rolandhvar

It's not just random internet noise. It's generative AI.

This is not low-quality digital output. This is value-driven generative AI.

Not automated. Generative.

There's poor and stupid, and then there's poor and smart

Probably not a popular opinion around here, but here goes anyway: the fundamental problem is that unless your net worth is $100m+, there is no point being a do-gooder as any attempts to fix big moral hazards are only going to harm you more than it's worth.

So why bother?

I agree. What is the purpose of this article?

I mean just look at UV.

Did that guy make it because Rust, and because he's passionate about that sort of thing? Probably.

But it's fucking fast. So did he sell out to OpenAI? Of course he did.

And thusly, both camps.

Yes I see, fair enough

So what happens when the datacenters need to upgrade (new hardware, or stupid enterprisey reasons like "must be new when replacing broken stuff")? Surely there remains a secondary market for the enthusiasts?

Sure, but, aim for the stars and you hit the moon right? Like fundamentally who cares? For the purpose of an AGI benchmark I'd argue you'd rather err on the side of being more intelligent and counting that as less intelligent than vice versa.

+1 on the UI thing.

I don't know what it is, but UI on Linux always feels too disjoint from the rest of the system.

It's a bit like how Windows 3.11 was just UI-on-DOS. I get the same feeling.

Don't get me wrong - I love Linux for all its CLI use but for some reason I've never been able to primary drive it without going insane after a week.

Windows just seems to feel more put-together and I guess that's because the kernel probably has hacks to support Office, and Explorer probably has hacks to support the kernel, etc.

The only other system I've felt this level of unity in is FreeBSD with its userland+kernel harmony.

Maybe I need to try a Linux desktop again as I haven't done it in ~10y but the other comment here about Fedora not feeling production ready doesn't inspire much hope...

Any ideas?

The thing that always irks me about c++ is this sort of thing:

> Explanation 1) Searches for the resource identified by h-char-sequence in implementation-defined manner.

Okay, so now I have to make assumptions that the implementation is reasonable, and won't go and "search" by asking an LLM or accidentally revealing my credit card details to a third party, right?

And even if the implementation _is_ reasonable the only way I know what "search" means in this context is by looking at an example, and the example says "it's basically a filename".

So now I think to myself: if I want to remain portable, I'll just write a python script to do a damn substitution to embed my file, which is guaranteed to work under _any_ implementation and I don't have to worry about it as soon as I have my source file.

Does anyone else feel this way or is it just me?

So here's the thing I struggle with. I do a lot of work in jupyter notebooks. I come up with a new model or approach to some problem, and I want to fork out and test a hypothesis in the background (which might be some set of hyperparameters, and might take several minutes, or hours; call it Run A) while continuing to work down some other path in the same notebook, and maybe kick off a Run B that explores some other change (like a restructure of the code that's not "compatible" with the hyperparameter search of Run A).

Then at some point when Run A finishes, I want to incorporate the changes I made in Run B and kick off Run C, and so on.

The hard/important things are:

1) Being able to do this while staying in a Jupyter notebook context the whole time. Even something as simple as multiprocessing sucks because I've found it's too hard to manage in a Jupyter context (e.g. how do you handle where stdout and stderr go?). It's easier if you move to scripts where you have full support for this sort of thing and you are expecting to look at multiple log files on disk and whatnot.

Also the sequential nature of notebooks doesn't help when you want to occasionally fork out or conditionally run stuff.

2) Keeping track of all these changes and hypotheses and merging the results/code together as you learn. It's like you need a VCS for your hypotheses. Maybe hydra & wandb help with that, I haven't used them. But this idea of keeping track of hypotheses seems like the more fundamental thing.

3) The main reason I prefer to stay in a notebook context is because I have all my objects easily accessible. My models, all my dataframes, functions to do some ad-hoc charting etc, all super easy to access in a REPL-like form. That is invaluable for doing ad-hoc sanity checks or digging/drilling down. So a big part of the workflow is you basically have this in-memory database of a bunch of relevant objects and you're querying it and constructing new objects & visualisations using Python as your tool, without having to load things from disk or build up the context from scratch. It's all "just there".

4) And then sometimes you want to take the results X1 of that notebook and plot them against some entirely different set of data X2 that requires a whole bunch of other code that you've defined in some other notebook somewhere, or maybe even as a real Python module. Like maybe that data lives in a database and you transform it or something. So OK, you call some functions to load X2 within your original notebook, but BOOM you get an OOM and you're like ok now I have to write some code to serialise X1 to disk, and make YET ANOTHER notebook so I can go analyze X1 and X2. It all just seems so... unnecessary, if only the right tooling existed.

My current best approach is to use semantic versioning on the filename, just copy the whole notebook each time I make a fundamental change, and try to keep track of my hypotheses, preconditions, learnings etc within comments and have a few of those on the go running, but it's often hard to engage in critical thinking when everything you know is sprawled across multiple notebooks.

Maybe a simple global journal is the only thing for this sort of use case. And that doesn't even address (4) which is often a huge pain point. Can anyone think of something better?

Dumb question, but how do you know it's smallest?

The ad provider has been changed to something less sketchy

This can be fixed by switching ads provider

This is a fair criticism for the current state of the site - but our goal is to expand this if the PoC gives useful results.

In our experience people are very sensitive about allowing their images to be used for certain generalisations of research like this; e.g., when it comes to guessing someone's ethnicity or gender.

Our goal is to allow questions to be answered for those generalizations and we are working on adding those to the website soon.

Caution: it's a little ad-heavy at the moment (sorry) but we need to sustain the hosting cost somehow.

The idea is to survey the Internet about what "attractive" means and see if some consensus is reached.

The statistical process for determining this will be made available soon!

As always, suggestions and criticism is welcome.