News for hackers. Stuff that matters.
Re:Opus 5.5 agents discover two room-temperature magnetic se... by nico on Monday October 05, 2026 @05:28PM (#49971067)
In a way, you can think of pretty much anything we express with language, especially things that are already modeled in scientific language, or logical language, or in equations or code; to be representable in a parametric/searchable space

Thus, you can build ai/ml models+agents to explore those spaces, at a speed and scope much larger than what any human can do

I can imagine findings like these are going to keep increasing in frequency to a point in which the bar for novelty goes a lot higher

Re:Opus 5.5 agents discover two room-temperature magnetic se... by nico on Monday October 05, 2026 @05:38PM (#49971190)
Anecdata: over the weekend, on a whim, I decided to download a real fly’s brain’s weights [0], run it on a simulated task like finding food, then train a logistic classifier using the fly’s decisions as the expert, then use the trained classifier as a decision model to simulate the fly on a 3d environment, running in real time on a website

It took me (using Claude code and some codex), about 3 hours to put it together

And even though it was a cool demo, it seemed so easy, that it also felt like it wasn’t worth sharing

0: ChessFly (not mine), uses the FlyWire connectome (the fly’s brain’s weights) to play chess https://huggingface.co/spaces/mlabonne/chessfly

Re:Opus 5.5 agents discover two room-temperature magnetic se... by binsquare on Monday October 05, 2026 @05:47PM (#49971311)
I think ai certainly raises the bar for those with taste
Re:Opus 5.5 agents discover two room-temperature magnetic se... by gradus_ad on Monday October 05, 2026 @09:06PM (#49972900)
It's like a mass proliferation of fast casual but damnit I want some steak.
Re:Opus 5.5 agents discover two room-temperature magnetic se... by tripleee on Monday October 05, 2026 @05:47PM (#49971314)
Many people wouldn't find that easy, even with AI
Re:Opus 5.5 agents discover two room-temperature magnetic se... by chneu on Monday October 05, 2026 @06:38PM (#49971784)
This can be said about anything.
Re:Opus 5.5 agents discover two room-temperature magnetic se... by caaqil on Monday October 05, 2026 @06:53PM (#49971910)
> This can be said about anything.

More generally, anything can be said about anything.

Re:Opus 5.5 agents discover two room-temperature magnetic se... by parineum on Monday October 05, 2026 @08:37PM (#49972691)
In a way, you can think of pretty much anything we express with language, especially things that are already modeled in scientific language, or logical language, or in equations or code; to be representable in a parametric/searchable space
Re:Opus 5.5 agents discover two room-temperature magnetic se... by anthonyrstevens on Tuesday October 06, 2026 @10:59AM (#49979504)
Just think of all the gaps in that space and ponder how we get from known to unknown
Re:Opus 5.5 agents discover two room-temperature magnetic se... by Towaway69 on Tuesday October 06, 2026 @03:39AM (#49975407)
Or anything can be said about everything and everything can be said about anything.

In the end, nothing can be said about many things and many things say nothing. Those things are then left to interpretation.

/s

Re:Opus 5.5 agents discover two room-temperature magnetic se... by polishdude20 on Monday October 05, 2026 @05:40PM (#49971228)
Oh please do share!
Re:Opus 5.5 agents discover two room-temperature magnetic se... by nico on Monday October 05, 2026 @05:50PM (#49971348)
https://playground.jeffyclassify.com/#fly

It's a small machine, so it might get bogged down

Re:Opus 5.5 agents discover two room-temperature magnetic se... by RivieraKid on Monday October 05, 2026 @08:48PM (#49972775)
The weights in the fly's brain are unknown, the connectome doesn't contain such data. Not sure what exactly these demos do, but it's certainly not a simulation of the fly's brain.
Re:Opus 5.5 agents discover two room-temperature magnetic se... by nico on Monday October 05, 2026 @10:28PM (#49973455)
You are technically correct. The weights of the connectome are not the same as the parameters of an ai model, but FlyWire absolutely does provide a weighted directed connectivity graph, where the edge weight is the number of synapses between neurons. The FlyWire literature itself calls those values “edge weights” or “connection weights”, and treats them as a proxy for synaptic strength

And agreed that this isn’t a faithful simulation of a fly brain. I’m using the connectome as the network structure/parameters for a computational model, then using its outputs as the teacher for the classifier. Not sure how the ChessFly uses it

Edit: in any case, these are just fun demos, they aren’t research papers trying to claim accurate physiological fly brain software simulations

Re:Opus 5.5 agents discover two room-temperature magnetic se... by cousinbryce on Monday October 05, 2026 @10:47PM (#49973580)
This is extremely impressive, but at the same time it has some pickle Rick plays chess vibes
Re:Opus 5.5 agents discover two room-temperature magnetic se... by gspr on Tuesday October 06, 2026 @04:48AM (#49975914)
> You are technically correct.

That sounds like you're saying that they got you on a minor technicality only. That's not the case; they're right, you're wrong – you did not "download a real fly's brain's weights".

Re:Opus 5.5 agents discover two room-temperature magnetic se... by KingMob on Tuesday October 06, 2026 @06:58AM (#49976740)
They're a proxy yeah, but it's also somewhat loose.

There's a lot more going on at synaptic clefts, so just knowing the number of synapses doesn't tell you enough to model anything. You still need to know which neurotransmitters are used, how much, second-order effects like G proteins, basal firing rates, distance to the axon hillock, the shape of the neuron's effect on potential decay, etc.

And that's all to predict whether one neuron will fire. You could get very different behavior between different two neuronal pairs having the same synaptic count.

Re:Opus 5.5 agents discover two room-temperature magnetic se... by jacquesm on Wednesday October 07, 2026 @07:32PM (#50000199)
It's a bit like having a metal layer of a chip and nothing else.
Re:Opus 5.5 agents discover two room-temperature magnetic se... by SkyBelow on Tuesday October 06, 2026 @08:46AM (#49977591)
I wonder if this should be considered some super-quantized version of the fly brain. Sub 1 bit level. Think like taking a 400GB model and reducing it to 400MB level of quantization. In a standard NN, this isn't really possible as it changes the shape of the weight rather than just the information for weight, so it isn't called quantization anymore. But with actual neurons, the amount of data needed to represent the actual brain enough to capture what is the fly really doing is so massive that reducing it down to the model might keep the shape but has quantized the model to a degree far beyond anything we see with LLMs. Maybe if we imagine an LLM trained where each weight is ~1KB (with sufficient training to actually make sure of those bits) is then quantized down to a 1 bit per weight model, but I'm guessing the IRL Fly to FlyWire is still orders of magnitude more quantization. So does it even count as quantization is a bit hard to argue.
Re:Opus 5.5 agents discover two room-temperature magnetic se... by tesnorindian on Tuesday October 06, 2026 @02:13AM (#49974863)
This was more interesting (A fly piloting a robot) - https://www.youtube.com/watch?v=j8RZHLJuwlI
Re:Opus 5.5 agents discover two room-temperature magnetic se... by nico on Tuesday October 06, 2026 @03:12AM (#49975226)
That's pretty impressive, if a bit cruel. For anyone who hasn't clicked, it's from 2008, they glue the fly's body to a little stick. The fly sees a screen and moves according to what it sees there. There are cameras that watch the fly's movements, which then get translated into control instructions for a remote controlled robot (a little battery-powered vehicle)
Re:Opus 5.5 agents discover two room-temperature magnetic se... by tesnorindian on Tuesday October 06, 2026 @03:18AM (#49975269)
Yes, that is 18 years old not in the current LLM era :)
Re:Opus 5.5 agents discover two room-temperature magnetic se... by numeri on Tuesday October 06, 2026 @06:12AM (#49976443)
I think this is the perfect example of how using Claude can trick you into thinking you've done something more impressive, when you're using it beyond your own understanding.

As far as I understand, you downloaded the structure of a neural network, ran it with essentially arbitrary weights, trained a classifier on the essentially arbitrary behavior, then simulated an approximation of that arbitrary behavior.

You could argue that there might be biases towards certain behaviors encoded in the connectivity, and I'm sure you'd be right, but your experiment is incapable of differentiating between those interesting behaviors and random noise. Especially because flies sorta act random anyway.

Re:Opus 5.5 agents discover two room-temperature magnetic se... by jacquesm on Wednesday October 07, 2026 @07:31PM (#50000187)
A connectome describes the connections between neurons. The weights are the degree to which the neurons are biased to fire or not and what you did with your training was to create those biases based on the training data. The actual weights are not in the connectome it is - as the name suggests - just a map. I wonder what the control experiment could look like, possibly a randomized connectome. And maybe it would work just as well as the fly brain. You can do the most horrible butchering of models and then train them back to functionality, this isn't all that different.
Re:Opus 5.5 agents discover two room-temperature magnetic se... by hgoel on Monday October 05, 2026 @07:11PM (#49972043)
Does that really work for superconductors when the mechanisms for superconductivity to emerge are still a major field of study and not something one can just simulate and engineer?
Re:Opus 5.5 agents discover two room-temperature magnetic se... by nico on Monday October 05, 2026 @07:25PM (#49972127)
We can always iterate on the model itself. So you can speculate on the physics (explore the space of physical models), and for each of those physical models, you can explore the space of materials

I have no idea about those actual models, but there are layers of models that you can create, and for each, you can explore with data and compute

It's not a free lunch though. Depending on the task, you might need to collect a lot of the data, or review it manually, or pay a lot for compute, or wait a lot for compute. And still have to iterate a lot on the results, and do your own explorations as a human operator/driver of the whole thing. And then create the materials, test them, get funding to do the whole thing... so theoretically, I think we are in a place where we can successfully apply models to a lot of things, but realistically, we won't be applying all the resources to everything

Re:Opus 5.5 agents discover two room-temperature magnetic se... by wholinator2 on Monday October 05, 2026 @07:21PM (#49972104)
Also, this is about a magnetic semiconductor, not a superconductor. I also got the wrong impression initially before re-reading it. I imagine 99% of non-experts are going to make the same mistake.
Re:Opus 5.5 agents discover two room-temperature magnetic se... by esafak on Monday October 05, 2026 @05:34PM (#49971134)
It is not at all obvious that merely because we have words for concepts, that a model should be able to do all these miraculous mathematical and scientific things.
Re:Opus 5.5 agents discover two room-temperature magnetic se... by nico on Monday October 05, 2026 @05:40PM (#49971227)
You are correct. My comment is not so much about that this is something elementary. But rather an observation that, given the current state of technology, it seems like we are being able to model increasingly more things, in increasingly more efficient and automated ways, to the point that there seems to be a pattern to it
Re:Opus 5.5 agents discover two room-temperature magnetic se... by jeremyjh on Monday October 05, 2026 @05:36PM (#49971165)
Right, it also has to model a substantial fraction of reality (or at least a true simulation of it) to accomplish these things.
Re:Opus 5.5 agents discover two room-temperature magnetic se... by SR2Z on Monday October 05, 2026 @05:58PM (#49971446)
It models our language, which is a flawed and imperfect way of describing the world. So far, it seems like a lot of these discoveries are "filling in the gaps" between the things we've written down and the things they imply (if you have the memory to think them through).

The story about OpenAI's Navier-Stokes solution is a good example of what I mean. I don't think it would have been possible without computer assistance because that proof is long and complicated. I'm also not sure that it would have been possible without a human proposing a new approach to the problem, because by all accounts that's exactly what led to the absurd amount of spending that OpenAI did to solve the issue.

I feel like that at least implies that there's some room left for humans in the new world.

Re:Opus 5.5 agents discover two room-temperature magnetic se... by jeremyjh on Wednesday October 07, 2026 @08:07AM (#49991613)
Math is really very different in an important way: its abstract to everyone. Inferring relationships through a world model is far more impressive to me. Hinton has a clip where he mentioned how impressed he was that an early ChatGPT could draw the parallel between compost heaps and nuclear reactors - totally different forces involved but still chain reactions that begin when a sufficient mass is packed into an enclosed space.
Re:Opus 5.5 agents discover two room-temperature magnetic se... by anthonyskipper on Wednesday October 07, 2026 @04:25PM (#49998326)
You sound like someone out of date on AI. A lot of people think they can't reason, when they clearly do that well enough now to replace a lot of people whose job it is to do that. The reasoning will continue to get better. They don't make intuitive jumps yet, and they lack good creativity, but all that will come before long as well.

But stop thinking that just because they were initially simpler solutions that it stayed that way. We are in the singularity now!

Re:Opus 5.5 agents discover two room-temperature magnetic se... by SR2Z on Thursday October 15, 2026 @07:31PM (#50111806)
I assure you I'm not out of date, but the reasoning these models do is obviously not yet sufficient to deliver groundbreaking proofs entirely on their own.

It took a human being to see the beauty of an alternate approach and purpose it, and only then could OpenAI spend millions of dollars to squeeze out a result.

I am very doubtful that we will see models capable of making intuitive leaps anytime soon, or being "creative." Neither of those are straightforward objectives to train against.

Re:Opus 5.5 agents discover two room-temperature magnetic se... by nater5000 on Monday October 05, 2026 @05:45PM (#49971289)
Yeah...?

That's the pitch of LLMs lol