I think it’s fairer because you can keep using the app without upgrading to new features, but you have the option to pay for the upgrade (a quasi-subscription) if you want or need the new features. Some apps essentially stop working if you don’t renew. Ulysses is a case in point: if your subscription lapses, you can no longer create or edit content, though you can still export it. With the new DT model, as I see it, you ‘own’ the app, whereas with a ‘pure’ subscription, you’re essentially renting it. And if you stop paying rent, you’re ‘evicted’.
2 Likes
Agreed -and the DT model makes eminent sense from the perspective of fairnes and economics. The fundamental argument for a subscription model is to both motivate and fund ongoing maintenance and feature development of software.
Keep in mind though that if Ulysses were to adopt such a model then they would need to establish a higher “purchase” price to allow permanent use of a given version. The eviction between Ulysses vs DT4 falis a bit because Ulysses subscriptions are priced such that their users are essentially “renters” rather than “owners.” DT4 users pay more up front which merits their being treated like “owners”; the ongoing fees are not rent but rather maintenance costs.
I suspect few Ulysses users would be wiling to pay the initial cost that would be realistic for them to be “owners” rather than “renters.” So we have a bunch of renters who protest being evicted when they fail to pay the rent for a property someone else rightly owns.
2 Likes
I’m not able to make out if it’s
- Free App
- One time purchase/upgrade from version 2
- Subscription
Gotcha. I believe it’s saying you paid the discounted upgrade one-time price. That’s what I paid.
2 Likes
“## Lastly…
You should be diligent in your backups, especially local ones, in general and while testing our application.”
—-
Are there instructions for this in the Forums when using CloudKit for sync?
Sync is not a backup, neither advertised nor advocated as such. We strongly advocate proper primary backups using TimeMachine (or similar snapshot-style backup applications) and external drives. Also, you should read the series of articles we posted: Backups
1 Like
I’ve been running the DT public beta with a copy of a testing database. It’s been solid, and best of all, it’s not even really a transition. I know y’all spent some time really thinking about how to approach incorporating AI, and I appreciate this first iteration. It makes sense with DT.
No “real” feedback, though. Just some words I hope make the hard work a little more fulfilling. Congrats on the upcoming release.
4 Likes
Great THANKS…just what I was looking for!
1 Like
However you are still at the mercy of their license server being around. If that goes your copy of the software (DEVONthink here, but really any with a license server) can never be reinstalled. This has meant the end of several programs that I had used when the vendor has gone belly-up or just decided to no longer support a product. Yes an archival VM can save the day if you remember to create one and you have a way to run it.
Much appreciated and also what we were going for, having it make sense within DEVONthink.
Hey … that’s very helpful Jim. I’ve been playing around with the AI features and I like them in principle. I’m still feeling a bit of friction trying to navigate around DT. That’s something I know frequent users won’t ever experience again, but I am! I will persevere though 
I like that there is versioning. I think that’s new. It’s very, very helpful for this kind of work.
Thanks @rms, but TBH … I think going to the support site would add to my overwhelm and make things worse!
@DEVONtech_Jim’s comments were really helpful, for me. And maybe others.
1 Like
Thank you, I’m aware of the forums and their uses.
I posted here to see if anyone else in the MPU group who is playing with the beta may have encountered the same issue.
Well, not my experience. There is more than support going on. Some wise discussions. But given your stance, probably best that you give DEVONthink a miss as it won’t help your work process, as you describe.
2 Likes
I was amused by the new look of the Markdown editor and the “Typewriting mode”.
Also, gave a try to the new IA chat and summarization with a local Llama model and it is very promising. With all its existing metadata & content indexing capabilities, DEVONthink is in a good start to become the best “chat with your docs” tool for Mac users.
Instant upgrade!
3 Likes
I’m not sure what you mean by “stance”?
Perhaps you misunderstand: I’m already overwhelmed by DT; It’s a big powerful beast and very impressive, but it’s hard for me to figure it out; Jim was really helpful.
1 Like
In my brief read of this thread, it looks like DT4 new AI is more for a local chat with a document. I am looking forward to this development as well as natural language search. What I would love to see is more of a RAG (Retrieval-Augmented Generation) functionality, or something that works like Perplexity.ai and can answer my questions across multiple documents. Of course is probably asking a lot out of the current technology.
- What did you find amusing ?
- What model and parameters are you using?
In my brief read of this thread,
You should instead hop over to our forums for information directly from us.
more for a local chat with a document
Can you clarify what you mean by a local chat?
RAG (Retrieval-Augmented Generation) functionality
RAG is used as needed by the LLM in use, e.g., if a document exceeds the context window.
2 Likes
The look and feel of the Markdown editor seems more… not sure how to describe it… well, DT3 felt like a text editor with Markdown syntax highlighting and DT4 feels feels more like a more specialized Markdown editor. Which, giving it a second thought, perhaps the main visible change is the default font. But it feels more polished now and I like it.
Regarding the chat with the local model, I spun an ollama server with the first model I had already downloaded for another PoC, which was codegemma with a 8k token window. Even though it is fine-tuned for coding tasks, it did a good enough writeup of a group of files I had of an Ethics of Technology philosophy course. The bad piece is that local models through ollama are not exactly fast even in a 128GB M4 Max Mac Studio. Will have to investigate more with smaller models and compare it to OpenAI or Anthropic via API, but this doesn’t seem to be a limitation related to DT4, more with the state of the art of running local models on Macs.
4 Likes