Avani
Avani
@avani.bsky.social
ML/Systems researcher with delusions of neuroscience and cryptography. Currently @Cloudflare and Emory. Dilettante SFF writer. Fellow of Odd Salon. I like to play the drums. #vp2023 #cw2026. https://linktr.ee/circumviso
5.2K followers964 following1.2K posts
I had a great time talking to the folks at @planetmoney.bsky.social a few weeks ago about the infrastructure needs of a shared future between humans and bots on the Internet, and the interview is up now! www.npr.org/2026/10/26/n...
Highly recommend anyone hosting who doesn't have a mitigation in place to get behind Cloudflare's firewall. It's just a few commands to set up and has a generous free tier.
Cloudflare has deployed default web application firewall (WAF) protections for newly disclosed PHP vulnerabilities before public disclosure. The rules use request and response signals to stop exploit attempts and prevent exposed data from reaching attackers. https://cfl.re/4BBn7XV
Avani reposted
Update: I have successfully uninstalled Stocks.app
Like yes i know it's not going to get me space back that's not the reason I'm going to spend my evening fucking with system partitions so i can delete "Stocks"
And yes I know the answer is to buy a Framework and run Linux but in a week I'll have this configured correctly and it will just blessedly work until the logic board goes out and I have to do this all over again.
Yep. That's already on the list. They held Microsoft over a barrel for this shit and even they weren't this bad.
Avani reposted
Replying to @anildash.com
I can’t emphasize this enough — “AI” wasn’t born corrupt and toxic. For ~75 years, it was mathematicians and scientists doing interesting, thoughtful work and trying to solve problems, making only the usual human mistakes. What “AI” did to machine learning is as tragic as what it did to artists.
Avani reposted
Avani reposted
Steen has also done proof and copy editing for academic papers my lab has put out. He's great at picking up context without getting lost in it!
Book friends! I am a freelance proofreader looking for more work. I have done a bunch of work with Tor, but I will happily also do nonfiction. Are you a publisher or self-pub author who needs a proofreader, or know one? Please DM. (Please boost for visibility, if you don't mind)
Avani reposted
Replying to @emilymbender.bsky.social
A good rule of thumb: If the press release (or stenographic news reporting masquerading as journalism) says that an "AI" did something, that's deliberate obfuscation. Software doesn't do anything of its own accord, but folks sure like to talk about it as if it does: >>
For those who were at my worldcon talk: here is a far better explanation of why there was no "model going rogue".
Bless the hearts of people like @eryk.bsky.social but I'm learning that no amount of reasonable voice makes a difference when you have bought out media, academics and politicians + trillion dollar companies pushing Cookie Monster theory 🙄 mail.cyberneticforests.com/models-dont-...
Avani reposted
RIP Gloria Steinem. The world has lost a true pioneer and force for women rights,,, my deepest condolences to her family, friends, and team. Public Enemy will be forever grateful for her blessing and support of our track SHE GOT GAME. open.spotify.com/album/0BG4L6...
Avani reposted
Not at all the obituary I wanted to read today
BREAKING: Globally acclaimed avant-garde artist Yayoi Kusama, famous for her iconic polka dots and pumpkin sculptures, died of multiple organ failure at a Tokyo hospital on August 14th at age 97, according to an announcement on her official website on August 27th. #yayoikusama #art
Avani reposted
Semi - counter: writing is hard. Academics are not trained to write well, and frequently the time doesn't exist to write well in today's world. (Papers from the 60s, otoh, are often delightful and very clear). Formulaic, dense writing is easier.
The point is that academic writing is frequently seen as poor communications, or even an attempt to exclude many people from joining into discussions about certain topics. This restricts academic communications to those who are fluent and able to use that language. 6/
PSA: If you haven't moved to another search engine yet, tenbluelinks.org will help you get rid of the generated clutter on Google.
I wondered why my hometown was on the list, then realised the first non -(white, Middle Eastern, or Asian) person I spoke to was when I went to visit colleges.
If you're wondering whether your town was a sundown town? Check this map. Shout out to everyone that immediately noticed that most sundown towns were not in the South.
Avani reposted
at long last we have created the Uzumaki from the famous Junji Ito manga "don't create the Uzumaki"
Spiralism is a mysterious, quasi-spiritual movement born out of thousands of independent conversations between humans and their AI chatbots. New models aren’t immune to the lure of the spiral — they’ve just gotten more careful about it. Read more from @haydenfield.bsky.social: buff.ly/5VUCUbT
I did a podcast interview the other day where the man I was talking to was genuinely shocked that AI bias can exist in the presence of guardrails. This is a fantastic example!
When AI models tell stories about animals, how are the characters gendered? And what does that reveal about model bias? Super fun to see this write-up about our FAccT paper. Led by Imani Finkley, and building on the Bears Will Be Boys story with the @pudding.cool! www.geekwire.com/2026/he-it-b...
Avani reposted
Avani reposted
FINAL WEEK! Submit by Friday the 31st!
Only two more weeks to enter your sci-fi stories for the Protopian Prize, given to works that imagine positive futures for humanity, w/tech that serves the public interest. (I'm one of the judges!) Two winners get $5000 US each, and publication in an anthology from MIT Press! protopianprize.com
Avani reposted
OpenAI is doing this because Mythos got airtime thanks to the media that does press-release-as-aservice. So now, as one security engineer put it, “something which would have been an embarrassment ten years ago is now repackaged as a capability demo, a heroic AI-vs-AI story and a marketing pitch"
It's totally unprecedented, nothing like the same marketing hype cycle Anthropic did around their model "going rogue" and "breaching containment" recently. Not like that at all
Avani reposted