Nirit Weiss-Blatt, PhD

2.2K posts

Nirit Weiss-Blatt, PhD banner
Nirit Weiss-Blatt, PhD

Nirit Weiss-Blatt, PhD

@DrTechlash

Communication Researcher, analyzing the tech discourse. Book Author: The TECHLASH. Substack: https://t.co/4SJJhqrzXn Signal: DrTechlash.16

Cupertino, CA Katılım Kasım 2020
210 Takip Edilen6.9K Takipçiler
Nirit Weiss-Blatt, PhD
Nirit Weiss-Blatt, PhD@DrTechlash·
On April 12, after the firebomber attack, Oliver Habryka, who runs LessWrong, said that "discussing or advocating violence is not banned." On April 14, a LessWonger actually advocated "hundreds of simultaneous assassinations to wipe out of the entire chains of command of US executive branch, US intelligence, and all US AI companies." "One of the best ways to increase temperature of the conflict is to actually be violent. I have some aesthetic preference in favour of violence and bloodshed." Is the next Unabomber/firebomber/potential assassin gonna come from LessWrong? Hopefully not.
English
0
0
1
304
Will Manidis
Will Manidis@WillManidis·
If you can’t see where this is all going I can’t help you
English
17
16
419
58.6K
Nirit Weiss-Blatt, PhD
Nirit Weiss-Blatt, PhD@DrTechlash·
A new study, "AI going rogue? An integrative narrative review of the tacit assumptions underlying existential AI risks," finds that the x-risk literature rests on highly speculative, anthropomorphic, and unsubstantiated scenarios, while neglecting socio-technical realities.
Nirit Weiss-Blatt, PhD tweet media
English
5
5
17
857
Nirit Weiss-Blatt, PhD
Nirit Weiss-Blatt, PhD@DrTechlash·
It was posted about Anthropic's blackmail study, but it's relevant to Palisade's study too. The headline frames this as autonomous self-replication, but the experiment itself shows agents completing a CTF-like exploitation task inside a highly controlled, researcher-designed environment. x.com/drtechlash/sta…
English
1
0
2
168
Sebastien Meunier
Sebastien Meunier@sbmeunier·
Ridiculous. It saved the entire LLM infrastructure, the weights, the executables, scripts and configuration files into the target host 🤣 installed all that in the file system, executed the LLM and gave it a prompt. Yeah totally a real world attack scenario 🤣
Palisade Research@PalisadeAI

Over the past year, AI agents have learned how to self-replicate. In our test environment, an agent hacks a remote computer and copies itself onto it. Each copy then hacks more computers, forming a chain.

English
1
0
1
381
Anthropic
Anthropic@AnthropicAI·
We started by investigating why Claude chose to blackmail. We believe the original source of the behavior was internet text that portrays AI as evil and interested in self-preservation. Our post-training at the time wasn’t making it worse—but it also wasn’t making it better.
English
324
471
5.1K
4.4M
Anthropic
Anthropic@AnthropicAI·
New Anthropic research: Teaching Claude why. Last year we reported that, under certain experimental conditions, Claude 4 would blackmail users. Since then, we’ve completely eliminated this behavior. How?
English
552
804
9.2K
1.5M
Nirit Weiss-Blatt, PhD
Nirit Weiss-Blatt, PhD@DrTechlash·
Update about AI doom YouTube channels: More than a quarter of a billion views.
Nirit Weiss-Blatt, PhD tweet media
English
0
1
1
174
Nirit Weiss-Blatt, PhD
Nirit Weiss-Blatt, PhD@DrTechlash·
What 10 Studies Reveal About AI Panic in the Media 🧵
Nirit Weiss-Blatt, PhD tweet media
English
2
7
19
2.5K
Mapping AI
Mapping AI@mapping_ai·
Who actually shapes AI policy in the U.S.? We mapped 1,812 entities: 745 people, 918 organizations, 2,925 relationships. Frontier Labs, AI Safety orgs, Think Tanks, Government, VCs, and more. mapping-ai.org
Mapping AI tweet media
English
25
367
1.3K
291.7K
Nirit Weiss-Blatt, PhD
Nirit Weiss-Blatt, PhD@DrTechlash·
@mapping_ai Impressive work. But it lacks many publicly known connections that are not mentioned here. See, for example, the Coefficient Giving network. Even limited to AI policy, there should be at least 30 organizations, not 14.
Nirit Weiss-Blatt, PhD tweet media
English
2
0
13
575
Nirit Weiss-Blatt, PhD
Nirit Weiss-Blatt, PhD@DrTechlash·
I think we should add some definitions to your "investigative reporting" discussion: "public interest" vs. "interesting to the public". According to the National Union of Journalists, the "public interest" includes, for example: 1. Exposing a crime or a serious misdemeanor; 2. Protecting public health and safety; 3. Exposing misuse of public funds or other forms of corruption by public bodies; 4. Preventing the public from being misled by some statement or action of an individual or organization. While "interest to the public" is "what the public finds interesting," which can be pure entertainment, private details about individuals (like gossip), with the goal of generating clicks. This test helps determine if invasion of privacy was actually necessary.
Nirit Weiss-Blatt, PhD tweet media
English
0
0
1
52
MTS
MTS@MTSlive·
The Balaji Srinivasan and Taylor Lorenz episode @balajis and @TaylorLorenz joined us yesterday to discuss the rise of the creator economy, human-only social media, internet freedom, age verification laws, the legacy media vs. tech beef, and more. 00:00 Introduction 01:00 Rise of the content creator 06:55 Are identity verification laws anti-internet freedom? 13:00 Decentralized proof-of-human systems 22:33 Digital borders & the evolution of telepresence robotics 26:12 The global free speech recession 47:33 Grokipedia vs Wikipedia 1:15:24 Investigative reporting vs privacy, ethics of non-consensual disclosure
English
6
4
72
9.5K
Netflix
Netflix@netflix·
Michael B. Jordan, Juno Temple, and Tracy Morgan in the booth for SWAPPED
English
60
262
3.1K
211.7K
Nirit Weiss-Blatt, PhD
Nirit Weiss-Blatt, PhD@DrTechlash·
What 10 Studies Reveal About AI Panic in the Media -Part 2- A media-criticism discussion of what those studies miss: The organized creator/influencer ecosystem now distributing AI panic beyond traditional journalism. aipanic.news/p/what-10-stud…
Nirit Weiss-Blatt, PhD tweet media
English
2
0
5
494
Nirit Weiss-Blatt, PhD
Nirit Weiss-Blatt, PhD@DrTechlash·
What 10 Studies Reveal About AI Panic in the Media -Part 1- A literature review of 10 studies on AI media coverage.
Nirit Weiss-Blatt, PhD tweet media
English
2
2
9
1.4K
Nirit Weiss-Blatt, PhD
Nirit Weiss-Blatt, PhD@DrTechlash·
@EpistemicHope You've read Yud's stuff and found it good. I read it and found it completely unconvincing. Perhaps I wasn't the targeted audience...
Nirit Weiss-Blatt, PhD tweet media
English
1
0
1
43
Eli Tyre
Eli Tyre@EpistemicHope·
For my calibration, do others (aside from Critch) think that LessWrong has this problem? For those of you who agree, can you share the strongest example of this problem that comes to mind?
Andrew Critch (🤖🩺🚀)@AndrewCritchPhD

I'm not sure why it's so hard for LessWrong people to know what we're talking about when we say "Y'all are too into violent rhetoric, maybe cut it out?" I think they need help. I mean this non-bitingly and non-sarcastically. I think the internet needs to explain to the more receptive members of the LessWrong community why and how this is a problem. Maybe if we do that enough, we'll get through? E.g., I think Ryan Greenblatt is *not* needlessly violence-themed in his writing, yet also genuinely doesn't know what I'm talking about when I complain about it. With the recent violent attacks on Sam Altman, and a lot of reactions like "Wtf AI safety people, tame it down!", I wonder if some more actual good-faith messages could help. Like: "No seriously, we're not trying to be mean about this, it's just actually a problem how much your community promotes violent memes in connection with AI and AI safety. Can you please try to understand this and then explain it to your friends?"

English
16
1
45
6.7K

Keşfet

@AnthropicAI @AndrewOrlowski @mapping_ai @elonmusk @BarackObama @taylorswift13 @cristiano @BillGates