howrar

joined 2 years ago
MODERATOR OF
[–] [email protected] 1 points 9 hours ago

There's a few scenarios where I think it's worthwhile to get into these discussions online. Listed below in order of how much effort I'd put into it.

  • I don't know enough to hold an opinion or my currently opinion stands on shaky grounds, and I either want the other party to convince me of their position, or use the discussion to flesh out my thoughts and come to a more solid conclusion.
  • The stance I currently hold differs from the other party and I want to understand where we diverge. Do they know something I don't? Did they consider a variable that I didn't think about? Did they just start from a different set of "facts" and neither of us have the means of verifying which is correct?
  • I agree with their conclusion but disagree with how they reached it. The intent is to help the other party strengthen their position so that they can go off and preach the good word to everyone else more effectively.

The last scenario is the only one where I'm actually trying to change someone's mind. I do recognize that it's unlikely, which is also why I wouldn't put much effort into this. For everything else, the exercise of putting your thoughts into coherent words and thinking in new directions is where the value lies. Having the discussion with someone else forces you to consider many things that you wouldn't otherwise think about on your own.

Handling hate speech is tangential to promoting logical and well thought out discussions. I believe that this kind of community necessarily has to be homogenous in terms of values, otherwise there's nothing to discuss. If I want to maximize the number of oranges growing in the orchard and you want to maximize the number apples, then that's two conflicting goals. If we get into a discussion on how to manage the orchard, it won't go anywhere because at the end of the day, I don't care for apples and you don't care for oranges. There's no amount of logic that can convince your taste buds to change how they respond.

[–] [email protected] 3 points 11 hours ago

Do they also have a playlist for his cook time? I feel like it might be useful in the upcoming years.

[–] [email protected] 3 points 11 hours ago (1 children)

So that must be what Spotify's doing with all the data they're collecting.

[–] [email protected] 2 points 11 hours ago

The cook time varies a lot based on how dry the pasta is, which depends on how old they are, how they've been stored, and the air humidity where they're stored. It can vary as much as ±5min. Even on a conventional stovetop, it takes about 30s at most to come back to a boil after you drop in the pasta.

The best way to get perfectly cooked pasta is to regularly check on them.

[–] [email protected] 17 points 21 hours ago (2 children)

Whatever you decide to do, make sure to keep backups. You don't want to lose years of work because of a hard drive failure.

[–] [email protected] 13 points 2 days ago

I'm assuming you're talking about the US leadership? Eminem has been putting out music critical of every (2) Republican president since getting big.

[–] [email protected] 9 points 4 days ago

I've always gotten the impression that he's the only person in the entire administration that actually gives a shit about the American people. Problem is that he doesn't have enough brain left to not make things much worse.

[–] [email protected] 11 points 4 days ago (1 children)

That's no accident. A job is considered "unskilled" (or "unspecialized" as I like to call it) if any adult who's gone through the education system and is reasonably healthy can do. Since society would collapse without these jobs, we want to do everything we can to make sure we always have people who can do them. How do you make that happen? By designing the education system to teach everyone the skills to do them and making it mandatory to complete your schooling. As a result, nearly everyone is capable of doing some of the most important jobs for our society.

[–] [email protected] 2 points 4 days ago

I don't think the question should be whether or not it's needed, but rather whether it'll make things easier and encourage more people to make the switch.

[–] [email protected] 8 points 4 days ago (2 children)

In your heroin example, when they do decide to break the addiction, giving them a different drug (Methadone) is exactly what you do.

[–] [email protected] 4 points 6 days ago

Just because an English word was originally Latin and is written the same way, doesn't mean it's pronounced the same way. It's an English word now. It has an English pronunciation, pluralisation and definition that can all be different from the original. "Kentawur" is not correct for the English word.

[–] [email protected] 20 points 6 days ago (1 children)

I believe ROC is the "rest of Canada" in this context.

5
Open Sourcing π₀ (www.physicalintelligence.company)
 

https://bsky.app/profile/natolambert.bsky.social/post/3lh5jih226k2k

Anyone interested in learning about RLHF? This text isn't complete yet, but looks to be a pretty useful resource as is already.

 

Apparently we can register as a liberal to vote in the upcoming leadership race. What does it mean if I register? What do I gain (besides the aforementioned voting) and does it place any kind of restrictions on me (e.g. am I prevented from doing the same with a different party)?

 

An overview of RL published just a few days ago. 144 pages of goodies covering everything from basic RL theory to modern deep RL algorithms and various related niches.

This manuscript gives a big-picture, up-to-date overview of the field of (deep) reinforcement learning and sequential decision making, covering value-based RL, policy-gradient methods, model-based methods, and various other topics (including a very brief discussion of RL+LLMs).

 

If there's insufficient space around it, then it'll never spawn anything. This can be useful if you want to keep a specific spawner around for capture later but don't want too spend resources on killing the constant stream of biters.

10
submitted 5 months ago* (last edited 4 months ago) by [email protected] to c/[email protected]
 

I'm looking to get some smart light switches/dimmers (zigbee or matter if that's relevant), and one of the requirements for me is that if the switches aren't connected to the network, they would behave like regular dumb switches/dimmers. No one ever advertises anything except the "ideal" behaviour when it's connected with a hub and their proprietary app and everything, so I haven't been able to find any information on this.

So my question: is this the default behaviour for most switches? Are there any that don't do this? What should I look out for given this requirement?


Edit: Thanks for the responses. Considering that no one has experienced switches that didn't behave this way nor heard of any, I'm proceeding with the assumption that any switch should be fine. I got myself some TP Link Kasa KS220 dimmers and it works pretty well. Installation was tough due to its size. Took me about an hour of wrangling the wires so that it would fit in the box. Dimming also isn't as smooth as I'd like, but it works. I haven't had a chance to set it up with Home Assistant yet since the OS keeps breaking every time I run an update and I haven't had time to fix it after the last one. Hopefully it integrates smoothly when I do get to it.

 

This is a video about Jorn Trommelen's recent paper: https://pubmed.ncbi.nlm.nih.gov/38118410/

The gist of it is that they compared 25g protein meals vs 100g protein meals, and while you do use less of it for muscle protein synthesis at that quantity, it's a very minor difference. So the old adage still holds: Protein quantity is much more important than timing.

While we're at it, I'd also like to share an older but very comprehensive overview of protein intake by the same author: https://www.strongerbyscience.com/athlete-protein-intake/

 

Ten years ago, Dzmitry Bahdanau from Yoshua Bengio's group recognized a flaw in RNNs and the information bottleneck of a fixed length hidden state. They put out a paper introducing attention to rectify this issue. Not long after that, a group of researchers at Google found that you can just get rid of the RNN altogether and you still get great results with improved training performance, giving us the transformer architecture in their Attention Is All You Need paper. But transformers are expensive at inference time and scale poorly with increasing context length, unlike RNNs. Clearly, the solution is to just use RNNs. Two days ago, we got Were RNNs All We Needed?

 

Recordings for the RLC keynote talks have been released.

Keynote speakers:

  • David Silver
  • Doina Precup (Not recorded)
  • Peter Stone
  • Finale Doshi-Velez
  • Sergey Levine
  • Emma Brunskill
  • Andrew Barto
1
submitted 6 months ago* (last edited 6 months ago) by [email protected] to c/[email protected]
 

OpenAI just put out a blog post about a new model trained via RL (I'm assuming this isn't the usual RLHF) to perform chain of thought reasoning before giving the user its answer. As usual, there's very little detail about how this is accomplished so it's hard for me to get excited about it, but the rest of you might find this interesting.

view more: next ›