Dan Kaminsky

128.7K posts

Dan Kaminsky banner
Dan Kaminsky

Dan Kaminsky

@dakami

We can fix it. We have the technology. OK. We need to create the technology. Alright. The policy guys are mucking with the technology. Relax. WE'RE ON IT.

Katılım Eylül 2007
414 Takip Edilen83.7K Takipçiler
Sabitlenmiş Tweet
Dan Kaminsky
Dan Kaminsky@dakami·
I'm increasingly thinking that every functioning system has two forms: The abstraction that outsiders are led to believe, and the reality that insiders actually and carefully operate. You don't incrementally learn a system. You eventually unlearn its necessary lies.
English
162
1.7K
5.7K
0
Cruxador
Cruxador@Cruxador·
@SwiftOnSecurity It's still bad writing if they don't even spend half a scene telling you why those precautions weren't taken.
English
2
0
2
0
SwiftOnSecurity
SwiftOnSecurity@SwiftOnSecurity·
Remember how you were watching that movie and you're like "this disaster was entirely preventable through a few obvious precautions" and you said "this is terrible writing."
English
16
369
1.6K
0
Locopells
Locopells@Locopells·
@SwiftOnSecurity Yeah mean like the way Jurassic Park only actually fails because it's sabotaged, not Life Finding a Way?!
English
2
0
9
0
mdowd
mdowd@mdowd·
@infosec_au @daveaitel In the past, there were some pieces of software I thought were harder to set up than find bugs in
English
3
5
29
0
shubs
shubs@infosec_au·
Half the reason why people aren't reverse engineering and finding vulnerabilities in SharePoint/(insert enterprise product here) is because of how annoying it can be to setup
English
13
30
271
0
Chelsea Klukas
Chelsea Klukas@chelscore·
He insists I wake up at 7:30am, even on weekends. Ignoring him just leads to drama. He is definitely not supposed to be on the nightstand or using tools.
Chelsea Klukas tweet mediaChelsea Klukas tweet media
English
4
1
102
0
Dan Kaminsky
Dan Kaminsky@dakami·
@gcampax Have you seen ngboost? Regression with explicit probability distributions. stanfordmlgroup.github.io/projects/ngboo… There’s a lot going on treating probabilities as first class citizens. But do you know of anything quantifying probability prediction error for various real world systems?
English
0
0
7
0
Giovanni Campagna
Giovanni Campagna@gcampax·
@dakami Most model probabilities are way off, and models are overconfident. There is a whole line of work on calibrating models so that output probabilities refect the actual probability of correctness.
English
1
0
15
0
Dan Kaminsky
Dan Kaminsky@dakami·
@blaktron I suppose. Inference is just running the balls through the pachinko machine and seeing where they land.
English
1
0
0
0
Chris Dupres 🇨🇦 🇺🇦
@dakami Thats literally what inference is, right? Having just enough data to reduce doubt of the conclusion below a determined level.
English
1
0
0
0
Brian🇲🇽🇮🇱
Brian🇲🇽🇮🇱@SpeakingBee·
@dakami @AlexShulepov7 @patio11 Out of bureaucratic inertia. The CDC and NIAID utterly failed every test COVID put to it and nevertheless its leadership gained power, prestige and money. The FDA screwed up constantly and gained power to obstruct even more next time.
English
1
0
2
0
Patrick McKenzie
Patrick McKenzie@patio11·
Largely agree with this taxonomy. There are (and I know this is surprising to many peers) a lot of people who are acutely sensitive to complexity/difficulty/mental effort required to get through things. We understand this in B2C conversion optimization and industry spends $$$$.
Ashish K. Jha@ashishkjha

When it comes to COVID vaccine, 4 types of folks 1st Desperately seeking vax 2nd Want vax. Won't jump through hoops 3rd May want vax. Have questions 4th Don't want vax We're about done vaccinating 1st 4th not that big Lets work on 2nd (make it easy), 3rd (answer questions)

English
8
16
135
0
Dan Kaminsky
Dan Kaminsky@dakami·
@SpeakingBee @AlexShulepov7 @patio11 Why? Out of spite? You think the flu shot is going to stay the same? Do you imagine the vaccine production lines will be shut down? Chronic lower respiratory disorders were the number three killer of humans before COVID came along, which is btw a chronic lower resp disorder.
English
1
0
2
0