upbeat_general

For the longest time I’ve wanted a way to record my location history in a durable way, without resorting to google maps history (which is great but has…obvious downsides), or some hacky short-term custom solution.

I wonder if someone could integrate this into a more coherent long-term platform.

I think a lot of these comments don’t properly capture the benefit. The more skilled workers, the more startups/companies, not to mention smarter people.

But the effect is bigger than that, by allowing skilled immigration, it makes US universities and tech companies the best in the world, at the very least seen as such, which has tremendous larger effects.

It’s not a coincidence that we have the largest tech industry, and it’s not because we magically have smarter people.

Maybe I'm missing something but are there any current regulations that make it difficult to operate?

My understanding was that in many states, it's mostly just paper requirements and not any strict testing, etc.

scaled_dot_product_attention isn’t CUDA specific, it even works on TPUs.

Honestly most of the things I would like food are very oriented toward research/sharing use-cases, which I realize that is distinct from a lot of your users, but can mean:

- Working with large remote datasets/clusters - Needing to share visualizations with others [both interactive, and pre-recorded/controlled] - Being able to easy switch between N runs to look at data/results over time

Concretely:

- Currently it seems you need to forward multiple ports, and for each new session you need to re-connect. My memory is foggy but I found it was a lot of hassle to just connect a thin-desktop client to a remote backend.

- The support for baked-html is great, but it has limits (namely I was hitting some size limitations) and it would be cool to have a server that can read rr files from disk and stream them on-demand [e.g., point to a directory and let the user select]

- Generally speaking, I found the workflow difficult to integrate to the typical ML setup. Obviously not a priority for rerun but I think a little improvement [like the point above] could help a lot.

- A built-in screen recording feature would be great. I spent a lot of time screen recording and re-sharing clips.

Again, these are mainly nitpicks, love the tool overall and have recommended it to several people.

I’ve been using this for visualizing some computer vision data/model prediction and it has been great.

The remote client/server streaming has a bit to go, but just the option to use it in a browser is a game-changer. All the previous web-based tools that I’ve come across have been either very limited, or require significant manual setup.

Down for me as well. Thought my SSH agent was broken.

Well CPython has a lot of backwards compatibility to deal with that RustPython doesn't, so "import subprocess" might result in very different behavior.

Does this have faster startup times than cpython?

Every time I want to rewrite a shell function in python, I always hesitate due to the slow startup.

Haven’t read the paper (or looked at the code), but not sure there’s a general obligation to cite anything that’s not a pre-print/published. If they knew of and used the code, that would be a different story, but otherwise it seems okay, if understandably frustrating for the person that wrote the code.

This is especially the case for something like this where it’s a relatively obvious idea and the evaluation matters more.

Adobe can either (A) charge the full year up front, or (B) specify the full year cost and give customers the option to pay that monthly (like the buy now, pay later plans).

The issue is that their current wording is deceptive.

I thought it was pretty well used pre-covid, especially when the rides were sometimes 50% of a regular ride.

At least I personally used it a lot, and knew several people that did.

The implementation seems to be pretty clean and modular here where transformers (and diffusers) isn’t, unless you take their modules standalone.

This repo has a lot of handy utilities but also a bunch of clean implementations of common models, metrics, etc.

In other words, this is more for writing new models rather than inference.

As someone who works on both, this is just not the case. If you're talking about general automotive R&D, sure.

If you're talking about the total employee makeup of a current self-driving company, the number might be ~10-50%, depending on the company.

But if you take the subset of employees that work on software, this number is much much larger. There's lots of typical SW dev needed for GenAI (think data pipelines) but a good chunk of ML-folks doing self-driving work can be useful on GenAI as well.

I think you misunderstood the NLRB article entirely. The change is that in cases of election violations, the default remedy is now forced bargaining. There don’t appear to be any open elections, only the increased possibility that a union is formed by default.

I discovered this yesterday and spent some time with it. Sadly, it’s a pretty thin wrapper, with no typing or IDE support. It’s more or less ‘evall()’ with helpful type conversion between languages.

I wonder how many people actually use many of the newer features. It seems there’s always new data sources and niche features, but I imagine it’s not deep enough if you work in the field [e.g., chemistry or astronomy].

For me it’s such a shame as I would definitely use it much more often if the language and environment were closer to a typical coding workflow. It’s just such a hassle to have to use a standalone desktop engine, a weird editor, and a quirky language [at least at first] with few 3rd party libraries.

I would imagine any military “GPS Jammer” would also block similar systems…

This is such great news, I've been annoyed with slow arXiv loads for a while and happy to hopefully see it fixed!

I have the same issue. From what I can tell it’s just network-bound and the Arxiv servers are slow. They theoretically allow for you to setup a caching server but after spending a while trying to get it setup, I haven’t been able to get it to work.

https://info.arxiv.org/help/faq/cache.html

Aren’t basically all modern cars steer by wire? I guess ethernet vs CAN is notable but isn’t that a pretty minor implementation detail?

In my opinion the root cause is pretty clear: they had a network protocol update that was not backwards compatible and didn’t verify the runtime versions or have versioning.

Everything else isn’t really core to the issue [even that the update script silently failed].

Yeah this is the real reason right here.

Updating your code: Easy. Updating 10 different libraries that you depend on but don't know anything about: Not so easy

Personally it’s not one specific function but the answer I’m looking for [e.g, code that just works] is significantly better from GPT-4 imo. The better the output is, the less time I spend and the more likely I am to default to using the model in the first place.

Don’t worry about the price. It’s absolute crap. [At least PyCharm remote is and I believe it has the same underlying Jetbrains architecture]

Windows randomly quit, take forever to open, there are always indexing issues, etc. Locally everything is great, but remote is completely unstable.

I had very similar issues culminating in losing a non-insignificant amount of data. Despite being locally stored on multiple computers, an issue with iCloud Drive caused it to be deleted off of all.

I’ve since moved all my data out of iCloud Drive (although I still use it for their other services like photos). I thought my case was just special but I’ve encountered similar issues helping friends/family.

It’s beyond me how they get away with people thinking it “just works” and how they can’t manage to make a simple cloud sync service.

This article doesn’t address the real challenge [in my mind].

Framework support is one thing, but what about the million standalone CUDA kernels that have been written, especially common in research. Nobody wants to spend time re-writing/porting those, especially when they probably don’t understand the low-level details in the first place.

Not to mention, what is the plan for comprehensive framework support? I’ve experienced the pain of porting models to different hardware architectures where various ops are unsupported. Is it realistic to get full coverage of e.g., PyTorch?

Not sure about the current situation but since Apple controls literally the entire stack, I can’t imagine it’s a big issue.

If they really need a new op, they add it.

Lots of ML applications don't generally care about latency, and throughput only matters per dollar (as you can just scale horizontally).

To my knowledge, most inferencing (at least for simpler models) happens on cheaper, slower GPUs that have better throughput/dollar.

Are they price-competitive for those applications though? My understanding is that the main focus is Memory/GPU-interconnect for large models.

Are recommendation/ranking models large enough to take advantage of this? I don't think these cards are generally competitive in throughput/$.