And 1967 was 58 years ago, which was when the first deep neural network was trained with stochastic gradient descent. Yet, DNNs didn't take off until the 2010s when the hardware became powerful enough and data became plenty enough to successfully train and utilize them such that they were practical.
> you cannot use them to solve real world problems
Doesn't waymo and other self-driving systems use reinforcement learning? I thought it was used in robotics as well (i.e., bipedal, quadrupedal movement).
I just want to give a hearty thank you for pixi. It's been an absolute godsend for us. I can't express how much of a headache it was to deal with conda environments with student coursework and research projects in ML, especially when they leave and another student builds upon their work. There was no telling if the environment.yml in a student's repo was actually up to date or not, and most often didn't include actual version constraints for dependencies. We also provide an HPC cluster for students, which brings along its own set of headaches.
Now, I just give students a pixi.toml and pixi.lock, and a few commands in the README to get them started. It'll even prevent students from running their projects, adding packages, or installing environments when working on our cluster unless they're on a node with GPUs. My inbox used to be flooded with questions from students asking why packages weren't installing or why their code was failing with errors about CUDA, and more often than not, it was because they didn't allocate any GPUs to their HPC job.
And, as an added bonus, it lets me install tools that I use often with the global install command without needing to inundate our HPC IT group with requests.
> The main thing conda doesn't seem to have which uv has is all the "project management" stuff.
Pixi[1] is an alternative conda package manager (as in it still uses conda repositories; conda-forge by default) that bridges this gap. It even uses uv for PyPI packages if you can't find what you need in conda repositories.
I've been using Plex (connecting via Tailscale) with their Plexamp music player.
It's been working pretty well, but I might have to give this a try to compare. Although, it's not clear from the GitHub README or the Apple App Store listing if the mobile app allows you to download music for offline listening.
Also using Plex and Plexamp, and very happy with that combo. Curious about why talescale is needed - I'm on a static IP, but I believe Plex also provides a forwarding service (?)
I think you were talking about Blackcandy in the second paragraph, but just to be clear, Plexamp does allow downloading for offline listening.
It's free, extremely easy (not that port forwarding is complicated) and you don't need to port forward.
I point DNS records on my personal domain to tailscale IPs so it some subdomains can only be accessed when connected to tailscale, I can do app.mydomain.com etc without exposing anything online.
- Cloudflare tunnel for public access
- Tailscale for private use and sharing over WebDAV
- Nextcloud for general file management
- Jellyfin for music and video streaming
Nextcloud's WebDAV has issues with filenames or at least how it works. A large amount of files in non-'standard' characters wouldn't show up, so Ampache/Subsonic wouldn't work. This is why I tried Jellyfin.
> What is the primary operating system in which you work?
It doesn't ask anything about preference and I wager most people don't have a say in OS for their jobs. Same reason that Microsoft Teams is the most "popular" synchronous tool, yet so low in the "admired" section.
That's not to say you're wrong, just that the data you linked to does not "back it up".
There’s a ton of other questions, such as about visual studio (on windows, where else), loved stacks (.Net leads a category…likely most of that is dev on windows). Look at C#, Azure, ASP.NET, ASP.NET Core, Teams over slack and all in that category, and on and on.
And calling teams low in the desired section is such an odd assessment. It’s 3rd out about 2 dozen, and you ignore other highly desired items like .Net being #1 in its category, visual studio 4th of about 2 dozen, and others.
So yes, the data backs up that a lot of devs are on windows and that a lot want to use tech that windows devs use. That you so misrepresent one category and ignore others is not a very good way to understand the evidence, which far outweighs any one persons opinion in this thread.
> And calling teams low in the desired section is such an odd assessment.
I said it ranks low in "admired" (i.e., those who use it and want to continue using it). Less than 50% of those using it want to continue using it despite it being the most "popular" by a significant percentage. If everyone had their choice, Teams would drop heavily in popularity.
Regardless, it was just an example of how "popularity" doesn't mean anything because most people don't have a choice in their day jobs.
The question isn’t wether they want to use windows but if they’re productive using it. Unless you want to claim that a large proportion of devs isn’t productive, having a proportion of devs use windows proves that it works.
> It seemed like "productive" was used in a relative sense. Would all those people be more productive on macOS or Linux? That's not clear.
I would assume that companies are semi-rational actors and would switch if they could improve productivity that way. Especially since some sectors (graphic design for example) seem to prefer MacOS while others don’t. Of course there are some other factors (support, network effect, purchase cost) but if windows was just plainly unproductive, surely it wouldn’t be as popular as it is.
> Also, I thought the parent was replying to the following part, considering they said "I know plenty".
True, the survey doesn’t prove they actually prefer windows, u missed that context.
Indeed, i have never worked for a single company bigger than about 10 people that I could call 'semi-rational' when it comes to maximizing their worker productivity vs costs.
> Nyxt differs fundamentally in its philosophy- rather than exposing a set of parameters for customization, Nyxt allows you to customize all functionality. Every single class, method, and function is overwritable and reconfigurable. You'll find that you are able to engineer Nyxt's behavior to suit almost any workflow.
