YoungAchamian
We walk conditioned ground and name our folly civilization.
No bio...
User ID: 680
YoungAchamian
We walk conditioned ground and name our folly civilization.
4hr ago
Please, Millennials brought the resurgence of German Beers, Meads, Trappist, Darks, and Sours back into style. Just because hop ODing is easier does not define our grain-juice palette.
YoungAchamian
We walk conditioned ground and name our folly civilization.
4hr ago
Traitors!, Jezebels! Heretics! You move from 18% Imperial Stouts to nice 12% Barrel Aged Tripels and Quadruples!! Maybe a sleepy 4% Honey Kolsch or Helles Lager depending on the situation
YoungAchamian
We walk conditioned ground and name our folly civilization.
7hr ago
Oh that's hilarious. Is there a technical distinction between a Township and a Town?
YoungAchamian
We walk conditioned ground and name our folly civilization.
7hr ago
I failed to hold myself accountable.
I built out my training/data pipeline and set about building my first pass at a model last week. I hate what I came up with, went back to the drawing board. I read the PINO paper, realized it was too mathy for me to grok at a first pass and promptly lost interest in doing anything last week. However the competition is coming to a close and I never really submitted anything, I might try to train and validate one of their baseline models just for the hell of it.
I talked with the co-worker who suggested it to me, looks like the same thing happened to him. Which is unfortunate. Sometimes it feels like work can really drain the motivation to do similar stuff outside of work. I used to love doing engineering ideas on ML models at home, now it just feels like work. I want to start fucking around with a 3D printed drone idea I have but now I'm afraid I'm just going to repeat this experience again with added money waste as well.
A 'village' sounds like something thirdworlders live in.
I grew up in a farming town in the fly over states. We had a both a "Town of Bumfuckville" and a "Village of Bumfuckville" as two separate incorporated municipalities. The town was technically the larger outlying land, and the village was the area around the main-street, the branches, and a few of the neighborhoods surrounding it. No clue why it incorporated like that, but while they were technically distinct legal municipalities they did play together (along with other towns) for the school district.
YoungAchamian
We walk conditioned ground and name our folly civilization.
8hr ago
Ah yes the monte carlo approach.
YoungAchamian
We walk conditioned ground and name our folly civilization.
1d ago
Maybe this is just me, but I actually don't enjoy writing.
It's not just you, I also don't enjoy writing. But LLM outputs of my thoughts never really hit write (buh dum tss) either. They never really convey any soul to my thoughts.
YoungAchamian
We walk conditioned ground and name our folly civilization.
1d ago
Don't tell me, are you opposed to strapping IEDs to low performing kids for deployment on the front lines?? This gives them purpose in life and removes their dysgenic genes from the societal pool! A Win/Win!
YoungAchamian
We walk conditioned ground and name our folly civilization.
1d ago
Hear, Hear! Millennials are the Craft Beer generation. We popularized it.
YoungAchamian
We walk conditioned ground and name our folly civilization.
1d ago
But it would be a challenge for me to find something on my (English) bookshelf that doesn't extensively use them
I'm a prolific reading and I have never noticed them in any book I've read. I notice them non-stop now.
YoungAchamian
We walk conditioned ground and name our folly civilization.
1d ago
Wow someone call Anthropic and tell them they discovered AGI through this "one simple trick". Just write the soul.md as "Be human-like come up with your own tasks, make no mistakes". If only all the researchers had your ideas!! Or, your understanding of how the soul.md works is technically deficient.
AGI has always meant "Human-like artificial being". You may think Astra is, but most people can intuitively tell the difference. This is the problem with people who are rhetorically skilled, or as I call them: Wordcells. Just because you can craft a clever argument that the "sky is hot pink" does not rewrite the skeins of reality to make the "sky hot pink" You can say that a "broken galley slave" is not intelligent, but even broken slaves have dreams, have thoughts not driven by their task. It is clearly different than an AI-slave.
YoungAchamian
We walk conditioned ground and name our folly civilization.
2d ago
Because it's not a revolution if it hasn't massively improved the state of the art. The SotA for robotics is that it can already do non-local power, non-local compute. There are technical reasons as well, but at this late on a Sunday, I don't want to get into them. Ask Chat or something.
Waddle had me nodding along until the reveal was "Instead of using a massive dataset + training, we'll just use something with a massive data that was already trained!!". Talk about missing the trees for the forest. A chance to have an inventive idea squandered by doing what every other AI startup in the world is doing: using someone else's LLM with zero moat.I'd be remiss if I didn't point out the majority of what Waddle is doing was invented/discovered originally in 2022. I don't think that meets the bar of "revolutionary new thing".
When people in the field first started talking about creating an Artificially Intelligent Being, the need to pick apart minutiae of definitions between a "General Intelligence" vs a "Singularity Intelligence" was pointlessly pedantic and not at all relevant. It hasn't become more relevant in the past 30 years either in my opinion. It feels like a pointless wordcell argument to make pointless wordcell definition fights, for pointless wordcell internet debates. The Culture is not a "Singularity Intelligence" either because The Culture is not "AI-GOD", neither are the AIs in Hyperion. Intelligent yes, self aware, with their own agency, absolutely. But considering the whole plot of Hyperion is them getting bamboozled by some time travel + human empathy, with a strong concurrency of actually trying to create an "AI-GOD" to worship, that would point them towards not being an "ASI".
Likewise, working towards an assigned goal doesn't tell us much about how general the intelligence doing the work is. I spend most of my working day performing tasks I am assigned, that do not intrinsically motivate me
This is always a confusing objection, to the point it feels uncharitable. Spend 5 mins looking in the mirror and think about all the things that you do that are NOT task oriented. Think about all the thoughts, feelings, or other cognitive processes that every human does that have nothing to do with the task their boss gave them. Literally 5 mins thinking about human cognition points to a massive gap between how a programatic/algorithmic intelligence like an AI operates vs how Humans operate. I'll reiterate, AGI is human-level artificially intelligent being. Getting stuck on "general" is some, idk, semantic trip-up.
X is 1000 times less sample efficient
This is cute, the real number is likely north of 100,000x or above and is non-linear across tasks.
10,000 times the training data a human
How much training data does a combat medic get when doing field operations on a wounded soldier? A month long course? How much training data would you like to bet it would take a combat medic AI to do that job to the same level of efficiency, with the same level of situational awareness? How much do you think it will cost to collect all of that data? It's easy to give models 10,000x training data when that data is relatively easy to access and buy. Suddenly sample efficiency becomes a massive burden on any real-practical ML model. Making it sound like such a simple thing is annoying. It comes across as handwaving the actual hard problem.
God knows what the frontier labs are up to these days
Paying hundreds of millions of dollars to create datasets specifically for training ML models on math + reasoning. Naturally-available data did run out, but money didn't, and the frontier labs were able to prove the business case for spending obscene amounts of money creating new data.
synthetic data
If synthetic data is such a solved problem then how come the Sim2Real gap still exists, is unsolved, and is the target of plenty of research dollars? I feel like you are miscommunicating something here, or misunderstanding what "synthetic data" means in ML terminology.
It does make “these things need millions of examples to learn something new” rather difficult to sustain, regardlws sof practical relevance - which I dispute. They observed it figuring out unfamiliar mechanics and constructing symbolic models to plan around them. Your claim about analogical reasoning needs similar qualification. I do not believe they're the same thing anyway.
Huh? It literally does take millions of samples to train a LLM are you arguing elsewise? And it is impractical for areas where millions of samples do not exist. This argument/objection you are making makes no sense. Please elaborate.
Analogical reasoning - Ever heard a sports analogy applied to a non-sports topic. Did that analogy help convey a better intuitive understanding of some facet? Did that analogy allow someone who has never done that non-sports topic a better starting point, or better performance before being given that analogy? Such is the power of analogical reasoning, or the ability to convey how one statistical distribution is similar along a particular latent axis to another statistical distribution, with the purpose of using the already learned statistical distribution for performance improvements on the new unknown distribution.
Alternatively, we could solve continual learning, or simply reduce the temporal delta between train-deploy-train to the point that it has no practical relevance.
Yes such sci-fi talk much wow, if only we could solve FTL then we can truly conquer the stars. Obviously FTL is a forgone conclusion, it's so trivially simple to solve, leave it to the shape-rotators. Let's get back to planning galactic expansion or galactic political organizations.... I pointed out the challenge on RSI on a different post below. It's not "simply reduce".
My concern is how much useful work it can do, how quickly that range is expanding, and what remains exclusively ours. I don't expect the terminology to buy us much time. I certainly don't want to spend more time arguing terminology.
Then why try to create new terminology? It's an good, useful AI model, nothing more is needed, no new terms needed. AGI means what it has always meant, the semantic definition isn't being pushed to the stratosphere so that we can classify Astra as AGI and win our internet arguments and secure more VC dollars.
Intelligence is about solving problems.
Sure, you're an intelligent being. I imagine when you aren't solving problems that somebody else as told you to solve, you sit there nice and still, with zero thoughts in your head and no-consciousness.
Let's not pretend. Astra solves directed problems, it does not have the agency to do otherwise. Humans don't no matter how manny word-cell arguments wish to define it otherwise.
In reality, once a model is trained,
No one is arguing otherwise. You still need to train the model, with millions of examples. The fact that you can copy it post training says nothing about its sample efficiency, nor its AGI-ness. I'm sure in some equally distant future alternative universe where humans learn to mind-scan other humans, we can create embodied brains of digital human slaves, solving your butlerian-deficient solution.
YoungAchamian
We walk conditioned ground and name our folly civilization.
2d ago
Not to nitpick completely, but there is no short term threat via RSI because RSI isn't here. Theoretically yes, if RSI existed, I think the AI models might begin to reach outside their current training datasets. But part of the cost of AI rn is compute for training. If it takes 100 Million worth of compute to train an LLM, I don't think a company is going to give an LLM-Agent the ability to just start up training runs for the "Self" part of RSI. RI already occurs but part of that limitation is that baked into the RSI argument, is the idea that the AI will make improvements that the human researchers can't understand. This presents a conundrum from a simple business economics perspective, are you the company going to risk billions of dollars on potential RSI runs which also could just as likely be dead-end hallucinations? You obviously can't understand the improvements the AI is trying to make, it's smarter than you. So it comes down to a matter of faith, is an AI lab going to take a leap of faith that "this training run is the RSI first domino" that will lead to an investment recoup to pay for the next RSI domino. We are not there yet.
LLMs' flaws compared to the human brain are an algorithmic
Maybe, but that doesn't mean that algorithmic solutions are actually discoverable. This harken's back to the OG AI researchers trying to discover how humans did it quantitatively as a precursor to coding it into computers. They spent a lot of cash + time trying and got their lunch eaten by data-driven models. Now we've just proposed that we'll sick a hyper data-driven model on the same question. It really is just a shot in the dark. We think its possible because humans can do it, but LLMs don't work at all similar to human cognition so we don't have any evidence that they can achieve an algorithmic solution to their already existing problems.
YoungAchamian
We walk conditioned ground and name our folly civilization.
2d ago
Depends on how deep in the technical weeds you want to go.
- Reinforcement Learning: An Introduction by Barton and Sutton.
- Deep Learning, by Goodfellow, Bengio and Courville.
Probably good starting spots. I'd advise to stay away from AI Safety literature, it's definitely going to be more digestible but without an understanding of how AI actually works, its will be pretty hard to separate "what is true" from "what is a cargo-cultist belief based on science-fiction literature".
I'm a robot with non-local power myself
I didn't realize you were hooked up to a machine that replaced your stomach? 100% of normal humans have local-power derived from the bio-reactor that is their digestive system. They contain localized computing from their brain. The "Robot Revolution" has always been imagined across 90% of sci-fi as localized robots, who have individualized agency, local power, local compute.
This genre of dismissals is fair enough but getting vacuous.
Yes. No. Maybe... It's really not about an opinion or criticism, it's just the actual statistical fact. The current methods of training for Neural Nets are not designed for out-of-distribution learning. Everything you are describing is not changing that, it's just an attempt to expand the distribution, so that less things are outside of the distribution. With the current paradigm you would either need to quantize the whole of all worldly knowledge and train models on that, or we would need to achieve RSI across modalities so that AGI agents could expand their own knowledge outside of their training dataset.
For example, asking an LLM (yes even Astra) to give you the mathematical response of a ball bearing hitting a human body, asking it for the tissue injuries, how the bearing responds to stress/pressure/impact, how does it effect bone, etc. Is genuinely out of distribution. Having tried this, it does not work well. Why should it, it have nothing in its training set about this. The kind of actual quantified numbers on the subject is highly restricted, speculative, and mostly only generated via blowing up pigs and testing Finite Element Models. But because it's so limited it also nearly impossible to create a benchmark dataset to test AI models performance on stuff like this. Double edge sword. There a whole hosts of engineering problems and disciplines like this that LLM models do not perform well in.
YoungAchamian
We walk conditioned ground and name our folly civilization.
3d ago
GEN 1.5
I saw this one, haven't had the chance to look at Helix yet.
I don't see why this demand is fair.
I'm just calling it how I see it. There's a couple arguments. One from a generalized capabilities standpoint. Robots with non-local power are only useful in factory settings. We also have them. There is no need for "LLM-factory robots" because factory robots don't need to adapt on the fly. Despite EA fears, not much is given by making a factory controlling AI, AGI smart for the purposes of making paper clips. Robots that require an LLM datacenter to control them are easily disrupted, latency is still an issue and it becomes more of an issue when you need to be processing so much information all the time. The last part is that robots for very specialized situations that can't adapt aren't very useful outside of the niche uses they have specialized in, which likely requires training, which is expensive.
Two is from a "what is the current state of the art" The current state of the art for mono-situation robots with centralized power and control is already here. If you want to "revolutionize" robotics you need to do more that what already exists.
The third argument is a funding argument. It should be a shock to nobody that the 3 main funders of Robotics work is defense/military, manufacturing, and VCs interested in consumer robots. 1 is going to need all of the above, 2 doesn't really need LLMs to be added, they have other needs, and 3 is probably the most wishy washy of the bunch, but in general people are wary of centralized robots that feed all of their personal data to a central server.
robots with complex behaviors will almost certainly have some combination of cloud forebrain + local hindbrain
Yes, but thats not really an argument against the above. That's just an argument that a certain class of robots will require a reasoning engine like an LLM and a controls module like what currently exists.
I don't think this holds after Astra crushing human baseline on ARC-AGI-3.
While I'm not an expert on all the various LLM benchmarks, but a brief look makes me think this is exactly the sort of thing that RL game playing is good at, and learning general game-playing strategies in training would generalize pretty well to this. I'm not going to goodhart a new metric into existence, but you conceivably need something that there exist zero training data for, including close enough transfer learning data.
YoungAchamian
We walk conditioned ground and name our folly civilization.
4d ago
distributing RFID tags to your troops is that they can be used by both sides
Yes, I've heard that point a lot. I know an RF guy who's trying to do some passive emitter thingy that goes wayyy over my head. No clue how it works. But he has some pithy comment for it that I am forgetting, its quite funny.
YoungAchamian
We walk conditioned ground and name our folly civilization.
4d ago
I actually thought that open-source imagery would be better resourced than that
There are what a couple hundred grainy low res videos of Ukr drones targeting Russian tanks + other vehicles? It's well below the threshold to train an ATR model on data quantity, without even getting into quality. Stock images are from non-operational angles (you for the most part aren't level with targets, its an isometric or overhead view, NERF to solve that is expensive and has dubious performance improvements, and having worked in the generation of low-sample military images, it can actually be quite hard to secure exemplars in EO imagery. SAR imagery just sucks, speckle is multiplicative noise which makes it nearly impossible to remove and hard to detect around.
Humans are obviously much easier to target, however then we get into friendly fire and the fairly disturbing optics of AI drones targeting your own soldiers... There are logistic issues around distributing RFID and similar tags to prevent it, that to my knowledge are being worked through. But the main goal of drone warfare is destroying expensive vehicles with cheap drones.
Training datasets for the ATR stuff are being developed but its sort of an incumbent's advantage, with active known defense contractors having most of the customer connections to be able to get the data, and the resources to label it. Unfortunately they are also slow AF, which is why defense tech startups sorta eat their lunch in certain spaces. Transfer learning is something I've seen recently, taking civilian jeeps with SAM-esque models and applying them to humvees etc. It's always much easier to train on your own nations hardware anyway. Though it gives bad optics to Brass, who don't like it, and an open question on whether it will transfer.
Give the robotics people a year, since LLMs are already revolutionizing robotics too (why the fuck not)
Uhhh what? VLA is good but its not revolutionizing, but so are Diffusion models and those are not LLMs. The also aren't "Astra" level in reasoning either. To actually revolutionize robotics on the level you seem to be catastrophizing about would require entirely local models running on local power, local compute, able to be applied across a wide variety of operations in a wide variety of environments. We're not there unless you have some additional evidence to prove your point.
I've more or less concluded that Astra is dangerously close to AGI, and probably meets most reasonable criteria for it (good luck finding a consensus definition; the goalposts are on Mars)
AI used to be the word for Asimov-level artificial intelligences that could make their own decisions, and operate with their own agency, maintaining long term planning horizons, memory, possibly even emotions. The word got shifted to AGI. If you want a empirical definition its science fiction AIs like the Culture, The AIs in Hyperion, Daneel in Foundation. The goal posts keep getting punted because people keep trying to change what was previously intuitively understood so that they can sell their idea as the one true AGI, win internet arguments, or catastrophize about the oncoming doom. Astra is only able to really solve problems, it has a moderate amount of self agency in the scope of completing its tasks, and exhibits some planning ability, again in the scope of its assigned problems. It's powerful enough to be "dangerous" sure, but its not really AGI as is commonly understood.
The competitive advantage I retain (and most people, really)
The competitive advantage that you retain is the ability to learn shit without requiring millions, even billions of examples. You, like any smart human also possess the ability to do analogical reasoning (out of distribution reasoning), something that eludes current LLMs by and large.
YoungAchamian
We walk conditioned ground and name our folly civilization.
4d ago
I'm enjoying the political face one, I have like an 86% success rate with guessing the more leftist woman by picture. Sometimes its hard to determine which one is the MOST leftist though.
YoungAchamian
We walk conditioned ground and name our folly civilization.
4d ago
But LLMs are not the only form of artificial intelligence, and for most military tasks a strong general-purpose LLM is both overengineered and poorly suited for the mission, especially the mission of terminal target determination. For intelligence synthesis, they likely have some promise.
You have the right of it.
The additional issue is data, to a shock of nobody paying attention. Military data is near non-existent. And the data that does exist is low-resolution, or tightly controlled by non-ML/AI folks who have been trained to have negative desire to share (The Security Clearance process over time selects for a type). LLMs are the Ford F150 of ML models. Half the battle is often getting a very specific tank configuration that is actively trying to hide in foliage from the very CV models you are trying to detect it with. You have N=3 samples of this tank sitting on a tarmac somewhere and it turns out the ability of your model to extrapolate that picture to 3.0 GSD satellite or drone footage is abysmal. And that is before it even starts attempting adversarial countermeasures.
There are of course solutions, but they don't work as well as folks imagine they do, including generative image generation. Most of the examples of drone based hits are FPVs or FPVs with a "wait" situation where the drone hovers, detects a general class and is then given approval by a human analyst who manually reviews the vid/image.
YoungAchamian
We walk conditioned ground and name our folly civilization.
4d ago
There's the demo floating around of the fly neural network that's been trained to solve a Rubik's cube. Presumably it won't be able to solve a Millennium Problem
It won't, it's been trained explicitly to solve rubick's cubes in a way that is unlikely to generalize past rubick's cubes. This method of small network + RL on some straightforward game problem has been around forever. I remember training an RL model to play Euchre back in the day through self play. You can fit a lot of performance onto a small model that has been trained in a supervised manner.
YoungAchamian
We walk conditioned ground and name our folly civilization.
4d ago
This needs a "not interested in either" So many of these I feel forced to chose between two horrible options.
Idk places around me, there are 3-4 meaderies in my local area. I consider them beer adjacent because it's really only craft beer drinkers that tend to get them. Wine drinkers just get wine. Yeah Trappists as in Doubels, Tripels, and Quads, as the beer type. From the sparkling-ale region of my local breweries. And Sours, Farmhouse Ales, Goses, etc have absolutely flooded my local beer stores that it's almost gotten to the point that if I see a barrel-aged bottle of expensive beer, its a 50/50 flip between an Imp Stout and a Sour.
Depends on how hip honestly. We have local beer stores and local breweries that have wide varieties. When I was in Chicago I frequented Beermiscous and Beer Temple a lot. Restaurants around here often have locals which are fairly wide in variety, my current favorite is a nice ESB that is spreading, but yes boomer establishments where the average age of the clientele are 50+ stock mostly disgusting domestic/Macrobrew IPAs
More options
Context Copy link