thinkzilla

9 ms·

1.
by thinkzilla 20d ago
c.f. the Berry paradox https://en.wikipedia.org/wiki/Berry_paradox
2.
by thinkzilla 2mo ago
I'm under the impression that the majority of startups go nowhere, so the fact that they don't have this property is not itself a major statement about the field as a whole.
3.
by thinkzilla 2mo ago
The Attention Is All You Need paper deserves a lot of credit, but it also builds on other ideas. Earlier versions of attention were introduced in 2014 papers from (not U.S.) universities.
4.
AI alignment research is unintentionally building a censor's toolkit(s-ball-10.github.io)
3 points by thinkzilla 2mo ago|0 comments
5.
by thinkzilla 8mo ago
Making estimates is fine, but we should be clear that they are (provably) just opinions. http://scribblethink.org/Work/kcsest.pdf https://news.ycombinator.com/item?id=13731975 https://www.am
6.
by thinkzilla 10mo ago
I generally agree with the point of the article ("Fourier transform is not magical"). However saying it is "just" curve fitting with sinusoids fails to mention that, among an infinite number of basis functions, there are
7.
by thinkzilla 11mo ago
The comparison to railroad infrastructure is interesting. I think the author is wrong on this point however: > Today’s tech just cannot do what will be required of it (AI shouldn’t be dispensing medication when it can’t even count to 7).
8.
by thinkzilla 1y ago
I am sympathetic to the motivations and argument in the article, but the analogy with bridge building is flawed. In engineering, if your assumptions are correct and you apply the formulas correctly, the bridge will not fall. This is _not_ t
9.
by thinkzilla 1y ago
While the post uses DPO to illustrate RL and RLHF, in fact DPO is an alternative to RLHF that does not use RL. See the abstract of the DPO paper https://arxiv.org/abs/2305.18290 , and Figure 1 in the paper: "DPO op