App rewrite vs value creation

What follows may in places sound bitter, but that is only because I used to like Paperpile. A lot. Please read it in the spirit of a formerly satisfied customer who would like to remain a customer, but may have little choice but leave.

I know many of us here have been waiting eagerly for many important new features. Items on the roadmap have repeatedly been pushed back from October to “a beta probably early next year”.

While I definitely understand the need to rearchitect a “legacy” application, this comes at the cost of zero additional value creation for customers. If I click on the “Announcements” button in Paperpile interface (top right), the last feature is from December 2021. “Settings > Feature previews” is also empty. It’s been a whole year in which paying customers have been waiting and waiting.

While the base app for organizing research papers is OK, it’s just that: OK. Paperpile was, several years ago at least, better than anything else on the market and way slicker. At this point, Paperpile is, for me, mostly a glorified BibTeX file generator with PDF downloader cum reader, for which all annotations remain stuck inside Paperpile, as there are no integrations or developer APIs through which such integrations could be built without the need for Paperpile’s team to spend resources on it.

In the meantime, I have adopted Logseq, which is vastly superior in linking highlights and notes from PDFs, although it does not grab bibliographic information as Paperpile does. As of yesterday, a private beta of Readwise Reader has become available, which blows the PDF experience of Paperpile out of the water as it comes with GPT3-based summarization out of the box and of course integration into Readwise, from where all notes would go to Logseq or wherever I need them. The ability to grab my research notes, combine them with everything else I have in Logseq, is crucial. As of now, I still use Paperpile to grab bibliographic information and papers, but I have already started moving PDFs directly into Logseq where I annotate them and link them up with my other notes. Why? Because I have better things to do than sit and wait. It’s not an ideal experience (with copying files around), but it is better than having everything locked in Paperpile with no way to get anything out of it without copy-pasting a thousand times, which is worse.

I have used Paperpile since the early days, but I am seriously concerned that feature development and, more importantly, new value creation for customers has ceased based on the lack of visible progress. That may or may not pick up again after the rewrite of the app is complete. I can only keep waiting. But while I do that, other apps will continue their development and perhaps take over what used to be Paperpile’s territory. If Logseq (without Zotero) or Readwise add the ability to grab bibliographic information from journals, store it, and generate a BibTeX file from it (or their developer communities), then I don’t really see any value in Paperpile any longer. Of course, they may not go in that direction at all.

This is only a single (loud) voice. I get that. I don’t expect anything to change because of that. I also don’t expect priorities to change because of that. I don’t expect a continuous stream of features I would love to see, as there are many people and use cases depending on Paperpile and their features may take precedence. That’s fine. That’s entirely expected. But the lack of visible progress is a major concern.

8 Likes

Very well said and you’re definitely not the only voice.

Part of what pushed me over the fence to migrate to Paperpile last month was all the potential I saw based on the promises and timelines, and everyone involved is frustrated about this I’m sure - us users and the devs.

That said I think Paperpile’s user experience still trumps all the other apps in this category, and that’s what initially got me hooked. It’s just very pleasant, intuitive and responsive. The major gaps are like you say, the limitations in viewing and organising annotations/notes and lack of integrations (imo).
(I’ve built a setup in Notion that gathers and sorts annotations and notes per paper, with relations to other databases to aggregate them per topic, concept, etc. A lot of copy-pasting, yes, but not massively more inconvenient than workflows in other apps (for now…). It seems Logseq does an even better job at this than Notion if it can also parse PDFs.)

I still have faith that we’re nearing the end of the progress drought but in the meantime…I should probably stop expecting more shiny features and just focus on doing the work :man_shrugging:

While I echo many of your concerns & have been considering moving back to Zotero for similar reasons, there are a few things that I disagree with.

One major point is that one of the things that originally attracted me to Paperpile is that much of the data is not locked up.

As an example, my library of documents in Paperpile is easily findable in a typical folder & file structure within my Google Drive account. Installing the Google Drive app & marking the Paperpile folder to “Keep Offline” makes it super easy to drag files into Obsidian, LogSeq, Notion etc without needing to make duplicate copies.

The auto-generated file names are long, but clearly delineate the key information of date, author & title.

Likewise, my highlights in Paperpile’s PDF viewer sync into the document itself, including colours and the like. That makes it relatively easy to extract into a PKM platform if desired.

There may be some issues with the linking structure if you work cross-platform, but there are usually workarounds for that.

The metadata & database of Paperpile is less accessible, but most of it seems to be present in a BiBTeX file generated across my entire library. Perhaps not ideal, but shouldn’t be too much of a nightmare to use elsewhere.

What I’m getting at is that Paperpile’s team have plus points for this ethos from day one, especially compared to some other popular tools I’ve ‘enjoyed’. Likewise, for a long time, they had the only functional workflow for adding papers to my library (with reasonable metadata!) while out & about via my phone. Zotero seems to have caught up on that front, although I haven’t tested it out.

For what it’s worth, I’ve been using Raindrop in the same way that Readwise Reader seems to work for several years. It’s low-cost & very functional, including website highlighting, storage of any link or file type, good apps for all platforms, and tons of cute icons to add to your collections - important for me as a visual person. It doesn’t integrate directly with LogSeq etc (yet) but for $35 a year, I find it very useful.

The whole GPT-3 summary thing being offered by serval tools is interesting but in testing various tools based on that concept, my field tends to throw the AI off to the point of useless or outright misleading summaries and conclusions.

I’ve been less active with research since completing my course so these problems have been less of a pain, but my Paperpile license is due to be renewed soon. I’m not sure I can see enough progress to justify the payment when other options have caught up in many ways, but would be happy to revisit later if these promised updates come to fruition.

3 Likes

Thanks for the comments. I will hold off a bit with my response. Paperpile will be 10 years in October of next year. It will be easier to explain our roadmap, strategy, and commitment on the basis of the new Paperpile which will make us ready for the next 10 years (including note-taking workflows).

We are down to a few bugs and the first beta should be imminent.

6 Likes

Hi Stefan, thanks for the update. Does imminent mean that it will be released before Christmas? Thanks.

My interpretation of imminent is some time during Q1, but that may become Q2. I don’t expect anything in 2022 tbh.

You’re not the only one.

I have started moving my workflow slowly to readcube papers. While it doesn’t bring about any external integrations with note-taking apps, it at least has a far better reading experience, and the built-in literature search and automatic article recommendations make for a significantly improved reading experience.

I do very much like your readwise/logseq workflow idea. And I think an intermediate solution might just be to capture citation details in Zotero, I guarantee someone will make a plugin that’ll allow for readwise integration, before paperpile finishes their promised new implementation which been promised for a few years now.

Honestly @stefan, it is becoming increasingly hard to believe the promised timelines…

The first version of the new Paperpile is already with the first private beta testers for more than four weeks.
It not only contains many new features but – as mentioned above – will also be the basis for any upcoming future directions like note taking which is high on our roadmap for this year.

2 Likes