- Today in Tabs
- Posts
- ChatGPT? Yeah You Know Me.
ChatGPT? Yeah You Know Me.
Are the robits coming for your jerb?
December is the month of year-end retrospectives and judging by how many retrospectives landed on the first weekend of December, none of us can wait for this year to end. The OED picked “goblin mode” as the word of the year, despite it being two words and also a hoax. But what’s more goblin mode than that? Film directors picked their new 100 greatest films of all time for the BFI, but shanked it for the fortieth straight year:
Brian Feldman collected Twitter trending topics all year and built whatshappening.online, a timeline and calendar of some of the nonsense that happened this year and how Twitter’s now-axed curation team chose to summarize it. I like the timeline view but I only made it to early February (through West Elm Caleb and the Sexy Lentil, both somehow this year) before I had to take a break. Curbed collected 39 reasons to love New York right now, which is a normal number of reasons for anything. The video for number one is spectacular. Ryan Broderick compiled the year in Garbage Day, using data to validate many of the same opinions about running a newsletter that I arrived at solely by way of ~vibes~. And Sabrina Imbler remembered the deep-sea octopus Mother of the Year 2014 in a truly gorgeous Atlantic essay / book excerpt. Imbler also talked to Delia Cai about crab memes, which is a hat trick of Tabs favorites.
Each essay in your book, How Far the Light Reaches, centers around one sea creature that then serves as a symbol for different stories from your personal experience. In terms of the writing process, I was curious if you picked the animals first, or if you had to “cast” the right one for each essay.
It depended on the essay. With the octopus essay, I knew that I wanted to write about that octopus that had a very clear parallel with my life. For the yeti crab one, I wanted to write about queer nightlife, and I was thinking about how, when I’m in a club, it’s dark, and we’re all here together, and you can’t really see. And I was like, oh, it’s like the deep sea. What’s a weird-looking creature that might be an interesting subject? So I found that yeti crab. Then once I started writing it, I read this paper about this one species of yeti crab that waves its hands and dances, and I was like, I’m a genius. I couldn’t have predicted it.
AI… AI… AI… Will Always Love Yoooooouuuu
There were only two kinds of social media feed this weekend: either you were flooded with Lensa selfies, or you were flooded with ChatGPT screenshots. Lensa is a paid app that makes it easy to use Stable Diffusion’s AI art tools to make your selfies look hot in a variety of art styles. ChatGPT is an OpenAI project that gives you a free text chat interface to an OpenAI language model that is noticeably better at staying on topic than past generations of text AI. And that was it! Everyone had fun and no one lost their mind.
J/k, the spectrum of unhinged online reaction ranged from “Lensa is going to put your local neighborhood portrait painter out of business” (after all the expensive art classes they took!) to “omg we literally just split the atom and the lamestream media refuses to cover it!” Egg started jabbering something about “World War Orwell.”
So are the robits coming for your jerb? Noah Smith called the new crop of AI tools “autocomplete for everything,” which is… kind of right. Sam from the Tabs Discord described ChatGPT’s vibe as "bored 12th grader who's extremely smart but has major senioritis and will kind of bullshit the assignment." One thing users quickly discovered is that ChatGPT is pretty stupid unless you specifically tell it not to be stupid:
Another discovery was that the easiest way to get around OpenAI’s content filters is to ask the bot how it would evade its own content filters. Or just ask it to write some code or make some tables instead.
Seeing people trick ChatGPT into getting around the restrictions OpenAI placed on usage is like watching an Asimov novel come to life.
— Dare Obasanjo🐀 (@Carnage4Life)
3:06 PM • Dec 1, 2022
What ChatGPT is unequivocally good at is generating bullshit. This thread is the best example I’ve seen of that:
…Noah imagines that at some point, his workflow will look like this: First, he’ll think about what he wants to say, and type out a list of bullet points. His AI word processor will then turn each of these bullet points into a sentence or paragraph, written in a facsimile of Noah’s traditional writing style.
It’s very funny for a professional writer to admit that his job is to come up with some ideas and then encase them in low-value bullshit that a bot could generate, but I do think he’s right, and that generating necessary bullshit is the worst part of many jobs. Writing grants, writing software project proposals, writing pitch decks—lots of jobs consist of coming up with an idea and then coating it in bullshit of a particular format. Lawyers do almost nothing else. A lot of people could nearly eliminate the worst parts of their job with a good bullshit generator, as long as everyone just stays cool and pretends we’re not all using AI to do that. So don’t be a narc.
“Andreessen Horowitz's buzzy tech publication Future is shutting down” vs. “Astra Magazine Had Creative Freedom and a Budget. It Wasn’t Enough.” You win some, you lose some.
Caroline Ellison spotted in NYC. 👀 Zeke Faux’s Bloomberg mea culpa with SBF in the Bahamas penthouse was a banger. “Raccoons Get a Reputation Makeover.” Nice try, scientists, but absolutely not. Stewart leaving Salesforce. Tin-smithing is in his blood.
Today’s Song: Cults, “Always Forever”
She’s not wrong! Buy a t-shirt. Play some tunes. Subscribe to join us in the Discord. I can’t believe it’s only Tuesday.
Reply