The complainers were FF users forced to deal with bloat they didn't use, those who are sad here are pocket users. They're just different people. Though, even those who didn't like the bundling of the extension probably didn't actively want the service to fail.
Right. I would be one of the people who saw pocket as an unnecessary distraction, but even I tested it and my opinion is partly based on pocket just not working in my Firefox at the time. I also just did not like that it was given space in the toolbar while a way more important rss button was denied that space. And despite that, I still think the shutdown now is bad - this should be spun out or be moved to a Foss project, and certainly not be killed for more ai nonsense.
BTW, fakespot (the service they also shut down) is or could be an applied ai project where that technology could be helpful, and they also shut it down. That also feels wrong, especially the combination.
In the sense I understand that comparison, or have usually seen it referred to, the compressed representation is the internal latent in a (V)AE. Still, I haven't seen many attempts at compression that would store the latent + a delta to form lossless compression, that an AI system could then maybe use natively at high performance. Or if I have... I have not understood them.
I don't find that "pessimistic" at all, though I don't think it's fair call the serious PKM crowd a cult. Different use cases and projects just have different requirements. The set up for someone like the author here who is doing historical research, will naturally require more organization, and thus naturally go through cycles of reorganization. I think this is similar to how engineering architectures sometimes cycle through idiomatic themes over the course of different refactors.
Though especially now with the potential of LLM-based retrieval I personally worry about keeping things well organized even less than it used to.
For clarity, I wasn't calling the PKM crowd a cult. My point is that after a certain level of engagement (i.e., consuming a ton of written and media-based information about the subject), I realized there was a certain cult-like following present. I'm unsure how much that following constitutes the overall PKM crowd, but I wouldn't say it's the entire crowd or even a large majority.
Like with many things, serious people tend to be a lot quieter (i.e., they are not running YouTube's content treadmill). So, the perception likely doesn't match the reality.
Congrats on finishing the novel! I also use obsidian for fiction and would be quite interested if you end up writing anything about the pipeline you mentioned.
I went through an unshipped app dev cycle that was fairly similar in many ways. Started on React, then RN, then Flutter, and eventually migrated away from graphql into sqlite.
My app had some similar aspirations -- to bridge the gap between habit motivation, goal adherence measurement, task scheduling & rescheduling. I worked on it for a few years, and my identity was very much wrapped up in eventually bootstrapping a company.
For me, the decision to let go of the project came in multiple phases, but one big closer was that I simply didn't want to be an app dev in the long run. While difficult to let go of, I currently feel good about the decision. Also, as evidenced by the resurrection part of the story, "nothing is ever fully lost" anyhow, though I doubt I'll ever return to this particular project.
One key idea I had for expanding beyond the "high cognitive load" nature of most productivity apps was to implement a "life module" marketplace of sorts that would let, say, a fitness influencer sell a workout routine + meal plan + journal template one could "install" into their life.
LLMs will also make detecting fall-off and attempting to attribute causes, or respond to "non-actions" much more feasible, which I think is important for anyone not type-A enough to use a productivity app consistently every day on their own.
Rebasing potentially allows for clean, clear documentation of changeset intent in a way that integrates easily with our existing tooling, at the expense of a potential footgun in highly collaborative scenarios. Ie, if someone forked off your branch or referenced a commit you might break a reference accidentally. So it can be nice, but really depends on the environment you're working in.
Gerrit solves the referencing issue by using git notes and maintaining its own ID for changesets across rebases. Without such a system, some seeming benefits of the rebase workflow, like atomically reviewable stacked commits, become fairly awkward. IE, if we want to review commits themselves, but their references get clobbered due to a rebase, that's no good (this was the case a few years ago with GH's rebasing support, maybe improved since).
IMO, that we have to make the "rebase or merge" tradeoff at all, or accept the deep limitations of pure commit-based history, is fairly sub-optimal. It'd be nice if more tooling/workflows were built more around notes. I envision someone/some bot going back and annotatinb a range of commits with a note that associates them into some coherent code documentation system, or amends some faulty assertion, links artifacts, etc. That way blame could take us to salient docs or surface behavior snapshot gifs or w/e. From directly within IDEs
This was my first thought as well but I don't think it matters and was likely more a result of using real training examples than limitations of the authors. The point of the work is that regardless of perturbations that degrade or complicate clarity, their model is able to extract and enact the motions that seem most implied that would be OOD with the other methods.
So this extends to barked under-contextualized commands like at the end with "Leaps forward then stands straight" but also these looser seemingly nonsense statements like "native motions" or w/e.
The big tradeoff here is that if it seems overly permissive. It would be very annoying to be talking of a third person and have your robot start dancing due to identity parsing issues.
It does all feel a little disorganized/wild-west-y compared to say, a .vimrc with a list of plugins and bindings, which is something that makes a system like Nix (or a fully containerized DE of some kind) appealing
I've been a SWE for >10 years & have been daily driving a Kubuntu desktop for more than a year now. I regularly run into basic issues day-to-day that require intensive troubleshooting, which would occasionally be catastrophic if I didn't know my way around a terminal. There are still minor issues/glitches I don't bother attempting to fix in preference of restarting.
I mostly feel these headaches are worthwhile in exchange for the freedom and other (mainly severely technical) perks I enjoy about the ecosystem. I can't speak for windows, but coming from macOS, calling the price paid for these perks today steep would still be an understatement to me.
