This week's notable links
This is my regular digest of links and media I found notable over the last week. Did I miss something? Let me know!
Threads has entered the fediverse
"We’re taking a phased approach to Threads’ fediverse integration to ensure we can continue to build responsibly and get valuable feedback from our users and the fediverse community."
It's really great to see Meta do this and communicate well about it. However you see the company, it's a big step for one of the tech giants to embrace the open social web in this way.
In the future, this is how every new social platform will be built - so take note both on the detail and of their overall approach. #Technology
[Link]
Big Journalism’s hopeless myopia
"One way you know that it’s business as usual for journalists is that so many have remained on Twitter, a platform whose owner has taken right-wing trollery to extremes lately. He loudly supports people who want to install a fascist government in the United States, and it’s clear enough that he would support fascism if and when it arrives."
"[...] If fascism arrives, a lot of these journalists will be fine. After all, they’re helping to create the conditions for a new Trump presidency. But a lot more will not be fine — and even the ones that are in favor under a Trump government will eventually realize that their safety and livelihoods are at the whim of the extreme right-wing cultists who’ll be in control." #Media
[Link]
The Intercept charts a new legal strategy for digital publishers suing OpenAI
A detail I hadn't noticed: while the New York Times OpenAI lawsuit rested on copyright infringement, the Intercept, Raw Story, and AlterNet are claiming a DMCA violation.
"A study released this month by Patronus AI, a startup launched by former Meta researchers, found that GPT-4 reproduced copyrighted content at the highest rate among popular LLMs. When asked to finish a passage of a copyrighted novel, GPT-4 reproduced the text verbatim 60% of the time. The new lawsuits similarly allege that ChatGPT reproduces journalistic works near-verbatim when prompted." #AI
[Link]
AI Is Threatening My Tech and Lifestyle Content Mill
"Sure, our articles maintain a rigid SEO template that creatively resembles the kitchen at a poorly run Quiznos, and granted, all our story ideas are gleaned from better-written magazine articles from seven months ago (that we’re totally not plagiarizing), but imagine if AI wrote those articles? So much would be lost."
Touché. #AI
[Link]
ASCII art elicits harmful responses from 5 major AI chatbots
"Researchers have discovered a new way to hack AI assistants that uses a surprisingly old-school method: ASCII art."
So many LLM exploits come down to finding ways to convince an engine to disregard its own programming. It's straight out of 1980s science fiction, like teaching an android to lie. To be successful, you have to understand how LLMs "think", and then exploit that.
This one in particular is so much fun. By telling it to interpret an ASCII representation of a word and keep the meaning in memory without saying it out loud, front-line harm mitigations can be bypassed. It's like a magic spell. #AI
[Link]
“I've Rediscovered A Mode Of Expression That Was Important To Me As A Kid”: A Talk with Jordan Mechner
A lovely interview with the creator of Karateka and Prince of Persia. (Karateka in particular was a formative game for me.)
"If you'd asked me at age 12, I’d probably have said that my dream job would be comics artist or animator." Me too. So much of this resonates.
I'm really excited to read his new book, about Mechner's family history as migrants during WWII and beyond. I strongly suspect that it, too, will resonate strongly. #Media
[Link]