Software Engineer, Physics enthusiast. We all have the same origin story on a long enough timeline… striving to be less wrong. Re-posts are 80% for bookmarking.
Australia
- Tweets 22,289
- Following 2,742
- Followers 1,428
- Likes 14,668
Pinned Tweet
In 2018, I had become uber-interested in control systems and in particular flight control for drones. I built my own drone from cheap parts, including power distribution, stabilisation and my own flight controller software running on an ATMega chip.
I did more work on a V2 of this system in 2019 & 2020 but tapered off anything further due to COVID restrictions making testing very difficult. Ill try and recover some of my records and share more on the journey in the coming months. It was definitely one of the cooler personal projects Ive worked on.
9
1
47
11,625
James Limmer retweeted
One of the things I’ve observed in our community is how friendship is used to subtly compromise key players in the field. How does this happen? First, someone is approached by a person we’ll call a “contact” (e.g. another researcher, a podcaster, etc.). Over several months, a bond of trust is established. Once that bond is established, the contact starts exerting influence over the key player. This might come with concrete demands. The key player has two options: agree or disagree. The moment they disagree about something (e.g. who they are friends with, who they talk to, or which professional decisions they make), the contact will turn on them and begin undermining their reputation, sometimes in public.
I find this behaviour lamentable. It reminds me of a film from the early 2000s called “Mean Girls” (great movie).
I have no idea how to inspire more dignified behaviour in the field, but how about we try to cherish genuine bonds rather than exploit them for the sake of information control? Be my friend, or don’t -- but be authentic in your intentions.
68
58
626
11,896
James Limmer retweeted
This is incredible. Wow!
In a certain very particular sense, this is now the world’s top mathematician.
Congratulations @sama! So exciting!!
We’re releasing a broad range of new mathematical results produced by an internal frontier model.
We’ve been consulting with the independent Advisory Group on Mathematics and Artificial Intelligence at the Institute for Advanced Study, and we have drawn on their advice and public recommendations to inform how we release these results.
github.com/openai/math
120
118
2,710
196,111
James Limmer retweeted
Ben Affleck reveals he writes Python, understands convolutional neural networks, worked extensively with GPUs, and used his celebrity status to get private looks at Google and OpenAI’s video models
“I’ve always been kind of into computers since I was young. Then, when film started to move from analog film to digital, I became more interested in that aspect of it. The visual-effects workflow for many years has included machine learning, so I can write pretty shitty Python scripts and stuff like that.
“With convolutional neural networks, which were the precursors to what the transformer can do, which is much more computation simultaneously, you would do things like look at what’s called a tensor. That’s the numerical translation of a visual image in numbers, like the batch number, the frame number and the red, green and blue values of each pixel in each frame. It’s just that simple. That numeric is called a tensor.
“You’d use a convolutional neural network to identify patterns that reveal what’s called edge detection or feature extraction, which is identifying patterns well enough to know, this is where the window ledge is, so we can more easily take the green-screen image out and replace it with something.
“That was familiar to me early on because, prior to Artists Equity, I had a small visual-effects company. I’ve worked with GPUs a lot too. The visual-effects guys said, ‘Hey, you should see. There are a couple: Google and this other company, OpenAI, are doing really interesting stuff with transformers in video.’
“I’ve learned that I can actually just call up and go, ‘Hey, it’s Ben Affleck. Can I come see what you’re doing?’ Sometimes people say yes, to my astonishment.”
$ANET Andy Bechtolsheim says AI has made optics demand about 10 times bigger in five years and the industry is only at the "very beginning"
"So I guess I don't need to tell you that AI has been driving this incredible increase in demand for high-speed optics, which is probably now 10 times bigger than it used to be five years ago..."
"And what I want to talk to you today is that we're not at the end of this journey, but rather the very beginning. There's easily another order of magnitude increase in bits needed for the next generation kind of data centers."
"So what I want to talk about first is what's happening at the data center level, and how we as an industry have to get together to solve these problems of this very rapid and high demand growth as fast as possible, since AI can't wait."
______
For full set of takeaways to Andy's AI Datacenter and Optics lecture: firesidealpha.substack.com/p…
819
2,421
32,317
11,446,257
James Limmer retweeted
The FMS70 Ducted Power Single-Axis UAV is an open-source, experimental vertical take-off and landing (VTOL) drone designed to mimic the physics and controls of a reusable rocket booster. Unlike traditional quadcopters that use multiple independent rotors for stability, this craft relies on a single central propulsion unit, making it a popular DIY challenge for testing advanced control theory.
12
33
595
45,567
James Limmer retweeted
reflecting on 1920s images of people working on massive generators for the electrification of the modern world makes me realize that the past was wildly futuristic too
121
474
6,377
140,079
James Limmer retweeted
I don't think people grasp research agents are completely different to an AI chatbot.
I've never seen a researcher do these type of data points to figure out if Maria is authentic.
2
3
24
712
Digital civilization
Introducing Kardashev-0.7, the world’s first trained swarm made of 32 distinct models
Trained with RL for Population Scaling (RLPS), 32 models organically develop specialization & complementary capabilities, delivering frontier performance at:
- 0.007x ~ 0.02x of the inference cost
- 0.03x of the required memory
Civilization advances through different minds specializing and working together. We’re bringing that principle into AI
@BanburyRoadAI, we’re scaling intelligence by model count, toward civilizations of models that learn to build on one another
36
James Limmer retweeted
You've got 12 weeks until January. That's 30lbs.
1. CUT ALCOHOL COMPLETELY
15
84
1,522
2,056,483
James Limmer retweeted
Peter Thiel on world-class entrepreneurs: "They're something closer to polymaths"
a company hands its founder product, psychology, culture and management decisions together, so range pays
go wide
9
55
604
61,602
Good advice
Peter Thiel on how you should think about your future:
"What you end up doing when you think of the future as indefinite is you pick a job that you think will look good on your resume, because it will lead to a different job later on. There's even a credentialing version where you go to grad school, you get a consulting job, and you think of it as this branching tree diagram where you never really do something that you want to do.
You can think of the future on three horizons. Very short-term is: are you going to be happy, and will you be learning a lot? Medium-term is: how good is it on your resume? And very long-term is: are you working on something important or meaningful? We're in a skewed world where we overweight the medium-term and underweight some combination of the short-term and the long-term. I would recalibrate to focus on the short-term and the long-term, and forget about the medium-term."
1
42
Yes.
Peter Thiel on how you should think about your future:
"What you end up doing when you think of the future as indefinite is you pick a job that you think will look good on your resume, because it will lead to a different job later on. There's even a credentialing version where you go to grad school, you get a consulting job, and you think of it as this branching tree diagram where you never really do something that you want to do.
You can think of the future on three horizons. Very short-term is: are you going to be happy, and will you be learning a lot? Medium-term is: how good is it on your resume? And very long-term is: are you working on something important or meaningful? We're in a skewed world where we overweight the medium-term and underweight some combination of the short-term and the long-term. I would recalibrate to focus on the short-term and the long-term, and forget about the medium-term."
2
118
When I watched this section of Peter Thiel’s interview - I had to rewind what he said a couple of times before I went “Ok, hang on - he’s describing a nuance which is never talked about”…
A contrary perspective worth using to help calibrate our moral compass.
Peter Thiel:
I think we moralize way too much.
And there's always a question I like to ask: What is the antonym of good? The moral antonym is evil. The functional antonym is bad.
A Ferrari is an evil car because it doesn't get many miles per gallon. A Trabant is a bad car. It works. It's less good than a Golf car, you know, the former East German car.
So, you know, if you're evil, you're at least competent. And if you're evil, you're not bad, and therefore you're actually maybe kind of good because you're at least getting something done.
I think it's better to be good than evil.
43
Recently learned that Schmidhuber’s Formal Theory of Fun and Creativity, covering compression progress and intrinsic motivation and their relationship to beauty, curiosity, art, science, music, and jokes, was published in a Japanese scientific journal in 2009.
jstage.jst.go.jp/article/sic…
It is wild that this framework: compression progress as the basis for beauty, curiosity, art, science, and even jokes, was laid out in full formal detail in a Japanese journal 15 years ago.
ALT Simple Algorithmic Theory of Subjective Beauty, Novelty, Surprise, Interestingness, Attention, Curiosity, Creativity, Art, Science, Music, Jokes Jürgen SCHMIDHUBER Abstract In this summary of previous work, I argue that data becomes temporarily interesting by itself to some self-improving, but computationally limited, subjective observer once he learns to predict or compress the data in a better way, thus making it subjectively more “beautiful.” Curiosity is the desire to create or discover more non-random, non-arbitrary, “truly novel,” regular data that allows for compression progress because its regularity was not yet known. This drive maximizes “interestingness,” the first derivative of subjective beauty or compressibility, that is, the steepness of the learning curve. It motivates exploring infants, pure mathematicians, composers, artists, dancers, comedians, yourself, and recent artificial systems.
Jürgen Schmidhuber proposed a Formal Theory of Fun and Creativity in the 1990s.🤔 people.idsia.ch/~juergen/cre…
49
119
1,411
116,058
James Limmer retweeted
Still trying to process what we saw last night here in the Stockholm archipelago. My wife, my friend Brian from Canada and I all saw the same thing. A large ball of intense light, almost like a plasma ball, suddenly shooting up into the sky right above my house.
My wife has always been skeptical and never really interested in this subject. Brian has never seen anything like this before either. But last night all three of us saw it.
I have seen things before and I know this phenomenon is real. But for them, this was a first.
So here is Brian this morning, the day after, talking about what he saw.
#UAP #UFO #UAPDisclosure #UAPPhenomenon #UFOs #Unexplained #ThePhenomenon #Disclosure #Stockholm #StockholmArchipelago #Sweden #UAPSighting #UFOSighting #BeyondInstitute
39
39
498
20,480
James Limmer retweeted
🚨 THIS IS A FALCON 9 LANDING TEST - IN A LIVING ROOM.
What you're watching is a custom EDF rocket. No propellant - a 90mm ducted fan inside a duct pushes mass downward, same physics as a rocket engine, just electric.
70mm fan on 7S makes 8kg+ thrust - thrust > weight = hover. Uses thrust-vectoring + PID like a real lunar lander.
DIY VTVL for $250. 🛸🚀
44
126
1,371
86,389