- Inclined.ai
- Posts
- 🤖 Confronting Singularity: Hope Or Doom
🤖 Confronting Singularity: Hope Or Doom
PLUS: Generative AI Still Has a Bias Problem
What's up? You're reading Inclined AI. Solidarity for all the Subreddits going dark today on Reddit to speak out against their new API policy.
Here's the AI news for today:
What in the world is the “Singularity”
Generative AI can highlight the bad parts of humanity
Meta announces new AI plans
will.i.am believes in everlasting Taylor Swift
THE TERM WE ALL DANCE AROUND
a book illustration of a silhouette standing near the end of a tunnel surrounded by white light --ar 2:1 --s 400 --chaos 1
Singularity is near. You’ve heard that term, right? The long history of AI research etched the idea into all things machine learning.
After all, what is this all for? Why do we feel that buzz of electricity when something like ChatGPT comes along?
The goal is a seismic change in how our society works. We are working towards singularity, but even that goal does not encapsulate that term.
Singularity is identifiable. The moment passed that point is not.
The New York Times published an article this weekend discussing the discrepancy. You should read it to learn about Singularity's key players and history (unpaywalled here because it’s a term that’s important).
But let’s try out best to give you dinner table material:
Singularity is when technology exceeds human capabilities and begins to self-improve with its agency. That’s the clinical term.
To say it another way, AI will grow uncontrollably and make choices on its own.
But that’s it. The outcome is subjective. Some people think that metaphorical moments will lead to the end of humanity.
If the computer examines us, it will demote us to canon fodder.
But the positive side flips the script.
If we align AI to understand us and work with us, it can help us advance. The singularity is a bright moment where we find a new purpose and life changes for the better.
We’re in the middle of a tunnel running towards a light. That light is the singularity at the end of the tunnel.
No one knows what happens on the other side.
But we expect the best and brace for the worst here at Inclined AI.
Let me tell you about magic. Have you ever witnessed magic?
I didn’t think I had until I tried Plus and their new AI for Google Slides tool. The results dazzled my senses and made me feel like a kid in a candy store. Let me explain.
Plus created a unique extension for Google Slides that harnesses generative AI to craft an entire presentation in seconds. All you do is type in the prompt, select a theme, the order the slides.
Bam! You’re all done making that would-be stressful presentation, and you have Plus AI for Google Slides to thank. But it gets better.
Once you generate the slides, their AI tool goes a step further and gives you tips on each one, highlighting things you should add to go above and beyond. It doesn’t make the slides then leave you hanging for the clean-up.
Plus goes above and beyond. Their tool is something you’ll actually use in your day-to-day work.
DON’T LET BIAS ISSUES SLIP THROUGH THE CRACKS
Generative AI images influence you and create standards and opinions. We don’t realize it’s happening all the time, but it is.
AI content is consuming the internet. These Large Language Models and Diffusion Models are overwriting us and reinventing most of the web in their image.
The issue starts there, not ends.
The people who train these models influence AI. That influence permeates through the system.
The trouble is that most of these builders come from one group of life experiences. It’s a common tech problem.
That’s why the Bloomberg piece about the Stable Diffusions bias issue is unsurprising (unpaywalled here because it is important). But the concern is the inertia in the development of AI.
Researchers are working to address this issue because it reflects itself in many ways. But that takes time—time that moves slower than AI learns and reinforces current biases.
That’s the rub. What will the next generation of models reflect? We can’t say for sure, but the response will take time to correct if it’s negative.
We can expect better, but the jury will be out for awhile.
Quick Nuggets
🤝 Sam Altman is considering a call to collaborate with China to promote AI safety
🏳️🌈 ChatGPT writes a web series called “Straight Best Friend” about a gay man and his straight friend
🏈 A sportswriter uses ChatGPT to predict his college football team’s record this year
💰 Oracle finds the answer to their Microsoft-OpenAI problem and its Cohere
💸 These AI jobs pay up to $375k a year
🧓 Older people are looking towards AI to feel less lonely
🤷 Oops, AI art destroyed this indie book cover contest
🤔 WIRED writes a headline with big words like “Hyperdimensional Computing”
✨ will.i.am thinks artists like Taylor Swift will never die thanks to AI
📢 Meta announces even bigger plans for their AI tooling
🔥 Fresh Products
WINN AI - takes the busywork out of sales (link)
VoiceGPT - talk w/ GPT-4 on your Apple Watch (link)
Zizoto - mixes generative art w/ a social network (link)
Balance - lower acute anxiety w/ the help of AI (link)
Deepshot - helps automatically resync lips for new dialogue (link)
AI Spend - monitor your API key usage on OpenAI (link)
caspa AI - product photos are done using AI (link)
mesha - reinvents their financial command center by adding AI features (link)
Good Content, Shrek Reimagined
I know what you’re thinking, “what does Donkey look like?” Trust me; you don’t want to know.
That’s it for today. I hope you enjoyed the latest edition of inclined.ai - Davis.
Check Out Our Sunday Edition
You can dive into more AI news and topics with us every week by subscribing to our premium edition.
We’ve written about the following:
If you’re not already subscribed, that’s okay. We’re offering a free 7-day trial so that you can read this one. That’s how excited I am to post it.