Views are my own, hopefully constantly updating
Joined May 2011
- Tweets617
- Following1.9K
- Followers1.8K
- Likes14.1K
aaron horowitz retweeted
I love @random_walker and @sayashk's idea that nuclear weapons are the worst possible analogy for AGI. They are different in almost every important respect and if you are reasoning based on this analogy you are very likely to reach bad conclusions.
aaron horowitz retweeted
My sense is that Trump 2.0 is heavily influenced by the Travel Ban Experience, which taught Trump that the Court will allow him to start off with the most extreme version of something and then gradually adjust its scope and rationale until the Court says "that's good enough."
aaron horowitz retweeted
Be the ACLU’s first Director of Web Strategy and innovate where it matters.
⬇️
Apply Below!!
NEW JOB: DIRECTOR OF WEB STRATEGY
The @ACLU is hiring for a *new* role destined to be one of the most important digital jobs this cycle. A real chance for digital innovation x advocacy. All candidates encouraged to apply!!
NY salary is $155k
aclu.org/careers/apply/?job=…
aaron horowitz retweeted
I don't think I'll ever get over the anti immagrant crowd cheering for the guy from South africa who came here illegally, made a fortune through government subsidies then bought an election so he could take away their benefits.
And Marissa (whos amazing) will be your manager too! A truly incredible opportuntity
This Post is from an account that no longer exists. Learn more
aaron horowitz retweeted
Hi! I'm hiring a Research Engineer to join my team at Google DeepMind for the year. You'd be working with a great, interdisciplinary team on AI evals. Please share if you know anyone who might be interested!
boards.greenhouse.io/deepmin…
Note: this is a fixed-term, 12 month position
aaron horowitz retweeted
In an election this close, every voter that the @ACLU adds back to the rolls matters.
aaron horowitz retweeted
This is incredible - each of these cases are AI scams that have been alarmingly normalized in the past couple years (including DoNotPay, a "robo-lawyer"; several AI enabled financial/business services, etc).
Glad to see all of this being finally called out!
Wow. Huge push by the FTC to crack down on deceptive AI.
ftc.gov/news-events/news/pre…
aaron horowitz retweeted
I really love this - it captures what most frustrated me when I took this class. Some problems are easier to formally model - these are the scenarios in which optimization methods "work". But there's so many other types of problems where we're pretty much just fooling ourselves.
Caveat Emptor: Mathematical optimization is always premised on approximate modeling. "Optimal solutions" will hence often be absurd. argmin.net/p/the-cost-of-mod…
aaron horowitz retweeted
This is such a unique opportunity for anyone working at the intersection of CS, policy & law.
Please come and join us here at Berkeley:
eecs.berkeley.edu/connect/fa…
aprecruit.berkeley.edu/JPF04…
Sep 16th application deadline.
UC Berkeley "seeks applicants for four tenure-track (Assistant Professor) positions and one tenured (Associate or Full Professor) position in the area of “AI, Inequality, and Society” (AIIS)...Areas of interest for the AIIS Cluster cut across disciplinary boundaries and include, but are not limited to: (i) employment, (ii) algorithmic discrimination, (iii) generalized surveillance, and (iv) data, information, and markets."
Why does bigquery **default** to case sensitive schemas? I get they are cool and want to implement the pipe in SQL, but get the basics right for g-ds sake
aaron horowitz retweeted
The @ACLU‘s Voting Rights Project is hiring a staff attorney or senior staff attorney! This is a critical time for our democracy. Come join us and the fight to protect voting rights! (And please share widely!)
aclu.org/careers/apply/?job=…
aaron horowitz retweeted
🚨Last year, South Carolina defended its racist U.S. House map by claiming they were only gerrymandering against Democrats, not Black people.
Seriously, they said it out loud.
Now the state's high court will decide if that's ok. Spoiler: it's not, no matter what party you are.
NEW: We're suing to end partisan gerrymandering in South Carolina.
See y'all in court. aclusc.org/en/press-releases…
ALT "See y'all in court. We're suing to end partisan gerrymandering." Text appears over blue and red boxes on top of a cracked black and white background.
aaron horowitz retweeted
All of my friends who think through the sociology and economics of software talk about how breakdowns like today's are a matter of when, not if. It's gonna happen. It's like monocultures in agriculture or relying on a single supplier or piece in mass production (see airbags).
aaron horowitz retweeted
This feature of @github has to be one of the biggest operational security risks.
Why when you're searching to add people do a repo does it search the entire corpus of users?!
Show me the people in my org first!
aaron horowitz retweeted
Excited to see the continued expansion and maturity of the "studying up" literature!
How can centering *systems of power* help us better understand AI design and adoption challenges in the public sector?
In our upcoming #CSCW2024 paper, we draw on Nader's 1972 call to *study up* – shifting our gaze upwards to examine those with power to shape public sector AI 🧵
ALT Screenshot of CSCW 2024 research paper titled Studying Up Public Sector AI: How Networks of Power Relations Shape Agency Decisions Around AI Design & Use
aaron horowitz retweeted
This chart is a great rebuttal of the view that RAG will solve hallucination. In the legal domain, it turns out that it does improve precision somewhat, but at the expense of recall. Really great paper by Stanford & Yale researchers assessing legal AI. dho.stanford.edu/wp-content/…
ALT The image is a bar chart comparing the proportion of hallucinated and incomplete answers across different legal research tools. Hallucination: Lexis: ~0.2 Westlaw: ~0.3 Practical Law: ~0.3 GPT-4: ~0.4 Incomplete: Lexis: ~0.25 Westlaw: ~0.3 Practical Law: ~0.6 GPT-4: ~0.1 Caption: Comparison of hallucinated and incomplete answers across generative legal research tools. Hallucinated responses are those that include false statements or falsely assert a source supports a statement. Incomplete responses are those that fail to either address the user’s query or provide proper citations for factual claims.