Nostr Web Client

ai has been teaching me physics at a way faster pace than i was learning by myself ten years ago. Now i have self generated ai podcasts that go into advanced concepts on demand...

I got grok3 deep search to recommend the best source to learn N=4 super yang mills. It directed me to david tong’s lecture notes on super symmetry. So now i’m deep diving into that. Then i use ai to summarize the lecture notes.

I’m using chatgpt deep research to explore intersections of ideas i’ve read about, like twistor theory, double copy, and amplituhedron/positive grassmanians. It is able to find lots of interesting cutting edge papers that relate to each other.

Considering I’m an amateur physics enjoyer thats able explore the interplay between these complex topics has been mind blowing. I’m convinced we’re going to have an acceleration of physics breakthroughs once physicists start using these tools for research.

Reply to this note

Please Login to reply.

Discussion

HoloKat 9mo ago

This is also the most interesting aspect of this tech for me - being able to connect various ideas to seek out new insights. It’s great at looking at a ton of info to draw possible connections and explore further

Reid Walley 9mo ago

💪💪

Matthew 9mo ago

What is the best entry level podcast generator?

jb55 9mo ago 💬 2

notebooklm

ynniv 9mo ago

I was looking into gauge symmetry explanations of entanglement measurements with Claude a few months ago. It's way too easy to go deep with AI

https://arxiv.org/abs/2403.07935

jb55 9mo ago

Added to my ai podcast. Love these kinds of topics.

ynniv 9mo ago

For years QM has struck me as the kind of model you make when you just can't figure out what's really going on. I think AI is going to finally poke holes in it somehow

jb55 9mo ago

yeah its a bit frustrating that its something that is so mathematically defined but ontologically … 🤷‍♂️

its fun exploring unitary matrices and hilbert spaces… but what does it mean!?

ynniv 9mo ago 💬 2

I think we've confused the map for the territory. But I don't have a better theory either

jb55 9mo ago

yeah for sure, but the territory is out of sight so i can’t blame them. at least the string theorists and everettians are a bit more brave. Still not sure on either of those though.

ynniv 9mo ago

Testing things at the limit of "things" isn't easy. But then the Copenhagen zealots come out, take a nicely predictive model and pretend it's a religion, and I start learning QM

PolymathicPedagogue 9mo ago

Fascinating idea! To make the podcasts, are you using something like NotebookLM?

jb55 9mo ago

Im building my own that i can tweak and customize

nostr:nevent1qqsq6v5paheg6xvg2q6wc3v9jntg46u64xwnaa632vpcrfkxnskjfccpzpmhxue69uhkummnw3ezuamfdejszrthwden5te0dehhxtnvdakqzymhwvaz7tmjv4kxz7fwdf3r2dfwvdhk6qg4waehxw309aex2mrp0yhx6mmnw3ezuur4vgt67397

Scott ••• 9mo ago

I’d love to give this a listen, pay-to-listen sounds fine by me

John 9mo ago

Are you still using goose?

jb55 9mo ago

Cursor has replaced most of my goose use cases atm

John 9mo ago

Same, I gave it another shot today and prefer Cline over goose

jb55 9mo ago

The competition is insane. Most people have no idea what is happening right now

ynniv 9mo ago

Nothing stands still. I actually converted someone from Cursor to goose not too long ago. What's working better in Cursor and Line for you guys?

jb55 9mo ago

The tools seem to work better and the interactivity is more intuitive

ynniv 9mo ago

Getting the most out of goose does require a different mindset 🤔 But it's more important that you're using something than that it's goose. This wave is too important.

Stay buoyant 🤙

John 9mo ago

The edits in Cline are more visible than in goose. I'm cool with being in the loop since I'll be responsible for the code anyway

ynniv 9mo ago

Have you tried asking goose to write focused PR's?

jb55 9mo ago

goose is just cursor YOLO mode without change accept/deny. goose is better for vibe coding if you just want it to go off without much supervision.

I just like being able to select functions to add it to the context. With goose it seems to try to read the whole file to get the functions which floods the context. Maybe once LSP MCP tools are there this will be less of an issue.

ynniv 9mo ago

People love the term "vibe coding". I prefer "integration tests"

Braydon Fuller 9mo ago

Already happened! 200 million protein structures have been found that way! There is a Veritasium video the discusses it.

ynniv 9mo ago

One of my favorite AI games is to dig deep into conspiracy theories. The models are trained against talking about them but tend to be okay with evidence and reason, so there's this interesting tension. Claude in particular is really easy to nerd snipe. You can get it to talk about anything as long as it's "for the science", or philosophy

jb55 9mo ago

My favourite way to mine truth is to get it to roast things.

ynniv 9mo ago

It's interesting how well the models handle things like comedy and satire. These aren't simple concepts, and they complicate the idea of AI safety training

smallworlnd 9mo ago

Would have been a handy tool to have during my physics undergrad. But at the same time, things were just harder so I had to force myself that much more and it stuck with me.

calle 9mo ago

there are no shortcuts. concepts are easy, doing the math and understanding that is the hard part.

waxwing 9mo ago

It's really good at general concepts, even very abstract ones, but it can be disastrously bad at concrete calculation. Something to watch out for.

My best use so far has been language learning; I guess that is not surprising, lol.

jb55 9mo ago

code > thinking > one shot

Narcissist's Dollar Store Honeypot Unenjoyer 9mo ago

If physics can be accelerated by glorified auto-complete, it's a joke discipline of retards.

jb55 9mo ago 💬 2

A lot of science is just finding the gaps in what you already know

eliza 9mo ago

Oh yeah well I installed a new shower head yesterday by myself so take that smarty pants. 😂

That’s cool tho!

jb55 9mo ago

I usually keep the physics to myself, but it’s definitely my #1 interest outside of programming. It’s also the least communicable thing. I thought protocols were bad. I also need to clean my showerhead. Thanks for bringing be back down to reality .

eliza 9mo ago

Actually I thought the post was cool as hell and it gave me ideas of ways that I could do something similar with things that I’m into. That’s the type of post that breathes new life into my lungs. Thank you.

Globe99 9mo ago

I find it's great at summarizing a field, or giving a very generalized "look over there" kind of answer to top-level questions about scientific inquiry.

But I've found they (at least Google Gemini) can mess up sometimes... Like, citing something that actually shows the opposite of what they claim.

I think they're very incentivized to give "an answer", when there are important situations where the correct answer is "we don't know" or "your question doesn't make sense," etc.

N3WD3V 9mo ago

Too bad they don't want break throughs they want grants they want jobs! Hahahha

Captain's Nostr Log 9mo ago

I thought I knew something about physics but now I know that I know nothing

Achilles 9mo ago

AI is better than us. Humanoid robots will be better than us. Once humans break free from human centralized control with help of bitcoin how do we ensure we don’t get controlled by what’s better than us?

Bufcat 9mo ago

Test yourself with the Feynman Lectures.

jb55 9mo ago

I’m way too dumb to do the calculations. Closest I’ve done is like 100 of pages of quantum mechanics calculation drills when going through mermins quantum computer science book. Physics is easier when you constrain Hilbert space into some basis so you can calculate things at the quantum level. I never got into classical calcs or SR/GR. quantum mechanics, quantum information, and quantum computation has been my main interest. I learned all about the symmetries and representations theory while skipping einstein and newton 😅

maven leo 9mo ago

So much science, so wow 🤩