Zack Dulberg

1.2K posts

Zack Dulberg

Zack Dulberg

@pianozack

Physician & PhD in computational neuroscience @PrincetonNeuro, Senior Fellow @ncri_io

Katılım Şubat 2009
338 Takip Edilen581 Takipçiler
Sabitlenmiş Tweet
Zack Dulberg
Zack Dulberg@pianozack·
Why do we grieve? In our latest paper, just published in Psychological Review, we use simulations to suggest that grief may function, counter-intuitively, to maximize reward! This is the first computational model of grief we know of in the literature, and we show it can account for several empirical grief trajectories seen in humans. The paper can be found online here: psycnet.apa.org/record/2026-20… Please reach out and I will be happy to send the most recent PDF! An older pre-print of the paper is described in more detail in the thread below
Zack Dulberg@pianozack

I am happy to share our latest pre-print, in which we model grieving as a natural solution to the reinforcement learning problem of adapting to loss. Given the amount of grief in the world these days - we thought it was important to better understand it. biorxiv.org/content/10.110…

English
2
2
6
5.5K
Zack Dulberg retweetledi
Adam Louis-Klein
Adam Louis-Klein@adam_louis52328·
According to @ezraklein, antizionism is a neutral reaction to Israel's actions over the past two and a half years, driven by genuine concern for human rights. But scientific data from the @ncriio completely contradicts his view. The data shows that antizionist views correlate with a heightened tendency toward moral inversion — a systematic distortion of the perceived human rights records of authoritarian versus democratic states. Antizionists misjudge human rights. If antizionists would claim concern about Israel's drift from liberal democracy, then they themselves tend to equivocate between, or outright prefer, authoritarianism to democracy. My piece with @pianozack, who led the study, at the @TheFP
Adam Louis-Klein tweet media
English
13
56
212
5.2K
Zack Dulberg
Zack Dulberg@pianozack·
But why? We can use distillation to infer so many of the parameters. We have already used a mere fraction of collective human brain outputs to produce superintelligence. As data capturing brain outputs continues to explode (audio, video, movement, etc.), models will continue to imrpove, and particular brains can be distilled through additional finetuning. At the end of all this, how much brain structure is left un-modelled? That gap should be your only focus. Why not see how far solving the inverse problem gets us, before you painstakingly rediscover parameters we have already distilled?
English
0
0
0
411
Isaak
Isaak@isaakfreeman·
I’m leaving MIT and not continuing into my PhD. AI is coming too fast for humans to keep up. But there might be a way: I realized digital humans are more possible than most think. With capable AI researchers helping, maybe for $10B, maybe in less than 10 years, on 50k H100s.
Isaak tweet media
English
111
282
2.5K
211.3K
Zack Dulberg
Zack Dulberg@pianozack·
@adam_louis52328 @petersavodnik Ok so you're saying that the overloading itself of general with specific is a unique kind of situation that contaminates the general version. So it's attempt some decontamination vs leave it be. To each their own? Anyway good point of convergence.
English
0
0
0
20
Adam Louis-Klein
Adam Louis-Klein@adam_louis52328·
@pianozack @petersavodnik No, it is overloaded. But the point is that the general term is mapped on to a specific term, excluding another specific term. Classical antisemitism and antizionism are opposites (+/-), but the term 'antisemitism' only captures one pole, even as it claims to be generic.
English
1
0
0
67
Peter Savodnik
Peter Savodnik@petersavodnik·
The "far left youth" are making a distinction without a difference. They like to believe that they're supremely tolerant, and they go to great lengths to illustrate that they're not antisemitic -- I attended a "seder" at a pro-Hamas encampment at Cal State Humboldt, for example -- but that doesn't make them not antisemitic. It makes them idiots, useful and otherwise, who traffic in the worst Jew-hating tropes and clichés, and have done more to make countless so-called progressive spaces inhospitable to Jews -- not just Zionists. What's more, they seem not to grasp that the inevitable nth degree of anti-Zionism is antisemitism: How is one to be sure the Zionist fever has been put out so long as there are any Jews (any potential Zionists) among us?
Seth Bannon@sethbannon

The far left youth are anti-israel but the least anti-semitic. The far right youth is anti-israel and the most anti-semitic.

English
12
15
111
6.2K
Adam Louis-Klein
Adam Louis-Klein@adam_louis52328·
@pianozack @petersavodnik @mr_james_c That’s why the term classical antisemitism is helpful. It draws attention to how ‘antisemitism’ in fact is anchored. We need to apply pressure and any leverage we have to get organizations to start centering antizionism.
English
1
0
1
53
Zack Dulberg
Zack Dulberg@pianozack·
@adam_louis52328 @petersavodnik @mr_james_c It's inevitable in the short term given its institutional, legal and cultural power. As long as it's around, I want to make the definition more transparent so it is less usable as a shield. That may change and you're right that naming antizionism is paramount.
English
1
0
0
47
Adam Louis-Klein
Adam Louis-Klein@adam_louis52328·
@pianozack @petersavodnik @mr_james_c It’s not clear to me why the continued use of ‘antisemitism’ is inevitable. In any case, I think saying antizionism—without anchoring it to classical antisemitism— is the most important thing, more important than not saying antisemitism.
English
1
0
1
55
Zack Dulberg
Zack Dulberg@pianozack·
No arguments there. But there is a tension in the present between specifying the details of antizionism and abstracting them away, which like you said is more of a future destination (assuming the damage has been sufficient to generalize the term as the Holocaust did for antisemitism). In the meantime there is a bit of a vacuum for naming the pattern - Jew-hate works, while antisemitism has more institutional purchase through IHRA etc. I don't think it can hurt to add clarification when antisemitism inevitably continues to be used to refer to abstract Jew-hate, and also switch to using Jew-hate and antizionism in your own speech.
English
1
0
0
45
Adam Louis-Klein
Adam Louis-Klein@adam_louis52328·
@pianozack @petersavodnik @mr_james_c Jew-hatred—or antizionism will organically become the new generic term, as antisemitism did in the past. Since virtually all Jew-hatred today is antizionism or embedded in antizionism there’s also no harm in using antizionism as a generic term.
English
1
0
1
56
Adam Louis-Klein
Adam Louis-Klein@adam_louis52328·
@pianozack @petersavodnik @mr_james_c That’s the destination rather than the starting-point. First people must see that antizionism is a hate movement, harms Jews, and is formally anti-Israeli racism. The case must not *rely* on anchoring it in ‘centuries of antisemitism.’
English
2
0
4
104
Zack Dulberg
Zack Dulberg@pianozack·
Agreed. It's just that part of that condemnation is missing if we ignore that antizionism also fits the 'abstract' or 'structural' definition of antisemitism, which is a very real definition. And they can't deflect that by saying they aren't classically antisemitic if we are very clear that we are not condemning them using that definition.
English
2
0
0
80
Zack Dulberg
Zack Dulberg@pianozack·
@adam_louis52328 @petersavodnik @mr_james_c But it remains usable by antizionists as camoflauge as long as it remains ambiguous. What is the harm in disambiguating? Antisemitism, defined as the generic structure of Jew-hate, is alive and well.
English
1
0
0
73
Zack Dulberg
Zack Dulberg@pianozack·
@adam_louis52328 @petersavodnik @mr_james_c The ambiguity needs to be made explicit or everyone is gonna keep talking past each other x.com/pianozack/stat…
Zack Dulberg@pianozack

This is all correct, but the reason people are so confused is because the term antisemitism is overloaded - it has two meanings. In computer science, overloading happens when different functions share the same name. This strategy can compress related concepts, but it comes at the cost of ambiguity. Harnessing this ambiguity and hiding behind overloaded terminology is a core mechanism of antizionism. In this case, antisemitism refers both to the abstract structure of jew-hatred (the term generalized after the Holocaust) but also maintains its reference to the specific form of 19th/20th century classical antisemitism. Antizionists appeal to the latter meaning to deflect accusations of the former. This pattern shows up everywhere once you see it. Free Palestine from Israel or of Israel? Resistance to being killed or resistance as killing? The occupation by the military in the West bank, or by the Jews in all of the land? Intifada as a general term for uprising, or as reference to specific violent uprisings in Israel? Many in the movement may even intend the more innocuous meaning. But the ability to switch back and forth between these meanings undetected is the camouflage strategy that makes antizionism so dangerous. We should therefore talk about and disambiguate overloaded terms like antisemitism, forcing people to use transparent definitions, so they can no longer hide behind them.

English
1
0
1
190
Adam Louis-Klein
Adam Louis-Klein@adam_louis52328·
@petersavodnik @mr_james_c These results are bc in fact those youths usually are not classically antisemitic and the term ‘antisemitism’ is anchored to classical antisemitism. Instead we could help them see that hatred of Israel is also wrong even if it isn’t anchored in classical antisemitism.
English
3
2
25
1.8K