Erick Galinkin
@erickgalinkin.bsky.social
Threat actor, artificially intelligent I do bad things to good models World’s #1 Regex Enjoyer
"𝑃ℎ𝑜𝑡𝑜𝐷𝑁𝐴 is unreliable for the detection of illicit content: it is easy to incriminate someone by sending them false content with a hash value close to illicit content (a false positive) and to avoid detection of illicit content with minimal modifications to an image (a false negative)."
Deirdre Connolly¹ ²@durumcrustulum.com · 7mo ago
This composition had eluded me for nearly a decade, but yesterday I was finally able to capture it. A train rolled through just as the Duquesne Incline was crossing over the tracks in #Pittsburgh, creating a picture perfect winter scene. It was worth the hours and hours I've waited over the years.
Is the point of mathematical research to know *what* is true or to know *why* it's true? Emily Riehl lays out a strong argument - and what it has to do with AI - along great advice for the field in @science.org www.science.org/doi/10.1126/...
Mathematics is hard for mathematicians to understand too
At a recent conference on mathematics in the age of automated proofs, mathematician and Fields Medalist Akshay Venkatesh presented “How do we talk to our students about AI?'' He quoted an email he'd r...
science.org
An issue we're seeing at all levels of university is that many students are simply refusing to do *anything*. They aren't reading the syllabus, aren't following assignment guidelines, aren't engaging with material, ignoring deadlines. And this might seem like old news, but it truly has ramped up.
HOPE has been banned from St. John's University. www.2600.com/content/hope...
HOPE CONFERENCE BANNED BY ST. JOHN'S UNIVERSITY | 2600
2600.com
There's some really big caveats to this. A thread.
Eric Geller@ericjgeller.com · 11mo agoNew: Google says it has discovered at least 5 malware families that use AI to rewrite their code and generate new capabilities on the fly, suggesting AI-powered malware is finally starting to take off. cloud.google.com/blog/topics/... Report also has interesting stories about state actors' AI use.
If you are white, not racist, not a supporter of ethnic cleansing, but are struggling to pay bills, or don't know how you will afford food next month... First of all, I am genuinely sorry for the stress you are dealing with. Truly. Second, this stress was placed on you intentionally. By racists.
Mekka Okereke@mekka.mekka-tech.com · last yr.We managed to avoid what they call "phase 2" in 2020, just to squander our near miss and fail in 2024.🤦🏿♂️ They put RFK jr in charge for a reason: to do exactly what he's doing. Promises kept. Check the date. Yes, this was 4 months before the 1st recorded US Covid death. They wrote this years ago.
I almost spit out my lunch when I got to HUMBERT® replacement.ai
Replacement.AI
Humans are no longer necessary. So we’re getting rid of them.
replacement.ai
This. Also, most of the time you get a “PhD-level”response, it’s just a paraphrase of Wikipedia which you can just… go to. Like, it’s right there. “But it can explain it in simple language” yeah man, check out Simple English Wikipedia, shit rocks.
Katie Mack@astrokatie.com · last yr."PhD-level experts in your back pocket" is a completely nonsensical description of AI but a pretty good description of social media if you follow the right people
Perplexity and Truth Social announced a partnership yesterday, both with exuberance. So, I cancelled my account just now, after figuring out a way to save my threads—they don't provide a way to do so and use that as a scare tactic to prevent cancelling (ref img 1). I also left them a "why" note.
When I was talking about “Fallout in real life”, I meant cool robot shit and Pip Boys, not fucking Cazadores.
Keith Taylor@mrkeithtaylor.bsky.social · last yr.Hot Wasps in Your Area
The FTC is seeking public comment about gender affirming care in order to go after providers and restrict access. Share your stories instead of how this care has helped you and how it is life saving. We need to counteract their false narratives. Please share! www.regulations.gov/docket/FTC-2...
Regulations.gov
regulations.gov
Why is Cathy every SYNner’s favorite comic strip? She’s always saying ACK
IT'S LIVE
Hulk Hogan, Trump idol and enemy of the free press, dead at 71
An anti-eulogy by Josephine Riesman
thehandbasket.co
Josephine beckbook.com Riesman@josie.zone · last yr.My Hulk Hogan anti-eulogy has been filed and will be running on @marisakabas.bsky.social's wonderful "The Handbasket" this evening Stay tuned
It’s probably hard to comprehend how huge this is if you haven’t rotted your brain on cybercrime forums for years but this is just… wow.
Allan “Ransomware Sommelier” Liska@ransomwaresommelier.com · last yr.I swear, I am going to have to add “new ransomware baddie arrested” to my Emmy-nominated series “newly-registered ransomware domain” series. Alleged administrator of Russian forum XSS, arrested by French police and Europol. via @reuters.com
important context for headlines like this: Altman also runs a company (Worldcoin) that he claims will solve this problem he also created
CNN@cnn.com · last yr.OpenAI CEO Sam Altman says the world may be on the precipice of a "fraud crisis" because of how artificial intelligence could enable bad actors to impersonate other people. Read more: cnn.it/3GVwPg7
if you have ever enjoyed one of my pieces, from the political jeremiads to the deep dives into sandwich history, pls become a paid subscriber! you get more content, i make rent, it's a win-win buttondown.com/theswordandthesandwich/buy
Stripe Checkout
buttondown.com
Tal Lavin@swordsjew.bsky.social · last yr.why you should consider becoming a paid subscriber to my newsletter: 1. support two journalists (me and @davidswanson.bsky.social) for the price of 1 latte per month buttondown.com/theswordandt...
Very excited to hear @kenwhite.bsky.social’s exasperated sigh followed by “Well, Josh…” on the next episode of Serious Trouble
Reuters Legal@legal.reuters.com · last yr.BREAKING: Donald Trump sues Dow Jones, News Corp, Rupert Murdoch, WSJ reporters for libel, court records show reut.rs/3TNpulG
I armed HF's "tiny-agents" w/mlx-community/Qwen2.5-7B-Instruct-1M-8bit & MCP servers providing access to DDG search, Wikipedia, and general fetch capability. I asked it a question I knew the answers to. Here's the session output: codeberg.org/hrbrmstr/gis...
My jokes aren’t setting the world on fire, but I feel like there’s an element of truth in this one.
Pro tip: you can basically work out as much as you want by watching videos of people lifting weights.
Joe Stieb@joestieb.bsky.social · last yr.You basically cannot.
in a crowded field this has to be one of the stupidest things ive read about LLMs and it astonishes me that this quote is from a humanist, a historian! by definition, what these tools cannot do is original novel research of the kind we do in monographs! www.newyorker.com/culture/the-...
Look, I have a lot of problems with age verification. I think it’s a camel’s nose designed to Balkanize the Internet by claiming everything is obscenity. But even approaching it in its best case, it is going to end up being a way to attach identity to every website visit. This is a government dream.
A few months ago, Nature published how-to guide for using ChatGPT to write your peer reviews in 30 minutes. This is, of course, a horrible idea. Here’s my response with @jbakcoleman.bsky.social .
AI, peer review and the human activity of science
When researchers cede their scientific judgement to machines, we lose something important.
nature.com
