Old man 1: Life is one trouble after another. I’d be better off dead, better yet, I wish I was never born
Old man 2: True, true, but who has such luck ?.. maybe one in a thousand.
My blog: https://cerebralab.com
I’m also building an open source generic ML library: https://github.com/mindsdb/mindsdb & https://github.com/mindsdb/lightwood …. which I guess might be of interest to some people here
I’m one of those odd “eat anything people”, I mean like, chug galons of milk, stop eating for 10 days, eat 5000+ kcal a day, go for months eating street food in poor parts of SEA and India, raw meet of various kinds, any random fruit peal on etc.. with two food poisonings in my life (both raw oysters) and feeling nausea once or twice a year tops, if even.
Granted FMT is about more than food tolerance, but I assume that’s a good proxy, and given the fact that I’ve never been to athletic nor health conscious food wise I have to think there’s some “hidden variable” influencing this and it’s probably more subtle.
That being said the Danish study looks promising and it’s better than nothing, but I’d love to have some stronger theoretical backing for what the secret sauce of microboime actually is, besides poorly fitting but non zero effect proxies.
Agree on that point, though I think the “crazy things are happening” scenario is likely one where money is less legible so countries bother about taxing such things less.
I can see that happening, and that’s part of “how do you find meaning once work is automated” which is the question I struggle with most. Props on brain extensions that you are conscious of/as, it does seem intuitive to me that something like that will soon be available.
Create an account with goose ai or open ai, go to “completion”, paste in text you wrote thus far, paste the completion.
I think that’s very relegated to LW and AIAF, but I think you don’t have to dismiss some probability of that happening in order to prepare for the probability where it doesn’t or where it does in a “hundreds of years soft take-off” way.
This is why I said I’m not certain it will have an impact, rather than I’m sure it will or won’t.
I can see self-driving having a big impact, but won’t that impact be outweighed by the kind of jobs people have to drive to being mainly automated?
Hence why I make mention of it in the article:
>The “important” bits of automation will happen in both. Germany and the US will both want their bombardier drone swarms. It’s just that the latter will have the engineers projecting them be driven to his co-working space by a self-driving car, getting their “Mediterranean wrap” from the claws of a drone; While the former will have said engineers take a subway to the office, then get this shawarma delivered by a guy on a bike
But the vast majority of automation doesn’t seem to be militarily relevant. Even if you assume some sort of feedback loop where military insubstantial automation leads to better military automation, world powers already have the trump card in terms of nukes for wars of aggression against them.
That could change in the future, but any government could theoretically institute a policy of international taxation for its citizens, so the US doesn’t seem particularly worse.
Would be rather difficult or impossible for most, since many intl trade treaties forbid double taxation among members. The us is an outlier because it can afford special treatment afaict.
If you assume the layman definition contains any information, which it doesn’t
maybe we should think of “explainability” as the AI’s lossy compression quality for its theories, in which case it must be evaluated together with our ability, as all modern lossy compression takes the human ear, eye and brain into account. In this case it could be measured by how close our reconstruction is to the real theory for each compression.
My view is along these lines, and see first link for an interesting example vis a vis this (start at min 17, or just read the linked paper in the description)
No reason in particular, wanted to switch fields to something more health/bio-related, wanted a new project to change things up, figured out I can probably get a better contract with other companies.
I do think it might not actually be the best field career prospect-wise, but that’s another topic I might write on later.
See my reply to Yitz, tl;dr I am hedgy on alignment being an issue due to being hedgy on algorithms getting “too powerful”, and I am hedgy transparency or interpretability is even a coherent concept for any algorithm complex enough to not be so (sans some edge cases).
But it’s not like I’m the best person to say, far from it.
To be clear you mean you think that progress will slow down, speed up, or just be anywhere but constant?
Anywhere but constant, i.e. currently it seems to be slowing down with a focus on productionizing what exists.
Current avenues of research seem bogged down by an inability to actually establish “what works”, which is a problem that I see no clear solution, interesting problems can no longer be scored with a simple number representing “accuracy” or “performance”.
The one solution I’ve seen to this is hand-waving around RL-like system that are left to train online or on an internet-like simulation coupled with control interactions with reality, but ideally in a simulated way whenever possible. E.g. see Gwern’s recent clippy post for a tl;dr on that. In practice that seems to have severe limitations to me but who knows, I’m not smart enough to rule it out entirely.
The one thing I’m sure won’t happen is the last decade again, which seems to be mainly an artificial of the mutual realization that ml algorithms can run on GPUs and GPUs can be designed with ML in mind, Culminating in things like the transformer, which e.o.d. can be tought of an an “unrolled” recurrent/hopfield NN lacking some abilities, but able to take advantage of massive parallelization and memory capacity via the trasnpose + multiplication. But my read is that this success has lead to the bottlenecks now appearing in areas that are not compute.
What are these?
Misspelling of “causal”
Wait, you were actually serious about the paying people part :)))
Eh, I think I came in a bit too strong trying to argue that even if he was 20% goofing around and thinking oddly, that 20% might me more than current academic slack allows.
I did try to clarify that in a comment in the original blog-post but I do feel like I’ve heard takes of this argument made before.
The biggest update for me are people claiming he actually was a top mathematician in every sense of the word, which might be true and would disprove a big part of my hypothesis, which is that he was a good mathematician, but good in the “top 10k people in the world” sense not in the “top 10 people in the world” sense.