- Big Brain
- Posts
- 🧠 OpenAI CEO says we are close to "scary AI"
🧠 OpenAI CEO says we are close to "scary AI"
PLUS: Humans are better than AI at Go again
Alex Valaitis
February 21, 2023
Happy Tuesday,
And welcome to another edition of Big Brain!
We are the largest AI newsletter in the world (and hopefully the most entertaining as well).
🔦 Today's AI spotlights:
OpenAI CEO says we are close to “scary AI”
Humans are better than AI at Go again
Roblox begins using generative AI
📢 Sponsorship opportunity:
Want to get your AI product, company, or job in front of 400K+ people? Fill out this sponsor intake form and we will get back to you in 24 hours!
🔦 OPENAI CEO SAYS WE ARE CLOSE TO “SCARY AI”
There’s been a running joke in this newsletter, that we’re waiting for our AI overlords to take over…
But there might be more truth to it than we realized.
At least, that is what OpenAI CEO Sam Altman believes.
In a series of tweets, Sam warned about the pace of AI adoption.
He stated that we are “not that far away from away from potentially scary” AI tools. 😬
He also stated that “society needs time to adapt to something so big.”
Some might point out the hypocrisy in these tweets. After all, OpenAI’s ChatGPT has greatly accelerated the rate of AI adoption.
In fairness, Sam has been calling for regulation of AI since 2015.
There are also other signs of companies throttling AI adoption. Microsoft recently limited their Bing AI chatbot to 5 questions per session.
🔦 HUMANS ARE BETTER THAN AI AT BOARD GAME GO AGAIN
In 2016, DeepMind shocked the world when it’s AI program beat the world Go champion in four games.
The champion, Lee Sedol, retired just three years later. His reason?
There was no point playing “an entity that cannot be defeated.”
Many people considered DeepMind’s victory, a monumental moment for AI. It was widely assumed that humans would never beat AI at Go again.
But 7 years later, it appears that humans are back on top!
Kellin Pelrine, recently beat a top AI program in Go, by taking advantage of a weakness in the program.
He was able to exploit the program, since it had not been trained on enough similar games to realize the vulnerability.
The irony is that this vulnerability was discovered by a different AI program.
The company called FAR AI, was launched to look for vulnerabilities in AI systems.
This story could give us a glimpse into the future of human-AI interactions.
Instead of constantly trying to figure out which is better, we should accept that we work best in tandem. 🤝
🔦 ROBLOX BEGINS USING GENERATIVE AI
One of the largest gaming platforms in the world is leaning into AI.
In a blog post last Friday, Roblox laid out their vision for generative AI.
According to the post, the company will be rolling out tools that allow users to alter in-game objects, characters, and more.
These alterations won’t require users to write code. Instead, the users will simply need to give text prompts.
Cheat codes have long been part of video game culture, but this is the first time I’ve seen a company fully embrace customization to this degree.
Could this be a sneak peek into a future in which everyone is able to run a slightly different version of a video game?
📰 OTHER HEADLINES & STORIES
🖼️ AI ART PIECE
NY night, light trail --ar 16:9 --q 2 --s 750 --uplight -by @iron28
🤖 CHATGPT PROMPT OF THE DAY
Can’t imagine why ChatGPT keeps forgetting to include his title 😅 (Via Reddit)
How would you rate today's newsletter? |