- AI Authority
- Posts
- Can OpenAI Solve ChatGPT's Biggest Problem? 🧠
Can OpenAI Solve ChatGPT's Biggest Problem? 🧠
Your Daily AI Rundown — One Headline, Two Tools and Three Community Creations
Ghost
June 01, 2023
Good morning, AI explorers!
Welcome to another edition of AI Authority, delivering all of the top groundbreaking AI news shaping our world into a bite-sized newsletter every weekday morning.
As always, we’ve got one daily AI headline, two exciting tools, and three pieces of content we love from the community for you today.
In this edition:
✍️ News: OpenAI Fights Hallucinations with Math
🛠️ Tools: Longshot, HeyGen
🎨 Creator Corner: Vegetable Fun, BroGPT, Terracotta Warrior Woes, Shaq Shaker AI Commercial
Let’s dive in 👇
P.S. — We added a few new sections to our website to keep you even more informed on the latest in AI:
Daily News Report — never miss a moment of the action with a roundup of all our Twitter posts.
AI Community — a curated list of all the top creators in the space, including resources for ChatGPT, Midjourney, and other great newsletters to check out.
One top headline in the AI space to know:
🧮 OpenAI Fights Hallucinations with Math
OpenAI is stepping up its game to combat the fabrications and missteps that AI models sometimes churn out in moments of uncertainty, often referred to as ‘hallucinations’.
How? By introducing a new training method dubbed 'process supervision'. This innovative approach rewards the AI for each individual correct step it takes towards an answer, not just the final conclusion.
The company released a new paper on Wednesday breaking down the study, which saw OpenAI train a model to achieve a new gold standard in mathematical problem solving using the new process.
OpenAI fighting hallucinations with math
Critics of AI have been quick to point out that AI-generated misinformation can pose some very serious issues — a particularly pressing issue for regulators with the 2024 U.S. presidential election on the horizon.
But by encouraging models to follow a human-like chain of thought, OpenAI hopes process supervision could be the key to more accountable, less misleading AI.
Still hungry for more AI headlines? Some other news that caught our eye yesterday:
— U.S., EU Working on Voluntary AI Code of Conduct (link)
— AI Identifies Gene Interactions, Speeds Treatment Target Search (link)
— AI Godfather Yoshua Bengio Feels Lost Over Life’s Work (link)
— Fight Intensifies over Fake Research ‘Paper Mills’ With AI (link)
— GoDaddy Product Launches Unlock Power of Generative AI (link)
Two AI-powered platforms to add to your toolbox:
Three of our favorite tweets, threads, AI generations and more:
Taking vegetables to a whole new level. A fun food prompt from @alillian:
"Don't play with your food"
Wanna play?
Basic prompt: food photography of [X] ultra realistic, ultra detailed, UHD image --style raw --s 750— Ashley McAuley (@alillian)
12:58 AM • May 30, 2023
This user unlocked BroGPT! Broseidon 💀
Terracotta Warriors doing mundane life tasks:
The daily life of a Terracotta Warrior! 🤣 #midjourney #aiartcommunitytwitter.com/i/web/status/1…
— Cakedroid (@Cakedroid)
3:21 PM • May 30, 2023
Bonus: A new ridiculous AI commercial. Are you buying a Shaq Shaker?
Ready for another terrifying (but hilarious) AI-generated commercial?
The Shaq Shaker is here:
— AI Authority (@Authority_AI)
7:44 PM • May 31, 2023
See you tomorrow!
If you haven’t already, subscribe to receive AI Authority in your inbox every weekday. And share it with an AI-curious friend!
