stoniejohnson

5 ms·

1.
An Automated Company 2.5M ARR(polsia.com)
2 points by stoniejohnson 6mo ago|1 comments
2.
Solid State Airflow [video](youtube.com)
2 points by stoniejohnson 2y ago|0 comments
3.
Attention as the management of electromagnetic field lines(qualiacomputing.com)
52 points by stoniejohnson 2y ago|23 comments
4.
by stoniejohnson 2y ago
Great post! I appreciate the diagrams.
5.
by stoniejohnson 2y ago
If you're always looking up but never looking down it becomes hard to appreciate things.
6.
by stoniejohnson 2y ago
Revisited this, and while it was tongue-in-cheek, it feels a bit ungrateful. A lot of tech roles are intellectually stimulating and have a wage in the upper distribution of salaries. I think we have a very nice situation going for us.
7.
by stoniejohnson 2y ago
While those are good reasons to keep the weights static from a business perspective, they are not the only reasons, especially when serving SOTA models at the scale of some of the major shops today. Continual/online learning is still a
8.
by stoniejohnson 2y ago
They were high in 2013.
9.
by stoniejohnson 2y ago
I get where you are coming from, and it is definitely an interesting thought! I do think it is an extremely inefficient way to have a swarm (e.g. across time through training data) and it would make more sense to solve the pretraining probl
10.
by stoniejohnson 2y ago
I was responding to the idea that an LLM would believe (regurgitate) untrue things if you pretrained them on untrue things. I wasn't making a claim about SOTA models with gigantic training corpora.
11.
by stoniejohnson 2y ago
I was responding to the back and forth of: > If you pretrained an LLM with data saying Moscow is the capital of Connecticut it would think that is true. > Well so would a human! But humans aren't static weights, we update continu
12.
by stoniejohnson 2y ago
Humans are not isolated nodes, we are more like a swarm, understanding reality via consensus. The situation you described is possible, but would require something like a subverting effort of propaganda by the state. Inferring truth about a
13.
by stoniejohnson 2y ago
They have a dark pattern around annual subscriptions, i.e. I had a monthly subscription I wanted to downgrade, and with no confirmation they charged me for the entire year after I selected the lower tier and hit next. This is probably why.
14.
by stoniejohnson 2y ago
I'm not really familiar with that technology space, but if you take that as true, is your argument something like: - We don't have limitless CPU cycles - Thus we need to split things into sub-problems If so that might still be ame
15.
by stoniejohnson 2y ago
I think the bitter lesson implies that if we could study/implement "how a machine with limitless cpu cycles would make our eyes see something we are currently thinking of" then it would likely lead to a better result than u
16.
by stoniejohnson 2y ago
Yeah that tracks, I just feel like the difference required to be an ASI might allow for a stronger claim here.
17.
by stoniejohnson 2y ago
I didn't mention it but I fully agree, I imagine ASI would have be to embodied. My reasoning is simple, there are a whole class of problems that require embodiment, and I assume ASI would be able to solve those problems. Regarding >
18.
by stoniejohnson 2y ago
I agree the premise of FOOM is very unlikely. But if you assume that we have created something that is agentic and can reason much faster and more effectively than us, then us dying out seems very likely. It will have goals different from
19.
by stoniejohnson 2y ago
obligatory IANAL, but seeing LLMs: - regurgitate entire passages word for word, until that behavior is publicized and quickly RLHF'd away - rip github repos almost entirely (some new Sonnet 3.5 demos Anthropic employees were bragging a
20.
by stoniejohnson 2y ago
On one hand, my brain kinda shuts off when people start applying structure to spirituality. On the other hand, doing an extensive search amongst possible head spaces you can occupy is a no-brainer for a consciousness implemented on a primat
21.
by stoniejohnson 2y ago
All people do is confabulate too. Sometimes it is coherent (grounded in physical and social dynamics) and sometimes it is not. We need systems that try to be coherent , not systems that try to be unequivocally right, which wouldn't be
22.
by stoniejohnson 2y ago
Language is the medium through which raw perspective refined itself. Language birthed social games and the sense of self. Yes, language evolved for communication. But without communication, thought would still be stuck in the land of instin
23.
by stoniejohnson 2y ago
Given your initial assumptions, that self-moderating end state makes sense. I feel like we still have a disconnect on our definition of a super intelligence. From my perspective this thing is insanely smart. We can hold ~4 things in our wor
24.
by stoniejohnson 2y ago
Totally fair points. > I cannot make the leap from "super intelligent" to "has access to all the levers of social and physical systems control" without the explicit, costly, and ongoing, effort of humans. Yeah this is
25.
by stoniejohnson 2y ago
Valid call out, this was firmly motivated by it being a super intelligence.
26.
by stoniejohnson 2y ago
I feel like if you are an intelligent entity propagating itself through spacetime you will have goals: If you are intelligent, you will be aware of your surroundings moment by moment, so you are grounded by your sensory input. Otherwise the
27.
by stoniejohnson 2y ago
So I think we have some differences in definition. I am assuming we have an ASI, and then going on from there. Minimally an ASI (Artificial Super Intelligence) would: 1. Be able to solve all cognitively demanding tasks humans can solve an
28.
by stoniejohnson 2y ago
> here is an ungrounded, non-realistic, non-representative of a potential future intuition pump to just get the feel of things: > (Yes, there are many holes in this, like how would it piggy back off of our infrastructure if it kills u
29.
by stoniejohnson 2y ago
I think the common line of thinking here is that it won't be actively antagonist to <us>, rather it will have goals that are orthogonal to ours. Since it is superintelligent, and we are not, it will achieve its goals and we wil
30.
by stoniejohnson 2y ago
they said they have a large action model, but allegedly they just have playwright scripts that break as rideshare apps update they said they are faster than chatgpt, but allegedly they are a chatgpt wrapper

More ›