LarsDu88

4 ms·

31.
by LarsDu88 1mo ago
This is the infamous "stack fallacy" https://techcrunch.com/2016/01/18/why-big-companies-keep-fai... The idea that the layer above your own is trivial to build (e.g. biology is just chemistry, or ch
32.
by LarsDu88 1mo ago
They finally did it. Someone decided to automate fun. Instead of having fun yourself, your AI agent can have fun for you.
33.
by LarsDu88 1mo ago
I'm sorry to say this, but the voxelization diffusion model I mentioned is effectively AI sculpting, and AI animation is already quite good (for humanoids at least) Virtually all the dominos have fallen except good topology at the mome
34.
by LarsDu88 1mo ago
The real algorithm that's needed is an intelligent AI mesh retopologizer. Honestly more of a reinforcement learning AI problem than a generative AI problem IMO
35.
by LarsDu88 1mo ago
The leading 3d gen techniques are latent diffusion models that work off voxelized intermediate representations of objects. When you convert these into meshes, very often they will have poor topology. Topology is important if you need to mod
36.
by LarsDu88 1mo ago
China is more authoritarian but the US under republican presidential administrations has been more more militarily aggressive. Every single Republican administration has initiated invasions of other countries and every Democratic administra
37.
by LarsDu88 1mo ago
Its not an old idea at all. Its from a 2012 article by Graham Alison followed by a 2017 book on the rise of China specifically that Xi Jinping read and referenced. Its claimed to be an old idea, but its actually only recently become popular
38.
by LarsDu88 2mo ago
RL can do behavior cloning, but really needs good simulations or verifiable environments to get to superhuman levels. That currently exists for math, coding, and a lot of videogames. Soon there will be good enough simulations for robotics.
39.
by LarsDu88 2mo ago
They will get sharply better in tasks with verifiable domains... math and coding Gradually the labs will start engineering verifiable sandboxes for wider domains like videogames This strategy will hit a plateau in about 18 months and then w
40.
by LarsDu88 2mo ago
You're take basically lines up with Francois Chollet: https://arxiv.org/abs/1911.01547 intelligence is more like polishing a ball smooth than growing the ball to infinity. For many tasks, it will be smooth enough.
41.
by LarsDu88 2mo ago
This whole episode suggests to me that people are misjudging the future of robotics. Many companies are building robots to exactly imitate humans, perhaps with the thought that the "egocentric" training data widely available will
42.
by LarsDu88 2mo ago
Yet the end result appears that people on their own accord go and "learn" things like vaccines are bad, fluoridated water is used for mind control, rectal enemas can cure bowel cancer, and that the moon landings were fake. At leas
43.
by LarsDu88 2mo ago
Run the sim many times... faster than it can run on actual humans and compute a probability density for specific events. Better yet use it to dimulate counterfactual phenomena like market manipulations ypu intend to enact...
44.
by LarsDu88 2mo ago
It would be extremely foolish for an early investor like Alphabet to liquidate SpaceX holdings en masse all at once. Also, for better or worse, SpaceX did increase by about 200B today. For all it's faults, it is forming a massive milit
45.
by LarsDu88 2mo ago
This is an architectural limitation that may be overcome by how you bake the MoE (mixture-of-experts) onto silicon. If you could manage a per-die expert somehow and keep the expert routing gate relatively fast (through an interposer interco
46.
by LarsDu88 2mo ago
It depends on how quickly you can bake new architectures. Text diffusion might be a disruptor here, but let me just say the most cutting edhe form of image diffusion (JiT and DiT) right now is just a big fat stack of alternating attention a
47.
by LarsDu88 2mo ago
Massive economic simulations with thousands if not millions of agents to front run the global economy and stock market. Fully interactive realtime NPCs in videogames at scale. Recommender systems that simulate individual consumers. Crazy sh
48.
by LarsDu88 2mo ago
I'm surprised neither OpenAI nor Anthropic made this move first. The Chinese open weight models are pulling ahead and commoditizing their value proposition. Baking models onto silicon would've been the next logical move to get a m
49.
by LarsDu88 2mo ago
What do you mean? Google bought SpaceX shares when it was a tiny startup. They are probably more than 100X on their initial investment. Even with a 99% drop in SpaceX stock, Google would still be positive. Edit. Google invested 900 million
50.
by LarsDu88 2mo ago
This is a very rose tinted take. AAA game development was far more brutal in the 90s. Go look up the development of Sonic Xtreme in the 90s. A brutal development process in the 90s that led to more thsn half the development team camping out
51.
by LarsDu88 2mo ago
This move was purely to pump up SpaceX stock price at its current absurd valuation b/c Google owns something like 6% of SpaceX
52.
by LarsDu88 2mo ago
Jony Ive wanted to chop down all the trees at De Anza Community College for something like 11 million dollars for an Apple event. There's a lot of ways he was demanding in the wrong ways for Apple
53.
by LarsDu88 2mo ago
Only if I can build a giant ass linear city in the middle of the fucking desert with a big ass maglev in the middle in Sim City 4000
54.
by LarsDu88 2mo ago
I'm almost certain the goal of this startup is to make physical automated research labs guided by RL
55.
by LarsDu88 2mo ago
But my GPU...
56.
by LarsDu88 2mo ago
Does this mean we'll finally get a new Sim City?
57.
by LarsDu88 2mo ago
I think the nuance that's being missed here is that the profit margin of inference serving when there's perfect competition is close to 0. Zero profit means no recouping investment costs effectively bursting the bubble. Technologi
58.
by LarsDu88 2mo ago
>Anyone who invests while convinced that only a monopoly would save his investments has got his brain screwed on backwards Whats the common trait between Warren Buffet and Peter Thiel? They only invest in companies that are monopolies or
59.
by LarsDu88 2mo ago
There may be a market for the super models, but the reality is, the super models and the merely ok models already saturate on lower levels tasks. You don't need Fields math PhDs to make you a web app or plan your vacation. The remainin
60.
by LarsDu88 2mo ago
Literally everything chatbots don't do along with everything chatbots currently do. They will gradually go from doing your dishes, to contructing an ADU in your backyard by 2040 System 2 (thoughtful slow planning) has already been crac

More ›