megaloblasto

This is great. Thanks for sharing. I'm always happy to make a donation for education.

Weird HN type question. On your site you mention paying $5 for hosting the site but didn't mention that here. We're you able to figure out a free way to host?

Figuring out what problem to solve is still hard.

Also bringing a product to market (a non-software product).

Also having good taste and cooking great food come to mind, but they are not really research problems.

In my understanding of the word intuition, it comes from a deep understanding of a concept that allows you to know why certain things happen with a "gut feeling", almost unconsciously. For example, I can intuit where my feet should move while jumping over river rocks, but I can't explain to you how to do it. I can only suggest that you spend a lot of time jumping on river rocks to develop the intuition.

I've spent a great deal of time studying llms and NN architectures in general. This allows me to intuit things about them and that intuitive understanding for my comes mainly from visualizations.

Are you able to visualize much about llms? I ask because if not, starting by learning to visualize high dimensional complex systems is what I would personally recommend.

Check out this video for a really nice visualization into NN architecture https://youtu.be/QgH9sr7G13Q?si=iTn1FiFsYCZCJKE_

Visualizing large dimensional matrices is a way of starting to develop intuition about llms. If you can't visualize high dimensional spaces, and you haven't studied llms, you're not going to have any intuition about how they work.

Depending on your knowledge of math, I recommend starting with linear algebra, building an understanding of the equations and try to visualize more and more complex systems, then study llms to see how you can apply your linear algebra intuition to your understanding of llms.

VTK is a great toolkit for visualizing complex systems. 3 blue one brown on YouTube has other visuals that might help you.

It takes time but it's possible. Good luck.

I think provable conclusions and building an intuition are different things to be fair. In my opinion it is entirely possible to build intuition about seemingly impossible to understand topics like infinitesimal areas, infinite limits, function mappings, and high dimensional spaces. For example, 3D spaces are easy to intuit if you can visualize vectors in your head. Some people claim it is impossible to understand dimensions higher than 3 but that's obviously not true because you could assign a color to each vector to visualize the next dimension. You could then assign a width to visualize a 5th dimension, and an arrow shape to signify the next, and so on and so on. That's more of a creative act that can start to build at least a visual understanding in you head of these higher dimension spaces. That's a way to start to intuit about those things.

Now with llms we need way more that 6 dimensions so we can start thinking of assigning matrices to each 3d point for example. That allows us to increase the dimension from 3 to 3 + whatever the matrix dimension is.

We can visualize the matrices instead of having numbers as having colors for each entry, so they can be a sort of cube with each vowel being a different color.

Now you can start to visually intuit about how these massively high dimensional spaces can be formed of these colored matrices that can react to some input training data.

That's a start of an idea for intuiting things that might seem impossible to have an intuition about. I think visualization is a great way to start.

I disagree. It's absolutely possible to develop an intuition about extremely complex mathematical ideas, including llms or high dimensional systems. Learning to build an llm is a great way to start building that intuition.

For what's its worth, any model that uses constitutional AI (like claude) does not use RLHF anymore

This is the thought process that I fail to see many people take (both here and other places). AI is changing many fields, ok what does that look like and how do we adapt? What can we do now that we couldn't before? What skills should I learn to adapt to this changing world? I think this is how we move forward. If you can get past the scary aspects of AI (not talking about datacenters or billionaires, that is a different subject) then you might be able to see how exciting this can be.

Whoa that's awesome. I didn't know California had one of those. Keeps animals, humans, and cars safe and looks beautiful too.

[flagged]

Personally I think the valuation is detached from the company itself. In theory Tesla's valuation is too high, but it doesn't seem to be coming down any time soon. Plus there seems to be ways to manipulate stock prices when you control such a highly valued company, to the extent that the stock price reflects actions taken in the stock market more than the underlying assets and balance sheets.

I strongly disagree. Just because compression is common in pop music (and perhaps overused in some genres) doesn't mean new music isn't innovative and dynamic. When I listen to music say from 1920 to 1950, it is so often so incredibly lame (not always). It's basic ideas and chord progressions and simple melodies with lyrics that don't say much.

Music is a way for people to express themselves and relate about how they see the world. People didn't stop doing that recently. In fact, I'd say people have been emboldened to say even more and push what music really means.

It sounds like you're saying that low will power leads to depression.

I hope this is a joke. Calvin's dad lying to him wasn't supposed to be idealized.

Looks like they missed Major League Cricket (MLC) founded in 2023

Everyone loves to compare AI with the dot com bubble. My question is, were there any policies put in place after the dot com bubble to mitigate a similar crash? Or did we learn nothing?

50 cent's intrinsic value doesn't decrease over time so to maintain that value, his name must go up.

How is thinking that a company might be doing PR work in a hacker news tread a conspiracy theory?

It takes the great minds at yale to figure these things out

Keebio LM is really nice. I've had one for about 3 months now and it's been great https://keeb.io/products/iris-lm-keyboard

It often seems that people think that the answers to the important questions must align with the subject they know best. Statisticians think life is all about probabilities. Economists think it's game theory. Computer scientists think it's computational.

"Everything can be understood through mathematics" is usually said by a mathematician.

HackRF One can go up to 6GHz ($400 new or $100 on alibaba for a similar device). Any higher frequency than that you'll be paying thousands.

They will continue to orbit for about 150 years, slowly falling towards earth until the drag from the atmosphere burns them up.

"Like many older satellites, the POES satellites do not have thrusters to support a controlled reentry into Earth’s atmosphere at the end of their mission life. Instead, once passivated, they are safely powered down, placed in a non-operational state, and left in a stable orbit. Without onboard propulsion or significant atmospheric drag at their current altitude, NOAA estimates they will remain in orbit for roughly 150 years before gradually reentering the atmosphere and disintegrating."[1]

[1] https://www.nesdis.noaa.gov/news/legacy-orbit-noaa-decommiss...

Sadly, you can't really get NOAA satellite images any more. NOAA-15 and 19 were decommissioned August 19, 2025, and NOAA 18 was decommissioned in June. It's my understanding that you'll need a much more powerful antenna to get images from the new satellites. Still, SDR is great fun. It's incredible to realize that all this information is stored in electromagnetic waves and passing through us all the time.

This is great info. Thanks.

Kind of. I don't use grapheneOS and I'd like to, but de-googling your phone by buying a Google phone seems a bit sketchy. I don't want to take away from a privacy focused project. I'm super thankful for this option and I can't stand android or iPhone. But in the back of my mind I wonder if I'm being tricked.

Someone clearly replied with the same link. I'm not sure how many people would keep replying the same thing without even doing a simple thread search.

Thanks

> These devices meet the stringent privacy and security standards and have substantial upstream and downstream hardening specific to the devices

It still seems strange. A big part of GrapheneOS is to provide a safeguard from Googles data hoarding, yet it works primarily on Google phones.

I've always found it strange that GrapheneOS only runs on Google hardware. Can anyone explain this choice?

I see what your saying but I would argue that vibe coding an ODE solver is an incorrect use of the tool. For something like and ODE solver you need to have a really solid understanding of what data structures you will use,and solid general knowledge of the numerical methods you want to implement. Then, you can use AI as an assistant when you get stuck, or to deepen your understanding, look over your implementation, etc.

It seems like developers used to always joke about how much they used stack exchange (even senior devs). Now it seems like there are suddenly so many people who claim to never need any help and can just smoothly bust out beautiful code all day long.