- Inclined.ai
- Posts
- 🤖 AI's Role In The Writer Strike
What's up? You're reading Inclined AI. A newsletter that’s like cake for breakfast—a treat that’s right for every occasion.
Here's the icing on your news today:
The WGA strike ties into AI anxiety
Emerging AI abilities might be all hype
Samsung bans generative AI
IBM & Chegg get an AI shockwave
❓ QUICK QUESTION❓
How often are you using AI tools? You can reply directly and let us know 😊
THE WRITER’S STRIKE IS ABOUT MORE THAN FAIR PAY
Hollywood’s brightest writers will sit out until studios meet the demands of the WGA union.
That headline may not sound like an AI news story, but there’s more than meets the eye.
Prompt: robot and human solidarity Style: Vintage Photo
The basics:
They want fair compensation and a change to their residuals agreement. If you don’t know, residuals are checks made to writers for extra earnings outside the scheduled show.
If they sell DVDs of a season or rebroadcast the show, writers get paid. But streaming changed viewer habits, taking a massive bite out of that benefit.
The new deal would give writers a cut of that residual revenue, but the change in pay makes studios hesitate (they want to cut spending, not increase it).
AI’s role in all this:
For starters, these writers aren’t stupid. They want guaranteed protection from new AI tools.
AI is changing the way we write and create content. Unions aren’t going to roll over and let it take all these excellent Hollywood jobs without a fight.
Moreover, the other side of the negotiation table is trying to claw back lost time from the strike. The new scab in town is ChatGPT, not non-WGA writers.
It’s not confirmed, but you have to imagine that these productions will consider using AI scripts to fill in the gaps.
The question is how soon they try it and what kind of content comes out on the other end.
RESEARCHERS AT STANFORD DON’T BUY THE AI HYPE CYCLE
AI’s emergent abilities get everyone excited. We like to imagine a world where Large Language Models learn amazing things without our guidance.
Sundar Pichai, the CEO of Google, claims AI models can learn information outside their programming. A 60 Minutes broadcast in April even mentioned that Bard learned Bengali without the language in its training data.
Is Chat AI teaching itself?
Maybe, but a group of Stanford researchers are skeptical. Their argument centers around measuring success.
Researchers calculate AI’s success at specific tasks in two main ways: continuous (linear) and discontinuous (nonlinear).
Continuous metrics are like judging a basketball player by their points per game.
Discontinuous measurements are a 1 or 0 score where the goal is 10 points. Players who reach that mark or higher score a 1, and the ones who don’t get a 0.
You can see where discontinuous metrics have flaws, so why are we using them to determine abilities in AI?
I recommend reading this article for a deeper dive. But let’s find a way to tie a bow to this news.
In essence, when the researchers examined claims of emergent abilities in AI like that “Spark of General Intelligence” paper from Microsoft. They found that switching to linear, continuous metrics eliminates that random jump.
It’s a predictable climb as the models increase in size, like a rookie basketball player getting 0s for three years jumping to a 1 out of the blue.
But, in reality, they improved with experience, and points per game would reflect that better.
Savvy?
Quick Nuggets
📉 Chegg’s troubles are growing now that ChatGPT is officially biting into their revenue
💧 An April data leak exposed Samsung's secrets and forced them to ban generative AI tools
📓Poor Wikipedia is being torn apart by AI developments
✋ Microsoft's private alternative to current AI may prevent data leaks
🧨 Automating control and giving it to AI is a temptation to avoid, says The Atlantic
📦 Inside the Box, the thinking is that AI tools are the future
🏗️ AI is nothing but a builder, but here it dreamt it was an architect
🚪 Nextdoor adds new Assistant feature powered by ChatGPT
😬 IBM pauses hiring of real people in favor of AI alternatives
⌛️ A quick rundown on the spam sites using AI to take over
📌 Get OpenAI to stop collecting your personal data and training models with it
🤔 Amnesty Norway uses AI images to portray real-life events
🤙 Casual reasoning from LLMs might bring up a new space to pursue
👔 Company execs are starting to realize how many jobs AI can replace right now
🍎 Sal Khan, the CEO of Khan Academy, explains how AI will change teaching
🔥 Fresh Products
Featured Pick: Orimon.ai is a no-code chatbot that lets you build & deploy a sales-enabling, generative AI bot for your website in under 2 minutes. It supports over 150 languages and dialects. (link)
JamGPT - AI that debugs your code with you (link)
Quadratic - adds AI into their collaborative spreadsheets (link)
TeamGPT - power your office workspace with AI (link)
Userpilot - grow faster with better user onboarding (link)
Byrdhouse - multilingual video conferencing (link)
Twifts - find the best gift using Twitter (link)
JourneyGenie - AI plans your travel for you (link)
Adventure Stories - CYOA for your kids (link)
Literally Anything - AI will make any tool you want (link)
ChilloutAI - use Midjourney in your browser (link)
Study Crumb - launches a new AI grammar checker (link)
PopcornAI - never watch the same movie again (link)
ES.AI - write your college applications with AI helping (link)
Chatty Cats - the cutest way to talk with AI (link)
Good Content, Galatic Gala
It’s a striking statement by Xenomorph, but I’m not sure the look is what the Met had in mind with their theme.
That’s it for today. I hope you enjoyed the latest edition of inclined.ai - Davis.
Check Out Our Sunday Edition
Every week you can dive into more AI news and topics with us by subscribing to our premium edition.
We’ve written about:
If you’re not already subscribed, that’s okay. We’re offering a free 7-day trial so you can read this one. That’s how excited I am to post it.