Federico
@federicovaggi.bsky.social
F_vaggi on Twitter. Senior staff scientist at Google X, previously Amazon.
November 10, 2025 at 9:06 PM
Like this is a pretty egregious misreading of what Brooks actually wrote, it’s just that Brooks is a useful synedoche for “hopelessly naive centrist idiot with stupid ideas” that a humanities prof can feel justified in the kind of misreading an undergrad would be failed for in a first year class.
November 10, 2025 at 1:47 AM
Like this is a pretty egregious misreading of what Brooks actually wrote, it’s just that Brooks is a useful synedoche for “hopelessly naive centrist idiot with stupid ideas” that a humanities prof can feel justified in the kind of misreading an undergrad would be failed for in a first year class.
It is outright misreading what David Brooks actually said because it is a convenient foil though: www.nytimes.com/2020/06/18/o...
November 10, 2025 at 1:45 AM
It is outright misreading what David Brooks actually said because it is a convenient foil though: www.nytimes.com/2020/06/18/o...
Awww, they are so precious! Here is our oldest taking care of our middle kitty when she was spayed and here is the kitty trio on top of my wife's bump a few months ago!
November 4, 2025 at 2:54 AM
Awww, they are so precious! Here is our oldest taking care of our middle kitty when she was spayed and here is the kitty trio on top of my wife's bump a few months ago!
have you considered getting another cat for him to monch on? It's very effective. 2 more is even better.
November 4, 2025 at 2:24 AM
have you considered getting another cat for him to monch on? It's very effective. 2 more is even better.
We need a genius problem solver to deal with Issues w/Various Countries
October 15, 2025 at 3:02 PM
We need a genius problem solver to deal with Issues w/Various Countries
September 27, 2025 at 9:38 PM
Here is the relevant portion of the transcript (me with Gemini's help, so it might be slightly inaccurate, but the gist is correct).
Note he specifically calls out that dehumanizing/refusing to engage the other side is bad both morally, but, importantly, also in terms of political outcomes.
Note he specifically calls out that dehumanizing/refusing to engage the other side is bad both morally, but, importantly, also in terms of political outcomes.
September 26, 2025 at 4:59 PM
Here is the relevant portion of the transcript (me with Gemini's help, so it might be slightly inaccurate, but the gist is correct).
Note he specifically calls out that dehumanizing/refusing to engage the other side is bad both morally, but, importantly, also in terms of political outcomes.
Note he specifically calls out that dehumanizing/refusing to engage the other side is bad both morally, but, importantly, also in terms of political outcomes.
I thought this article was very bad, specifically because it was extremely sloppy with understanding statistics:
For example - here, the paper puts together 3 forecasts that are completely different. Nate Silver's forecast is extremely good (given the data), while the Princeton one is garbage.
For example - here, the paper puts together 3 forecasts that are completely different. Nate Silver's forecast is extremely good (given the data), while the Princeton one is garbage.
September 16, 2025 at 10:49 PM
I thought this article was very bad, specifically because it was extremely sloppy with understanding statistics:
For example - here, the paper puts together 3 forecasts that are completely different. Nate Silver's forecast is extremely good (given the data), while the Princeton one is garbage.
For example - here, the paper puts together 3 forecasts that are completely different. Nate Silver's forecast is extremely good (given the data), while the Princeton one is garbage.
Matt Levine here:
September 10, 2025 at 8:15 PM
Matt Levine here:
On phone, so cleaned this up very slightly with Gemini:
Informally: are there instances where it's very difficult to get a posterior distribution over the parameters, but, easy to get a posterior distribution over predictions? Trivial example - linear model with perfectly colinear features.
Informally: are there instances where it's very difficult to get a posterior distribution over the parameters, but, easy to get a posterior distribution over predictions? Trivial example - linear model with perfectly colinear features.
February 23, 2025 at 7:57 PM
On phone, so cleaned this up very slightly with Gemini:
Informally: are there instances where it's very difficult to get a posterior distribution over the parameters, but, easy to get a posterior distribution over predictions? Trivial example - linear model with perfectly colinear features.
Informally: are there instances where it's very difficult to get a posterior distribution over the parameters, but, easy to get a posterior distribution over predictions? Trivial example - linear model with perfectly colinear features.
December 16, 2024 at 3:01 PM
This is the precursor to the Nobel prize only parking spot at Berkeley
December 11, 2024 at 3:38 AM
This is the precursor to the Nobel prize only parking spot at Berkeley
Show him this dessert:
December 3, 2024 at 10:55 PM
Show him this dessert:
Return to tradition
November 18, 2024 at 5:52 PM
Return to tradition
I am so incredibly disappointed by the authors of this paper: they couldn't find a single arbitrary constant to set to 420?
November 18, 2024 at 5:26 PM
I am so incredibly disappointed by the authors of this paper: they couldn't find a single arbitrary constant to set to 420?
First post on Bluesky should definitely be cats
May 9, 2023 at 6:32 PM
First post on Bluesky should definitely be cats