It strikes me that if Notion is a nice wrapper for a database, and the agent is being tasked with interfacing with that wrapper, why not skip the wrapper entirely? If they’re trying to offload most of your interaction with the application to an LLM agent, it seems like it doesn’t matter where the data lives. So why not use a Claude Code agent to do the same things for you locally?
I assume the alternative here is Obsidian.
It does have features that Obsidian doesn't have (like better URL preview). But it's mostly UI stuff. If you literally just want to manage some .md notes I'd say there isn't a reason to use Notion.
Notion seems have a lot of hype lately, and Microsoft tries to be king of the hill when it comes to productivity apps, buying tools encroaching upon their turf of Outlook and Excel (6 Wunderkinder, Yammer, Ally.io) or competing vigorously if they cant buy them (Teams). Seems like this Notion v3 could tip it over the edge into full blown productivity powerhouse.
Microsoft do already have their own Notion ripoff/inspired product (Loop) though. It is a bit half-assed and the development pace is glacial so perhaps a new team behind it would be something they'd be interested in.
Oh dear, more AI slop that's going to try to force itself on you, like a creepy uncle at a party. This "agents" thing seems to be a meaningless buzzword that every product must now use. I'd rather they focused on polishing the product or left it as it is, not contaminating it with trash that just gets in the way.
* they have decided to "have our own model" (always garbage model)
* they have decided to "collaborate with X" (and this model rapidly gets outdated)
What on god's green earth sort of a line is this?
It would've really helped if they worked on improving their subpar mobile apps, but instead they are focusing on AI features.
(Which, I don't see much incremental benefit in paying for separately, if I already pay for other AI subs like chatgpt).
How could the AI possibly know what I want to put in? The whole point of note taking and ordering and rearranging data is that I have the control over it. And by that a better understanding.
The base product was originally great: very smooth wysiwyg collaborative document editor with wiki-like linking. The problem is you don't need to do much on top of that. But clearly investors demand some "results" so PMs need to keep coming up with features that can be shipped in a quarter. Meanwhile bugs in the basic UX are plentiful.
Any really interesting work to improve the basic "collaborative document" experience is going to take time and experimentation, and I'm sure there's something to be found there. But the investor fueled focus on constantly doing something new and shiny means these really interesting spaces will never be explore and the product will continue to degrade with bloat each quarter.
I prefer one tool for one job approach.
Could someone please explain benefits of using one-does-all tools?
Most people got hooked before it was that convoluted.
Great reminder to export all Notion data to markdown and use a different tool.
saejox•2h ago
raincole•2h ago
pembrook•1h ago
matznerd•1h ago
After a few rounds of AI generating AI content from AI content, I'm sure it could eventually become slop...like the model collapse lol idk.
"AI models collapse when trained on recursively generated data" - https://www.nature.com/articles/s41586-024-07566-y
trinix912•32m ago
notrealyme123•2h ago
But instead more ai slop.
And thats not the worst. Every time when a company adds ai features, i know they want to train on my data sooner or later.
So hard pass for that one.
edit: seems like webhooks are here now. Will give them a try, but knowing notion, expect wild limitations
piokoch•39m ago
Now you are welcomed by a "AI chat" that wants me to specify what a great application I want to create. Total madness...