jessetemp

How do you validate this kind of work to weed out any confabulating by the LLMs?

I kind of agree with you. I think music theory would be way more approachable if it was taught using intervals instead of all the weird naming of western notation. For example, everyone learns major and minor scales which are the interval sequences (in half steps) 2212221 and 2122122 respectively, but the names major minor don't really help you know other scales (excluding modes, maybe). If someone asks you to play hungarian minor, you'd first have to learn and memorize it. But instead, if you understand intervals and are asked to play 2131131, you immediately know how to play it. For me it also encourages experimentation, because there are obviously way more possible interval sequences to explore right?

The problem as others have pointed out is that most musicians in the west already know some degree of western notation, so if you're collaborating, you'll have to translate back to western notation at some point. Even if you invent the perfect notation, it's like asking everyone to switch to esperanto because english grammar is flawed. And you'll still get people defending english "well actually, it's like that because the greeks blah blah blah".

My favorite music notation flaw is C flat. It's a hack. It's an ugly fucking hack and anyone who defends it is defending an ugly hack. The only reason it and double flats exist is because there are some key signatures (this happens with hungarian minor sometimes) where you end up needing to define 3 notes in the span of one space and one line on the staff, and you can't, so you have to borrow from an adjacent space or line. And so sometimes that C is actually a B. It's super annoying but uncommon enough that it's not worth everyone learning a new notation.

Anyway, don't let the nay sayers stop you from learning music however makes the most sense to you. Have fun

I get it now. Had to play around with the code a bit to see it. Very interesting and unintuitive problem. Thanks for the thorough replies

You might be right. It’s been a while since I read snow crash. I just remember the metaverse as a sort sad state of society, but I don’t remember if the evil corp stuff was in there or just in the world at large

I had no idea it was a lord of the rings reference. Here’s an apt quote from wikipedia: “The [palantir] stones were an unreliable guide to action, since what was not shown could be more important than what was selectively presented.”

Makes me wonder what it’s like to identify with the villains in media. Zuck looking at the metaverse and thinking hey that’s a good idea! Or Thanos wasn’t so bad. Those rebel scum had it coming. Homelander is the good guy!

That would actually be a great feature. Opening two file browsers to move stuff around is a really common workflow. Although with current trends we might instead get a chat bot prompt “tell me how you feel about where you want your files to be”

Not quite. I'm saying there are 256 discrete numbers (0-255) and 255 intervals between those numbers. Most of the real values will fall into the intervals and get mapped to 0-255 somehow, maybe by nearest neighbor, but I'm not trying to define how they get mapped. The point is that 255 is the largest number that can be represented with 8 bits, so you should normalize by 255.

I wrote a longer replay to alterom but it looks buried for some reason.

https://news.ycombinator.com/item?id=48365800

Judging by your other comment in this thread, you might agree with my rational [1] more than you realize

[1] https://news.ycombinator.com/item?id=48365800

Thank you for the thoughtful reply. Maybe bins is the wrong word to use, so I'll try with intervals. Starting with 1 bit data, there are two numbers and one interval. I think where bins makes it confusing is that inside the interval there are two big rounding errors mapping everything to either 0 or 1 and many people seem to be considering those the bins.

Taking a step back, remember we're ultimately mapping these discrete numbers to some real world continuous variable like the saturation of red, frequency, mass on a scale, whatever. And our digital device can only represent a finite amount of numbers. For 2 bit data, we can represent 0-3, and for 3 bit data we can represent 0-7.

The important part is that 0 represents the minimum and 1,3, and 7 all represent the same maximum real value, and everything that can be measured by the device will fall within those ranges. So comparing 1, 2 and 3 bit data on a linear number line looks like this:

  0                    1
  0      1      2      3
  0  1  2  3  4  5  6  7

The question is about how to normalize the data. 1 bit data is already normalized. If you normalize 2 bit data by 3 you get [0, 1/3, 2/3, 1]. LGTM. If you normalize it by 4, you get [0, 1/4, 2/4, 3/4] and you're effectively throwing away some of the range of the ADC. You can try to get it back by offsetting by 0.5 then normalizing but now you get [1/8, 3/8, 5/8, 7/8]. And you could stretch that with some clever formula to fill from 0 to 1, but if you do it right then it's the equivalent to normalizing by 3, so why not normalize by 3?

So the answer is, if you have N bit data, you normalize by 2^N-1.

By counting the edges

The author is confusing bins with bin edges. In their first plot, the standard approach looks strange because 0-7 should be the bin edges, not the center points as shown in the plot.

You can see this confusion again in the histogram example. There are only 255 bins, not 256. If you fix that mistake and remove the 0.5 offset, then the histogram is distributed correctly at both ends.

Creating a throwaway for this comment is telling

Jeff Lowenfels has a series of books about fungi, microbes, bacteria, and nutrients that I really enjoyed. They’re all fairly short and accessible for someone without a bio background.

https://www.jefflowenfels.com/

It’s a craft like anything else. Some people enjoy building a table and feel a sense of accomplishment telling their friends “I built this.” Other people just want a table and buy one from Ikea

That explosion plot is pretty bad. No y axis. Unclear if it includes tuition and loans (which are paid or owed by students). Loans are ~50% of "ED appropriations", and only ~$21 million was distributed to students in 2021. But in the plot is looks like spending was around 150 billion (hard to say with no y axis) but ~50% was loans? And the source is just a vague Dept of Ed, with a link at the very bottom of the page to every single table published by the Dept of Ed, so have fun checking the source.

I'm not criticizing your take, although I suspect teachers might lose more than their lunch, just pointing out how terrible the plot is.

do you think that there are limits to who can or should use ai?

I don't think there should be imposed limits, but there might be an upper bound where expertise becomes atrophied by depending on AI too much.

if the plumber's use of ChatGPT improved outcomes, isn't that preferable?

In the short term sure, and maybe even in the long term for the customer. I think the risk to the plumber is losing some of their expertise by outsourcing to AI. But who knows, maybe the plumber has excellent memory and only accumulates knowledge each time they use AI.

Some of the article is lost in the plumber example. I doubt plumbers are spending much time exploring new ways of solving problems, and might even benefit from having a narrower range of outcomes. Other fields that require both expertise and novel solutions will be at a disadvantage if they become more homogenized by depending on AI. Not only is the range of solutions reduced, but getting there is faster, so people end up in a local maxima. Maybe they get stuck there, maybe not, but that's the risk I see.

You don't imagine any long term risks by outsourcing expertise to AI?

The plumber obviously. Not everyone needs to know how to be a plumber, but a plumber should know how to be a plumber

Higher cost doesn't always indicate negative environmental consequences. It could be (and seems likely to me) that harvesting one cow's worth of plant protein is more labor intensive which isn't necessarily bad for the environment. If you compare two soy crops, one that uses herbicides and another that uses manual labor to pull weeds, the latter will be more expensive and better for the environment

It doesn't need to be cheaper than the cheapest meat to be competitive. If there's some social or moral incentive to avoid real meat, that adds value to plant based alternatives.

Fungi protein sounds cool though. I would totally add that to my diet. But I also think insects are an underutilized protein source, so I might be an outlier

Laughter isn’t a perfect signal, but is the only signal in all the noise you mentioned

If two people disagree on whether something is funny, who's wrong? You can't say either of them is. There's no reward function for funny.

Laughter is the reward. N of 2 is a small sample size, but if one person laughed you could say it was 50% funny.

a really good joke is recent, relevant, and shows deep understanding of its subject

These can help, but it ultimately doesn't matter how recent, relevant, or deep a joke is. If no one laughs, it wasn't funny.

Ideally, hold Google liable until their AI doesn’t confabulate medical advice.

Realistically, sign a EULA waiving your rights because their AI confabulates medical advice

Only true if the alien plants are using the same chlorophyll as earth plants, which seems like a colossal fucking assumption to me

Sorry to ruin your ruining, but if you read past the abstract and look at the data, you'll see it tends to correlate with whether a democrat or republican is in office. Immigration policy is also mentioned in the discussion.

Given these findings, a corollary question is what attracts foreign graduate students to the US and leads them to stay. Prior research points to immigration policy—a subject of perennial public interest—having a large effect on stay rates

fewer safety guardrails in pretraining and system prompt modification that distort reality.

Really? Isn't Grok's whole schtick that it's Elon's personal altipedia?

I always wonder how much of HN is just bots debating bots

I didn't find 2 surprising either, but I'm a little surprised you never see it. If you want to treat the args to a function as immutable, what can you do besides copy, modify, and return a new object?

I hadn't heard of the author before this. I'll definitely read more of their stuff, but I thought the bottom line for part three was a little incomplete.

Bottom line: the more uncertainty, indeterminacy, ambiguity in your game, the more depth it will have.

Sure, starting from 0%, adding uncertainty adds depth. But the player needs to maintain some influence over that uncertainty. If you crank the uncertainty up too 100% then its pure random which isn't deep or fun.

I've noticed a similar more-is-better trend in a few sequels I've played, where the first game had say 5 mechanics which were fun. Then the sequel has 10 mechanics, and because 10 is more than 5 it therefore must be more fun. But it ends up being too much shit to juggle and less fun as a result.

More isn't always better

The title contradicts the article. Our eyes are averaging information that is up to 15 seconds old, but we can obviously react to visual stimuli in less than 15 seconds so it’s not all that old. They’re probably also confounding the research by using faces as the thing that changes over time, since our brains are sort of hardwired to respond to faces differently than other stuff