It definitely is solvable though. Data versioning is a thing and it can work quite transparently to the mutations done on the data.
To not know who made and who approved a set of mutations on data can easily become equally as mind-blowingly stupid as not knowing who made mutations to code. Code is a subset of data after all and search over (provenance of) data can be implemented as DAG traversal.
Not tracking data changesets like code changesets is certainly a choice, not really a constraint anymore. A similar choice I feel is implied by "extreme scale of text searching".
> no they will not all add the telemetry you wish they did
...is just a failure of the corporate policy surrounding data handling. Is git-for-data already considered telemetry?
Of course the truth is provenance of data is something best institutionally forgotten as quickly as possible. The only thing that matters is it's there, that the data has no history, and that's why it can be used in whatever way deemed necessary.
All of your points are valid, and believe me I was trying to make them. The problem is one of culture. Most of the people doing this kind of work didn't like version control, and their work was really just running notebooks (like iPython or Google Colab) until a number was good enough and they'd submit the file for inclusion into training runs.
You can call it a policy failure, but these people were in very high talent demand and so top down dictates would risk "X people leaving lab Y for lab Z" headlines and morale hits.
I am not saying this is good. I am telling you that on the ground it is so much messier than it should be.
Yeah, vanilla MacOS is super focused on the set of "all windows of a given Application" as the useful "unit of work". Whereas in real work, you actually handle individual windows of a set of applications (e.g. firefox window for docs, vscode for code, finder window for project files) as the single unit you are working on.
MacOS is, in my experience, incredibly clunky due to this mismatch. You can fix some (e.g. use AltTab for sane alttabbing), but not others (how can i prevent the next application window popping up when minimizing one)
It is indeed very strange and annoying. And I don't remember it being so bad ten years ago (although I had to use witcher for alt-tabbing - which actually worked well, unlike AltTab now). Since Linux, and both OG Mac Os, I am used to having multiple Desktops, in total multiple windows of the same applications (browsers, pdf reader, office, file browser). Mac Os is completely hostile towards doing multiple tasks at the same time, and compartmentalizing them into desktops -- you keep getting thrown around to random windows on random desktops (example if you close a window), it's super difficult to find your windows, it's just crazy. You're afraid of ever moving away from a window.
I think this paradim is also what's makes the Mac OS feel snappy. If basically having multiple windows for one software open is "holding it wrong", the OS is pushing you into one application, one window paradigm. As long as you do that, everything is pretty quick. But don't dare open two browser windows, or even worse, two windows of a word processor. The performance will degrade (and you'll keep searching for your windows...)
I don't know, while I miss the super+right click drag to resize window and super+left click drag to move window from KDE on my work machine, I also kinda miss the smoothness of macOS's handling of many windows and desktops utilizing the trackpad on my gaming desktop. The swipe up and down/ctrl+arrow up and down to juggle windows within the desktop, and then left/right to move between desktops even makes their at first awkward handling of fullscreening apps feel right and efficient. What I really miss in KDE though is multiple virtual desktops per display I can switch independently per display. Actually makes me enjoy using multiple monitors slightly slightly more on macOS. What I sadly can't see happening is a merging of KDE, macOS, and the superior i3, as there would be too few few safe shortcut combinations available for programs to rely on.
vanilla MacOS is super focused on the set of "all windows of a given Application" as the useful "unit of work". Whereas in real work, you actually handle individual windows of a set of applications (e.g. firefox window for docs, vscode for code, finder window for project files) as the single unit you are working on.
Thank you for articulating this so well.
When I bought a new M1 MacBook Air for my kid, I hadn’t used OSX in years, and was excited to try a super-polished Unix. Holy heck I couldn’t believe how work hostile it was. I want apps tiled so I can quickly jump betwee - not to go to the top of the screen to get bottom tile’s menu.
And the state of finder (other than preview) is something out of Redhat 5.1.
And the response is always “you’re holding it wrong.”
My knowledge is a few years outdated by now, but I remember digging into this and realizing that most of the chinese open-source libs were license-washing software. E.g. PaddleOCR is licensed under Apache 2.0, a very permissive license, however its models were often-times built on/fine-tunes of less permissively licensed foundation models such as Microsoft's LayoutXLM (Creative Commons Attribution Non Commercial Share Alike 4.0). (Which in my laymans understanding is also a kind of viral license in that changes need to be shared back under a similar license?)
The link is annoying enough to find that I can imagine "Mea Culpa" being an effective enough strategy for businesses moving into the ML/AI field, changing their tune after they get caught, but matured their own software to stand on its own feet.
Yes, but expect those prison sentences will be more popular than you might expect. The current admin expends an insane amount of energy pumping the size of their supposed fanbase.
Yep. In particular, ARC-AGI-3 is a series of games where if you fail, you keep trying again (until eventually hitting a timeout). So the sooner you succeed, the sooner you stop spending tokens retrying. If it was a benchmark where everyone got one attempt with no retries, you wouldn't see it bend backward.
This default-to-auto-mode and the misleading marketing is begging for a class action once damages accumulate. Especially considering the Auto Mode even can actively prevent the clean-up!
With every newly released open weight model, the clock on the issues you describe is reset. I can see a marketplace arising for paid updates to common lines of open weight models, which will incentivize those with the hardware to train to fix the problem for those who only have the hardware for inference.
I would say when this comes to pass, we are already 5 years along?
> Part of it is knowing that whatever sort of enshittification the cloud providers do, my local programming environment won’t ever be less effective than it is today locally.
I think this is quite understated. It basically is freedom from a growingly antagonistic relationship between you and some remotely hosted API managed by faceless corporates at the whims of their board, shareholders and governments.. It really is such a mental burden to need to constantly manage this relationship (watermarks, silent downgrades, random false refusals, downtimes, model sunsets, changing ToS's, fucking ads). These companies will need to squeeze you for every cent that they can before open-weight models are simply good enough for the valuable tasks we can throw at them.
To have your own hardware is to no longer have this mental burden.
Not sure. It didn’t do “cat” so I guess that the file itself did’t leave the computer? Uploading all files in my ~/ would have taken noticible time and didn’t happen I think.
> You will need at least 75 GB of RAM or unified memory to run the model. Its smallest 1-bit quantized version is larger than usual because of the model’s architecture so 1-bit isn't really 1-bit at all. However, this also means the quantization is less aggressive, allowing the model to retain more of its original accuracy than more heavily quantized models.
Lots of RAM required even for the 1-bit, which is already downloadable. Interested to see how well this one works compared to Ornith1.5-35B-A3B I've been running (and quite happy about).
To not know who made and who approved a set of mutations on data can easily become equally as mind-blowingly stupid as not knowing who made mutations to code. Code is a subset of data after all and search over (provenance of) data can be implemented as DAG traversal.
Not tracking data changesets like code changesets is certainly a choice, not really a constraint anymore. A similar choice I feel is implied by "extreme scale of text searching".
> no they will not all add the telemetry you wish they did
...is just a failure of the corporate policy surrounding data handling. Is git-for-data already considered telemetry?
Of course the truth is provenance of data is something best institutionally forgotten as quickly as possible. The only thing that matters is it's there, that the data has no history, and that's why it can be used in whatever way deemed necessary.
reply