Brian Lindauer (@[email protected])

Security, data, machine learning, food, urbanism // Lead teams at @duosec, but these words are mine, not my employer’s.

[François Chollet @fchollet]

Don't score AI using tests designed for humans. In particular because, with humans, the default assumption is that *they haven't already seen* the content you're giving them. With a LLM, the default assumption should be that, *if it's on the Internet, it's already been memorized*

[Brian Lindauer (@[email protected]) @lindauer]

@vboykis 3. Starting from a specific domain is a winning strategy (for me, security). What I didn’t know about data science work then: problem definition, communication, and organizational navigation are at least as hard and important as the rest of it. 3/3

[Brian Lindauer (@[email protected]) @lindauer]

@vboykis The bits of my experience that are still applicable are: 1. Software engineering skills are super valuable, 2. Stats/algo fundamentals are needed to progress past a certain point (and this is part of why AutoML isn’t coming for your job), 2/

[JosH100 @josh_wills]

This is the best thing I have ever read about the transformative power of schemas for data teams; I haven't enjoyed a blog post this much since "I ❤️ Logs"

Well done @aerialfly 👏👏👏

https://t.co/XXlPqZ0Pl9

[Brian Lindauer (@[email protected]) @lindauer]

@wendynather @boblord To add to @wendynather’s reply, the Trusted Access Report population is @duosec customers – largely business users. For data about the general population, you could check out our State of the Auth report. https://t.co/EI6Q9NQS8z

[Brian Lindauer (@[email protected]) @lindauer]

Verified Duo Push raises the bar on attackers, but also adds friction for users. I'm excited to share @duosec's new Risk-Based Authentication feature, which invokes it at just the right times. Easy 🤝 Effective
https://t.co/Oodqg1Rv3I

[Brian Lindauer (@[email protected]) @lindauer]

@jondmatthews I've been wondering the same thing. This is the only information I've found so far. In short, choose Pfizer if you're at higher risk for myocarditis. Otherwise, mixing vaccines might help but probably doesn't matter. https://t.co/4F1KXuBCqY

[Jeremy Jordan @jeremyjordan]

machine learning systems often require deploying a decent bit of infrastructure. i wrote a blog post about terraform and how you can use it as a machine learning engineer to manage your infrastructure, let me know what you think! https://t.co/qArrrpdigd

[Brian Lindauer (@[email protected]) @lindauer]

@kpelaez10 @duosec Welcome to the team!

[Brian Lindauer (@[email protected]) @lindauer]

Folks in the tech industry looking for a concrete action you can take today, here's a place where we hold more power than most. Ask yourself what you are doing to protect your users' privacy.

[briankrebs @briankrebs]

1/ So you go shopping for a PIV card reader, because the US govt gave you one and you're curious to look at what's on it. You settle for this "DOD military USB common access smart card reader," because it's compatible with Mac OS. Cool! Only $15! What a bargain!

[vicki @vboykis]

And this embedding, is it in the multidimensional vector space with us now?

[Brian Lindauer (@[email protected]) @lindauer]

@sedward5 I know it's less fun than adding a resistor, but the usual low-tech solution in our house is a piece of tape over the speaker.

[Brian Lindauer (@[email protected]) @lindauer]

@rharang Case in point
https://t.co/rWfo4GvIyq

[Jamal Greene @jamalgreene]

What speech "the law" protects depends not only on which laws one is referring to (US? Germany? etc.) but also on a bunch of other contextual factors: is it a subsidy? are you an employee? is it a limited forum? gov't speech? broadcast media? time/place/manner reg? & so on. (1/3)

[Brian Lindauer (@[email protected]) @lindauer]

@MaggieEngler Congrats, Maggie!

[Brian Lindauer (@[email protected]) @lindauer]

@adamjodonnell You hit all your OKRs this quarter, then?

[Brian Lindauer (@[email protected]) @lindauer]

This article is a fascinating interdisciplinary summary of why user choice is not always a great optimization objective, and what we can do about it.

[Eric Colson @ericcolson]

Good articulation of why I feel humans are not qualified to make decisions on behalf of the earth. (thx @peteskomoroch for the find).
What Does it Mean to Give Someone What They Want? The Nature of Preferences in Recommender Systems by Luke Thorburn https://t.co/kKYwmKRrVQ