that’s an incredibly lazy and short-sighted take from the GP then. The training set will definitely get updated at some point… or you could just ask Bing, or Bard
For a different approach, but still in the handwavey space of presentation apps that are actually pleasant to use, I'm a big fan of iA Presenter (https://ia.net/presenter). Short version: write Markdown, then Presenter makes it look nice. Export your deck to a PDF when you're done so you can pass it around. Write CSS for styling, or use their online theme builder to do it for you.
And while it's commercial software, it has a pay-once option that I instantly purchased.
We've been using a non-commercial alternative, marp, at work to great success and make slides where PowerPoint usually sucks: code blocks (we present on our data format frequently).
I love this, seems really great for giving presentations on prototypes.
Is there any reason why this only works for Python?
Seems like you could just give it a variable to specify the comment delimiter and a variable for the syntax highlighting and then use it for any language right?
Cool stuff. This would solve many of my use cases of beamer if it had support for animations. E.g to click through each step of an algorithm for sorting a list in place.
Not sure this is a great answer anymore. O365/Powerpoint supports multi-user collaboration, change detection and automatic version control; however, no judgement on the overall efficacy.
Sync is poor and colleagues and I broke at least two presentations soon before preventing them so we all agreed to stop using PowerPoint and O365 this way. Now we use marp to great success see in another comment.
Is it bad that I look at every library now, no matter how awesome, and think "Well this didn't exist on or before September 2021, so do I really wanna learn this?"
I mean surely an interface of the form blah blah blah, "Now, take the content and turn it into slides using a document creation API" is the best right?
Should every library include a documentation endpoint for Scraper Plugin of the right context length? Is prompt injection an attack, or an opportunity? This is new ground for adoption.